As someone who works in a University, I am amazed UNIX vendors don't
give away free copies of their operating systems to students and home
users. If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
their OS releases, home users would be more likely to buy cheap used
HP kit. When those home users go in to industry, they are more likely
to buy **EXPENSIVE** new HP kit. So it's to HP's advantage to give
away HP-UX to students and home users.

The same can be said for any UNIX hardware vendor - I strongly suspect
it's in their long term economic interest to encourage people to learn
their operating systems. Once you know one commercial UNIX operating
system, you are likely to want to stick to it. 

Why on earth Sun don't change the licence conditions of their
so-called free licence for Solaris on SPARC, I don't know. Solaris can
be freely downloaded, yet its illegal to use it on virtually all used
Sun kit sold. Their previous licence, which allowed for its use on
machines of up to 8 CPUs seemed more sensible. 

HP have a reasonably attractive package that allows non-commerical
users to gain the latest copy of the Tru64 operating system and the HP
C compiler for Tru64 for $100. Yet HP have no such policy with their
other UNIX operating system (HP-UX). 

I'd like to know HP's logic of the differing policies of Tru64 and
HP-UX. It seems very odd to sell Tru64 and compiler for $99, yet give
no discounts on the HP-UX.

One can't expect Microsoft to give Windoze away for free, as Microsoft
will never gain any extra revenue from hardware sales. But SGI, HP,
IBM, Sun etc. will gain extra sales of hardware, by allowing home
users to learn their operating systems, rather than that of their
competitors. 

One can't expect SCO to give away SCO UNIX, as again they will gain no
extra revenue from hardware sales, although they might have done with
software sales had Linux never existed. But I don't think any home
user or student would care less whether or not SCO was given away now. 



-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

On Sun, 27 Jul 2003 03:44:04 +0100, Dr. David Kirkby wrote:

> HP have a reasonably attractive package that allows non-commerical
> users to gain the latest copy of the Tru64 operating system and the HP
> C compiler for Tru64 for $100. Yet HP have no such policy with their
> other UNIX operating system (HP-UX). 

And the compiler?  As of V5.1b, the compiler has a -separate- license.
 
> I'd like to know HP's logic of the differing policies of Tru64 and
> HP-UX. It seems very odd to sell Tru64 and compiler for $99, yet give
> no discounts on the HP-UX.

HP-UX has a "future" if their idiotic devotion to Itanic does not kill the
server branch of the company.  Tru64 has no future at all; if HPQ can get
$99 for the 3 basic licenses it is a windfall.
 
> One can't expect Microsoft to give Windoze away for free, as Microsoft
> will never gain any extra revenue from hardware sales. But SGI, HP,
> IBM, Sun etc. will gain extra sales of hardware, by allowing home
> users to learn their operating systems, rather than that of their
> competitors. 

Microsoft's student discounts to almost all university students
essentially amounts to giving away the "OS" and many apps for free.
 
> One can't expect SCO to give away SCO UNIX, as again they will gain no
> extra revenue from hardware sales, although they might have done with
> software sales had Linux never existed. But I don't think any home
> user or student would care less whether or not SCO was given away now.

SCO is irrelevant and has been so for 10 years.  The company will not
survive their legal action against IBM and every other user of Unix and
Linux, Solaris excepted for good financial reasons.

Scott Howard <scott@hunterlink.net.au> wrote:
> In comp.unix.solaris Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
>> As someone who works in a University, I am amazed UNIX vendors don't
>> give away free copies of their operating systems to students and home
>> users. If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give

> Sun do.  Sure, it's (legally) limited to one CPU, but that's not generally
> going to be a big restriction.

It's legally limited to one CPU _AND_ , in case of SPARC systems, the
machine must have been sold by either Sun or an authorized
Sun-reseller as stated at the bottom of this page:
http://wwws.sun.com/software/solaris/licensing/policies.html

Moreover, this page states that the right to use Solaris is not
transferable when the machine is resold with the exceptions of a few
authorized vendors. This means that nearly all used Sparc systems sold
on say on ebay today do not come with a valid license to use Solaris
and you legally CANNOT use the software licensed under the Free Binary
Solaris License program on those machines. You have to pay for the
Solaris Upgrade RTU license on such machines, even for a system with
one CPU, regardless of what you're using it for. The RTU license for a
single-proc system costs about $100. This is why the Solaris/sparc
enthusiast community is outraged. In many cases, the price of this
upgrade RTU license costs a large fraction of or even more than the
price of many used Sun systems that are resold today. I think it's
fair to slap this cost on the businesses and other organizations who
use this software to run their day to day tasks but it's unfair
towards the hobbyists and enthusiasts who don't derive a commercial
gain from using those systems and who actively promote the usage of
Sun hardware and software outside of their homes.


-- 
Akop Pogosian

This space has been accidentally left blank.

"Dr. David Kirkby" wrote:
> 
> As someone who works in a University, I am amazed UNIX vendors don't
> give away free copies of their operating systems to students and home
> users.

	Why? Do Microsoft do this? Perhaps you should widen your
	job experience.

> If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> their OS releases, home users would be more likely to buy cheap used
> HP kit.

	How do you conclude that?

> When those home users go in to industry, they are more likely
> to buy **EXPENSIVE** new HP kit.

	What makes you think they would be in a position to do so?
	Exactly how many home users go from there immediately to
	management positions with buying power?

> So it's to HP's advantage to give
> away HP-UX to students and home users.

	I can't see how you can draw that conclusion.

> The same can be said for any UNIX hardware vendor - I strongly suspect
> it's in their long term economic interest to encourage people to learn
> their operating systems. Once you know one commercial UNIX operating
> system, you are likely to want to stick to it.

	That sounds more like conjecture. What evidence do you have
	to support that claim?

> Why on earth Sun don't change the licence conditions of their
> so-called free licence for Solaris on SPARC, I don't know. Solaris can
> be freely downloaded, yet its illegal to use it on virtually all used
> Sun kit sold. Their previous licence, which allowed for its use on
> machines of up to 8 CPUs seemed more sensible.

	Why? How many home users/students have 8 CPU systems?

> HP have a reasonably attractive package that allows non-commerical
> users to gain the latest copy of the Tru64 operating system and the HP
> C compiler for Tru64 for $100. Yet HP have no such policy with their
> other UNIX operating system (HP-UX).

	Your point being?

> I'd like to know HP's logic of the differing policies of Tru64 and
> HP-UX. It seems very odd to sell Tru64 and compiler for $99, yet give
> no discounts on the HP-UX.

	Is it $99 or $100?

> One can't expect Microsoft to give Windoze away for free, as Microsoft
> will never gain any extra revenue from hardware sales.

	But by your previous arguments, said home users of such
	free software would go on to use it commercially.

> But SGI, HP, IBM, Sun etc. will gain extra sales of hardware, by allowing
> home users to learn their operating systems, rather than that of their
> competitors.

	You're argument seems specious to me. One could just as easily
	conclude that it would be better to do as MS does and stick to
	software and forget about hardware.

> One can't expect SCO to give away SCO UNIX, as again they will gain no
> extra revenue from hardware sales, although they might have done with
> software sales had Linux never existed.

	How do you conclude that?

> But I don't think any home
> user or student would care less whether or not SCO was given away now.

	Again, how do you conclude that?

-am	� 2003

Hello Anthony,

sorry you completely misunderstood what David Kirby wrote about the free
license for Solaris 8.

Solaris 8 was absolutely free for all Systems (SPARC and x86) with the
limitation of 8 cpus max, you had only to register the number of systems.
There was *no* restriction regarding second hand systems.

Solaris 9 is *only* free for SPARC with at least 2 substantical limitations:

1. No second hand systems except those from the authorized dealers. You have
to buy the RTU.

2. No multi-processor *capable* systems, that means no SPARCstation 10/20,
Ultra 2/60, etc. To use them you have to buy a RTU.

The x86 version of Solaris 9 is not free at all, you have to pay at least
$20 for download and even more for the media kit.

Michael
Remove the _MAPSON_ from my eMail address.
"Anthony Mandic" <nu@hotmail.com> wrote
news:3F2368E5.ED5037A0@hotmail.com...
> "Dr. David Kirkby" wrote:
> >
> > ....
> > Why on earth Sun don't change the licence conditions of their
> > so-called free licence for Solaris on SPARC, I don't know. Solaris can
> > be freely downloaded, yet its illegal to use it on virtually all used
> > Sun kit sold. Their previous licence, which allowed for its use on
> > machines of up to 8 CPUs seemed more sensible.
>
> Why? How many home users/students have 8 CPU systems?
>
> -am � 2003

Michael Lehmann wrote:
> 
> Hello Anthony,
> 
> sorry you completely misunderstood what David Kirby wrote about the free
> license for Solaris 8.

	I understand the license requirements as per Sun's website
	(the link was posted by someone else in this thread). What
	I don't understand is what you are trying to reply to. You
	don't answer the question I raised.

-am	� 2003

"Dr. David Kirkby" <drkirkby@ntlworld.com> wrote in message news:<3F233C74.4F4C3BAB@ntlworld.com>...
> As someone who works in a University, I am amazed UNIX vendors don't
> give away free copies of their operating systems to students and home
> users. 

They all have licensed bits that they must pay for.  Giveaways would
be at a loss to them.

>If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> their OS releases, home users would be more likely to buy cheap used
> HP kit. When those home users go in to industry, they are more likely
> to buy **EXPENSIVE** new HP kit. So it's to HP's advantage to give
> away HP-UX to students and home users.

Agreed here.  Though it is easy enough to get an OS if you network
around a little.

(snip)

> One can't expect Microsoft to give Windoze away for free, as Microsoft
> will never gain any extra revenue from hardware sales. But SGI, HP,
> IBM, Sun etc. will gain extra sales of hardware, by allowing home
> users to learn their operating systems, rather than that of their
> competitors. 

Agree here, but you forget Linux.  

(snip)
> But I don't think any home
> user or student would care less whether or not SCO was given away now. 

Their legal stunts lately will likely create this attitude everywhere
no matter the product quality.

Your best free UNIX is Linux.  Students that learn to use Open Source
code will be able to do much of their computing for free on just about
any hardware they can find.  If said students also realize the
implications of Open Standards, they will own what they create as
well.

I understand your position regarding educational licenses, but do not
expect to see things change much in the future.

Anthony Mandic wrote:

|>> Why on earth Sun don't change the licence conditions of their
|>> so-called free licence for Solaris on SPARC, I don't know. Solaris can
|>> be freely downloaded, yet its illegal to use it on virtually all used
|>> Sun kit sold. Their previous licence, which allowed for its use on
|>> machines of up to 8 CPUs seemed more sensible.

|>         Why? How many home users/students have 8 CPU systems?

Dual or Quad processor Sparc-10/20 systems are not _that_ uncommon.

-- 

"I'm a doctor, not a mechanic." Dr Leonard McCoy <mccoy@ncc1701.starfleet.fed>
"I'm a mechanic, not a doctor." Volker Borchert  <v_borchert@despammed.com>

In article <3F233C74.4F4C3BAB@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:


>Why on earth Sun don't change the licence conditions of their
>so-called free licence for Solaris on SPARC, I don't know. Solaris can
>be freely downloaded, yet its illegal to use it on virtually all used
>Sun kit sold. Their previous licence, which allowed for its use on
>machines of up to 8 CPUs seemed more sensible. 

If you buy a used Sun, this machine would have been run wich a license before.
Even if the seller used a corporate license, the machine definitely did come
with a valid license from Sun when bought from Sun.


At least this license could be sold together with the HW. If Sun states 
differenty, this is illegal and thus void.

So the real problem would be that Sun does not allow home users to upgrade
the Solaris release to a recent current version for free.




-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

In article <pan.2003.07.27.03.16.21.84665@yahoo.com>,
Dave Uhring  <daveuhring@yahoo.com> wrote:
>On Sun, 27 Jul 2003 03:44:04 +0100, Dr. David Kirkby wrote:
>
>> HP have a reasonably attractive package that allows non-commerical
>> users to gain the latest copy of the Tru64 operating system and the HP
>> C compiler for Tru64 for $100. Yet HP have no such policy with their
>> other UNIX operating system (HP-UX). 
>
>And the compiler?  As of V5.1b, the compiler has a -separate- license.

From SCO, there is a student licese for the SCO C-compiler for free,
The SCO compiler is comparable to the Sun compiler if you compare both
against GCC. Sun's compiler creates x86 code up to 30-35% faster than
code from GCC and SCO's compiler gives a few percent more speed.

So there is an example for a free compiler home license.
 
>Microsoft's student discounts to almost all university students
>essentially amounts to giving away the "OS" and many apps for free.

Which is the real problem.

Win32 is not POSIX and dfrom my understanding of research/education, 
universities should not teach industry proprietary OS kow how but
the official standard.

>> One can't expect SCO to give away SCO UNIX, as again they will gain no
>> extra revenue from hardware sales, although they might have done with
>> software sales had Linux never existed. But I don't think any home
>> user or student would care less whether or not SCO was given away now.
>
>SCO is irrelevant and has been so for 10 years.  The company will not
>survive their legal action against IBM and every other user of Unix and
>Linux, Solaris excepted for good financial reasons.

Sure? 

-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

In article <bfvn5p$2p4u$1@agate.berkeley.edu>,
Akop Pogosian  <akopps+usenet@ocf.berkeley.edu> wrote:
>Scott Howard <scott@hunterlink.net.au> wrote:
>> In comp.unix.solaris Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
>>> As someone who works in a University, I am amazed UNIX vendors don't
>>> give away free copies of their operating systems to students and home
>>> users. If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
>
>> Sun do.  Sure, it's (legally) limited to one CPU, but that's not generally
>> going to be a big restriction.
>
>It's legally limited to one CPU _AND_ , in case of SPARC systems, the
>machine must have been sold by either Sun or an authorized
>Sun-reseller as stated at the bottom of this page:
>http://wwws.sun.com/software/solaris/licensing/policies.html
>
>Moreover, this page states that the right to use Solaris is not
>transferable when the machine is resold with the exceptions of a few
>authorized vendors. This means that nearly all used Sparc systems sold
>on say on ebay today do not come with a valid license to use Solaris
>and you legally CANNOT use the software licensed under the Free Binary
>Solaris License program on those machines. You have to pay for the

Sun's statements are illegal and thus void!

Sun is not allowed to prohibit people to sell a machine _with_ a valid license.

-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

Doug Dingus wrote:

> Agree here, but you forget Linux.

	Yes, I agree, you can forget Linux.

> Your best free UNIX is Linux.

	Oh? And I suppose you can point us all to Linux's UNIX certification?

> Students that learn to use Open Source code will be able to do much
> of their computing for free on just about any hardware they can find.

	How do you figure that?

-am	� 2003

"Anthony Mandic" <nu@hotmail.com> wrote in message
news:3F2368E5.ED5037A0@hotmail.com...
> "Dr. David Kirkby" wrote:
> >
> > As someone who works in a University, I am amazed UNIX vendors don't
> > give away free copies of their operating systems to students and home
> > users.
>
> Why? Do Microsoft do this? Perhaps you should widen your
> job experience.
>
50 UK pounds for Windows 2000 student license, compare to what I paid
for HP-UX 11i for 100pounds from eBay + shipping, Irix 6.5 base with
6.5.2 and 6.5.13 update for 75pounds + shipping, both WITHOUT license,
media kits only.
My second copy of Windows 2000 was given to me for free by MS UK.

> > If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> > their OS releases, home users would be more likely to buy cheap used
> > HP kit.
>
> How do you conclude that?
>
Since Dr. Kirkby told me how fast his HP C3600 is compare to his
quad 450MHz U80, I have been trying to find a HP machine too.
For 1000US, you can find an HP C3600 from eBay with faster
processor than a single 450MHz U60/80 which tend to have higher
price tag.
BUT, you can downloads Solaris ISO from Sun but not HP-UX.

> > When those home users go in to industry, they are more likely
> > to buy **EXPENSIVE** new HP kit.
>
> What makes you think they would be in a position to do so?
>
People get used to the hardwares and softwares they use, I can't get over
using Outlook Express as my default news client, even got IE5 on my
Sol8 machine, now guess what I will use as my news reader 10 years later.

> Exactly how many home users go from there immediately to
> management positions with buying power?
>
Not a lot, but guess how MS become so dominent.

> > So it's to HP's advantage to give
> > away HP-UX to students and home users.
>
> I can't see how you can draw that conclusion.
>
Quote from Uli:
"If the students become consultants, they surely will prefer something they
know by heart.
One big advance of Linux these days."

> > The same can be said for any UNIX hardware vendor - I strongly suspect
> > it's in their long term economic interest to encourage people to learn
> > their operating systems. Once you know one commercial UNIX operating
> > system, you are likely to want to stick to it.
>
> That sounds more like conjecture. What evidence do you have
> to support that claim?
>
Same quote...


> > Why on earth Sun don't change the licence conditions of their
> > so-called free licence for Solaris on SPARC, I don't know. Solaris can
> > be freely downloaded, yet its illegal to use it on virtually all used
> > Sun kit sold. Their previous licence, which allowed for its use on
> > machines of up to 8 CPUs seemed more sensible.
>
> Why? How many home users/students have 8 CPU systems?
>
Not a lot, lets suppose we all buy from a reseller,
a SS10/20 and U2 can take more than 1 CPU, even
a dual CPU U60 goes for under 1000pounds now.

> > HP have a reasonably attractive package that allows non-commerical
> > users to gain the latest copy of the Tru64 operating system and the HP
> > C compiler for Tru64 for $100. Yet HP have no such policy with their
> > other UNIX operating system (HP-UX).
>
> Your point being?
>
> > I'd like to know HP's logic of the differing policies of Tru64 and
> > HP-UX. It seems very odd to sell Tru64 and compiler for $99, yet give
> > no discounts on the HP-UX.
>
> Is it $99 or $100?
>
$1....WOW! that's a lot of money, you can buy a botte of mineral water.


> > One can't expect Microsoft to give Windoze away for free, as Microsoft
> > will never gain any extra revenue from hardware sales.
> >

> But by your previous arguments, said home users of such
> free software would go on to use it commercially.
>
Microsoft = software company = sell softwares for $$$.
Sun, HP, IBM, SGI = hardware companies who sell their own softwares =
selling hardwares for living, not softwares.

> > But SGI, HP, IBM, Sun etc. will gain extra sales of hardware, by
allowing
> > home users to learn their operating systems, rather than that of their
> > competitors.
>
> You're argument seems specious to me. One could just as easily
> conclude that it would be better to do as MS does and stick to
> software and forget about hardware.
>
Irix works on MIPS, HP-UX runs on PA and Itanium, Solaris on Sparc with
limited support on x86.
Now who makes MIPS, PA and Sparc based machine?
If SGI do as MS, you would have possibily the prettiest OS without any
hardwares
to run it on.

> > One can't expect SCO to give away SCO UNIX, as again they will gain no
> > extra revenue from hardware sales, although they might have done with
> > software sales had Linux never existed.
>
> How do you conclude that?
>
The whole software and hardware company thingy.


> > But I don't think any home
> > user or student would care less whether or not SCO was given away now.
>
> Again, how do you conclude that?
>
I am the only person in the entire department of EE.Eng with a copy of SCO
UnixWare


Those are just my opinions

Yuan

Hello all, just to let you know I have a 6 CPU with 2.5GB RAM HP9000/T500 as
my home machine!
And yes, it will currently only run HP-UX as theres not enough effort going
in to get Linux running on it. And no, I didn't pay the $500 licence fee per
CPU (oddly enough!).
From my own view point its a fantastic way to learn HP-UX which surely is in
HP's best interest? If I hadn't saved this monster it surely would have got
scrapped and whats the point of that? The fact is that the machine would
have originally had a HP-UX licence with it so I have no problem with
"illegally" running HP-UX on it now. What possibe use is HP-UX without the
HP hardware? nothing at all. Are there any HP hardware clones? I don't think
so!
It's not like I'll be phoning HP for tech support either so it costs HP
nothing.

So if you have HP hardware then HP-UX really ought to be FREE as theres just
no point otherwise.

"Anthony Mandic" <nx@hotmail.com> wrote in message
news:3F23A005.83B7C2DA@hotmail.com...
> Volker Borchert wrote:
>
> > |>         Why? How many home users/students have 8 CPU systems?
> >
> > Dual or Quad processor Sparc-10/20 systems are not _that_ uncommon.
>
> Exactly! 8 CPU home systems would be rarer still.
>
> -am � 2003

Yuan wrote:

> > > As someone who works in a University, I am amazed UNIX vendors don't
> > > give away free copies of their operating systems to students and home
> > > users.
> >
> > Why? Do Microsoft do this?
> >
> 50 UK pounds for Windows 2000 student license,

	Which isn't exactly free.

> My second copy of Windows 2000 was given to me for free by MS UK.

	That's sounds more like it. So they do do it.

> > > If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> > > their OS releases, home users would be more likely to buy cheap used
> > > HP kit.
> >
> > How do you conclude that?
> >
> Since Dr. Kirkby told me how fast his HP C3600 is compare to his
> quad 450MHz U80, I have been trying to find a HP machine too.
> For 1000US, you can find an HP C3600 from eBay with faster
> processor than a single 450MHz U60/80 which tend to have higher
> price tag.

	Ah! So that explains his troll post.

> BUT, you can downloads Solaris ISO from Sun but not HP-UX.

	That's good. So what?

> > > When those home users go in to industry, they are more likely
> > > to buy **EXPENSIVE** new HP kit.
> >
> > What makes you think they would be in a position to do so?
> >
> People get used to the hardwares and softwares they use,

	So? That doesn't explain how they go from being a home user/student
	to being in a management position with purchasing power. Unless ...
	whoever hired them was a complete and utter idiot.

> I can't get over using Outlook Express as my default news client,

	My condolences.

> > Exactly how many home users go from there immediately to
> > management positions with buying power?
> >
> Not a lot, but guess how MS become so dominent.

	I'm not here to guess. So don't patronise me with specious
	statements. Show me facts and figures.

> > > So it's to HP's advantage to give
> > > away HP-UX to students and home users.
> >
> > I can't see how you can draw that conclusion.
> >
> Quote from Uli:
> "If the students become consultants, they surely will prefer something they
> know by heart. One big advance of Linux these days."

	Linux isn't HP. And quoting someone else's misinformed statement
	doesn't make it any less misinformed. Show me some proof. How do
	those who come to prefer HP do so when they don't use it as a
	student or home user?

> > > The same can be said for any UNIX hardware vendor -
> >
> > That sounds more like conjecture. What evidence do you have
> > to support that claim?
> >
> Same quote...

	Same quote what?

> > Why? How many home users/students have 8 CPU systems?
> >
> Not a lot,

	Exactly.

> lets suppose we all buy from a reseller,
> a SS10/20 and U2 can take more than 1 CPU, even
> a dual CPU U60 goes for under 1000pounds now.

	So? How does that constitute an 8 CPU system? At most its only 2.

> $1....WOW! that's a lot of money, you can buy a botte of mineral water.

	What's a botte?

> > > One can't expect Microsoft to give Windoze away for free, as Microsoft
> > > will never gain any extra revenue from hardware sales.

	Looks like you've just proven "Kirbu" wrong.

> Microsoft = software company = sell softwares for $$$.
> Sun, HP, IBM, SGI = hardware companies who sell their own softwares =
> selling hardwares for living, not softwares.

	Your point being?

> Irix works on MIPS, HP-UX runs on PA and Itanium, Solaris on Sparc with
> limited support on x86.
> Now who makes MIPS, PA and Sparc based machine?
> If SGI do as MS, you would have possibily the prettiest OS without any
> hardwares to run it on.

	Your point being? MS don't appear to have any problems with that.
	Why do you think SGI would?

> > > One can't expect SCO to give away SCO UNIX, as again they will gain no
> > > extra revenue from hardware sales, although they might have done with
> > > software sales had Linux never existed.
> >
> > How do you conclude that?
> >
> The whole software and hardware company thingy.

	You haven't presented anything by way of a concrete argument
	yet. All I'm seeing is conjecture. How about you show me the
	money?

> > > But I don't think any home
> > > user or student would care less whether or not SCO was given away now.
> >
> > Again, how do you conclude that?
> >
> I am the only person in the entire department of EE.Eng with a copy of SCO
> UnixWare

	My, aren't you lucky. But so what?

> Those are just my opinions

	They certainly are. And that's the trouble. Too much baseless
	opinion and no facts.

-am	� 2003

On Sun, 27 Jul 2003 09:36:28 +0000, Joerg Schilling wrote:

> In article <pan.2003.07.27.03.16.21.84665@yahoo.com>,
> Dave Uhring  <daveuhring@yahoo.com> wrote:

>>And the compiler?  As of V5.1b, the compiler has a -separate- license.
> 
> From SCO, there is a student licese for the SCO C-compiler for free,
> The SCO compiler is comparable to the Sun compiler if you compare both
> against GCC. Sun's compiler creates x86 code up to 30-35% faster than
> code from GCC and SCO's compiler gives a few percent more speed.
> 
> So there is an example for a free compiler home license.

But that is -not- the C/C++ compiler for Tru64 which the OP mentioned.

>>SCO is irrelevant and has been so for 10 years.  The company will not
>>survive their legal action against IBM and every other user of Unix and
>>Linux, Solaris excepted for good financial reasons.
> 
> Sure?

About which statement?  The Solaris exception is explained here:

http://news.com.com/2100-1016_3-1024633.html?tag=fd_top

Hello Anthony,

your question regarding home users with systems with more then 8 CPUs didn't
meet the point. You could not expect to get an answer !

The difference between the Solaris 8 and Solaris 9 license agreements, is
that the former is less restrictive in important aspects (using second hand
systems or those that are dual processor capable). That's the point Mr.
Kirkby made and I commented on.

Michael
Remove the _MAPSON_ from my e-mail address
"Anthony Mandic" <nw@hotmail.com> schrieb im Newsbeitrag
news:3F237F56.85C71B9@hotmail.com...
> Michael Lehmann wrote:
> >
> > Hello Anthony,
> >
> > sorry you completely misunderstood what David Kirby wrote about the free
> > license for Solaris 8.
>
> I understand the license requirements as per Sun's website
> (the link was posted by someone else in this thread). What
> I don't understand is what you are trying to reply to. You
> don't answer the question I raised.
>
> -am � 2003

Michael Lehmann wrote:
> 
> Hello Anthony,
> 
> your question regarding home users with systems with more then 8 CPUs didn't
> meet the point. You could not expect to get an answer !

	Why did you elevate it from 8 to more than 8?

> The difference between the Solaris 8 and Solaris 9 license agreements, is
> that the former is less restrictive in important aspects (using second hand
> systems or those that are dual processor capable). That's the point Mr.
> Kirkby made and I commented on.

	Its not the point "Kirbu" was making. The real point he was
	making is that he's a cheapskate who wants everything for free.

-am	� 2003

Martin Stenzel wrote:

> What I see from your postings here on c.s.hp.hpux is that you want
> _everything_ for free.

	Its not the only place where he wants this. But this appears
	to be the first time he's trolled big time across a lot of
	disparate newsgroups.

> It is everybody's/company's own decision what to to give away for free,
> at what price at what conditions.
> Sounds you are striving for the illegal part....

	It does, doesn't it?

-am	� 2003

"Anthony Mandic" <o0@hotmail.com> wrote in message
news:3F23A822.97A8CCF6@hotmail.com...
> Yuan wrote:
>
> > > > As someone who works in a University, I am amazed UNIX vendors don't
> > > > give away free copies of their operating systems to students and
home
> > > > users.
> > >
> > > Why? Do Microsoft do this?
> > >
> > 50 UK pounds for Windows 2000 student license,
>
> Which isn't exactly free.
>
Cheaper than most Unix OS.


> > My second copy of Windows 2000 was given to me for free by MS UK.
>
> That's sounds more like it. So they do do it.
>
So? it's a special case.


> > > > If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> > > > their OS releases, home users would be more likely to buy cheap used
> > > > HP kit.
> > >
> > > How do you conclude that?
> > >
> > Since Dr. Kirkby told me how fast his HP C3600 is compare to his
> > quad 450MHz U80, I have been trying to find a HP machine too.
> > For 1000US, you can find an HP C3600 from eBay with faster
> > processor than a single 450MHz U60/80 which tend to have higher
> > price tag.
>
> Ah! So that explains his troll post.
>
No, it's just a simple fact that second hand HP kits are not as popular as
Sun kits,
hence the seller can only demand less.


> > BUT, you can downloads Solaris ISO from Sun but not HP-UX.
>
> That's good. So what?
>
You don't like to excersise your brain cells do you?
So you can downloads the ISO, use your favourite CD buring software, and put
a few blank CD-R into your CD-RW, burn from those ISO into bootable CDs as
installation medias, and install Solaris from those CDs, it works out much
cheaper
than buying the media set from the vendors.

> > > > When those home users go in to industry, they are more likely
> > > > to buy **EXPENSIVE** new HP kit.
> > >
> > > What makes you think they would be in a position to do so?
> > >
> > People get used to the hardwares and softwares they use,
>
> So? That doesn't explain how they go from being a home user/student
> to being in a management position with purchasing power. Unless ...
> whoever hired them was a complete and utter idiot.
>
So you were never a student and you were never a home user?


> > I can't get over using Outlook Express as my default news client,
>
> My condolences.
>
My consolences to you too.


> > > Exactly how many home users go from there immediately to
> > > management positions with buying power?
> > >
> > Not a lot, but guess how MS become so dominent.
>
> I'm not here to guess. So don't patronise me with specious
> statements. Show me facts and figures.
>
The same as your non-constructive "so?", "your point being?" comments.
Students buy cheap MS products ---> students grow up only knowing
how to use Windows.

> > > > So it's to HP's advantage to give
> > > > away HP-UX to students and home users.
> > >
> > > I can't see how you can draw that conclusion.
> > >
> > Quote from Uli:
> > "If the students become consultants, they surely will prefer something
they
> > know by heart. One big advance of Linux these days."
>
> Linux isn't HP. And quoting someone else's misinformed statement
> doesn't make it any less misinformed. Show me some proof. How do
> those who come to prefer HP do so when they don't use it as a
> student or home user?
>
He only uses Linux as a fact to show you how it becomes popular.
That's the proof you want.


> > > > The same can be said for any UNIX hardware vendor -
> > >
> > > That sounds more like conjecture. What evidence do you have
> > > to support that claim?
> > >
> > Same quote...
>
> Same quote what?
>
As above.

> > > Why? How many home users/students have 8 CPU systems?
> > >
> > Not a lot,
>
> Exactly.
>
> > lets suppose we all buy from a reseller,
> > a SS10/20 and U2 can take more than 1 CPU, even
> > a dual CPU U60 goes for under 1000pounds now.
>
> So? How does that constitute an 8 CPU system? At most its only 2.
>
Up to 8 CPUs means you can use it on systems with up to 8 CPUs,
means you can installed it on a dual CPU or quad CPU system, as the
bew license only allows you to install Solaris on a single CPU machine
without paying extra.

> > $1....WOW! that's a lot of money, you can buy a botte of mineral water.
>
> What's a botte?
>
bottle, typo.

> > > > One can't expect Microsoft to give Windoze away for free, as
Microsoft
> > > > will never gain any extra revenue from hardware sales.
>
> Looks like you've just proven "Kirbu" wrong.
>
> > Microsoft = software company = sell softwares for $$$.
> > Sun, HP, IBM, SGI = hardware companies who sell their own softwares =
> > selling hardwares for living, not softwares.
>
> Your point being?
>
Are you wasting people's time?
So Sun, SGI, IBM, HP don't need to rely on software sales for their profit.

> > Irix works on MIPS, HP-UX runs on PA and Itanium, Solaris on Sparc with
> > limited support on x86.
> > Now who makes MIPS, PA and Sparc based machine?
> > If SGI do as MS, you would have possibily the prettiest OS without any
> > hardwares to run it on.
>
> Your point being? MS don't appear to have any problems with that.
> Why do you think SGI would?
>
Are we short of x86 vendors? are we short of manufacturers producing Intel
based
machines?
Of course not! Windows runs on x86, with plenty of x86 vendors out there,
they don't
need to worry about having no systems to run Windows on.
But Irix runs on MIPS, if SGI drops hardware production, and they are the
main
manufactorer for MIPS machines, what's the use for developing the OS when
there
is clearly no market for it.

> > > > One can't expect SCO to give away SCO UNIX, as again they will gain
no
> > > > extra revenue from hardware sales, although they might have done
with
> > > > software sales had Linux never existed.
> > >
> > > How do you conclude that?
> > >
> > The whole software and hardware company thingy.
>
> You haven't presented anything by way of a concrete argument
> yet. All I'm seeing is conjecture. How about you show me the
> money?
>
http://www.hp.com/hpinfo/investor/financials/annual/2002/graphics2002.pdf
starting from page 5.

> > > > > > >
> > > Again, how do you conclude that?
> > >
> > I am the only person in the entire department of EE.Eng with a copy of
SCO
> > UnixWare
>
> My, aren't you lucky. But so what?
>
Only supporting "But I don't think any home user or student would
care less whether or not SCO was given away now." statement using
the "figures" you asked for.
1 out of 400+ EE students, researchers and staffs use SCO UnixWare.
Without this 1, department will still be running, we will still produce
research works.
No harm done as it can be replaced by another flavour of OS, mainly Solaris.

> > Those are just my opinions
>
> They certainly are. And that's the trouble. Too much baseless
> opinion and no facts.
>
m.....interesting, aren't you all about your own baseless opnions and no
facts?
Don't see you saying anything constructive in this thread at all, only "so
what?",
"your point is?", "How do you conclude that?"


Yuan

Dave Uhring wrote:
> 
> On Sun, 27 Jul 2003 03:44:04 +0100, Dr. David Kirkby wrote:
> 
> > HP have a reasonably attractive package that allows non-commerical
> > users to gain the latest copy of the Tru64 operating system and the HP
> > C compiler for Tru64 for $100. Yet HP have no such policy with their
> > other UNIX operating system (HP-UX).
> 
> And the compiler?  As of V5.1b, the compiler has a -separate- license.
> 

I was not aware of that. I thought my $100 included a compiler. The
compiler software is on the CD and the licence keys provided enable
the compiler to work. 

> > I'd like to know HP's logic of the differing policies of Tru64 and
> > HP-UX. It seems very odd to sell Tru64 and compiler for $99, yet give
> > no discounts on the HP-UX.
> 
> HP-UX has a "future" if their idiotic devotion to Itanic does not kill the
> server branch of the company.  Tru64 has no future at all; if HPQ can get
> $99 for the 3 basic licenses it is a windfall.

I must admit I was unaware of these differences between Tru64 and
HP-UX. In that case, I think it makes more sense to encourage students
to use HP-UX.
 
> > One can't expect Microsoft to give Windoze away for free, as Microsoft
> > will never gain any extra revenue from hardware sales. But SGI, HP,
> > IBM, Sun etc. will gain extra sales of hardware, by allowing home
> > users to learn their operating systems, rather than that of their
> > competitors.
> 
> Microsoft's student discounts to almost all university students
> essentially amounts to giving away the "OS" and many apps for free.

I know universities get MS products at huge discounts. I can legally
use most MS products at home for no charge. However, our licence does
not allow students to take home MS products - only staff. I was not
aware students could buy MS products cheaply, but have never bothered
looking into that one. 
 
> > One can't expect SCO to give away SCO UNIX, as again they will gain no
> > extra revenue from hardware sales, although they might have done with
> > software sales had Linux never existed. But I don't think any home
> > user or student would care less whether or not SCO was given away now.
> 
> SCO is irrelevant and has been so for 10 years.  The company will not
> survive their legal action against IBM and every other user of Unix and
> Linux, Solaris excepted for good financial reasons.

I made that comment more as a joke. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Anthony Mandic wrote:
> 
> "Dr. David Kirkby" wrote:
> >
> > As someone who works in a University, I am amazed UNIX vendors don't
> > give away free copies of their operating systems to students and home
> > users.
> 
>         Why? Do Microsoft do this? Perhaps you should widen your
>         job experience.
>

Someone else has said they do. I know they encourage the use of MS
products at universities by giving huge discounts. 

 
> > If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> > their OS releases, home users would be more likely to buy cheap used
> > HP kit.
> 
>         How do you conclude that?

There are a large number of  

> > When those home users go in to industry, they are more likely
> > to buy **EXPENSIVE** new HP kit.
> 
>         What makes you think they would be in a position to do so?
>         Exactly how many home users go from there immediately to
>         management positions with buying power?
> 

How on earth do you expect me to know the *exact* number? I think
anyone who runs Solaris/AIX/HP-UX etc at home is likely to be
sufficiently interested in computers to take that career path. Or like
myself, have no professional involvement with computers, yet still
infulence what kit is bought. 

> > So it's to HP's advantage to give
> > away HP-UX to students and home users.
> 
>         I can't see how you can draw that conclusion.

I stated my logic for that, so I'm unwilling to repeat it. 

> > The same can be said for any UNIX hardware vendor - I strongly suspect
> > it's in their long term economic interest to encourage people to learn
> > their operating systems. Once you know one commercial UNIX operating
> > system, you are likely to want to stick to it.
> 
>         That sounds more like conjecture. What evidence do you have
>         to support that claim?

Becuase I know what products students are confortable with when they
leave univeristy and what one they don't know. 

> > Why on earth Sun don't change the licence conditions of their
> > so-called free licence for Solaris on SPARC, I don't know. Solaris can
> > be freely downloaded, yet its illegal to use it on virtually all used
> > Sun kit sold. Their previous licence, which allowed for its use on
> > machines of up to 8 CPUs seemed more sensible.
> 
>         Why? How many home users/students have 8 CPU systems?
 
Probably very few. 4 CPU systems are much more common. I don't see
nothing wrong with them changing their old licence conditions and
changing 8 to 4. But the free Solaris licence that prevented you using
it if the machine was capable of holding more than one cpu (not even
having more than one cpu), is silly. That excludes sparcstation 10's
and 20's, but of which are common amongst home users. 

> > One can't expect Microsoft to give Windoze away for free, as Microsoft
> > will never gain any extra revenue from hardware sales.
> 
>         But by your previous arguments, said home users of such
>         free software would go on to use it commercially.

Yes, okay, MS should do so. Apprently they do so anyway, although I
was unaware of it. I know they encourage its use in universities, by
making it so cheap. 

I suspect where I would would switch to StarOffice for students if MS
products cost a fortune and StarOffice did not. Unfortunatley, even if
Sun gave us StarOffice,  there would be no inclination to switch to
it, since Windoze and all the MS products cost us next to nothing.  

> > But SGI, HP, IBM, Sun etc. will gain extra sales of hardware, by allowing
> > home users to learn their operating systems, rather than that of their
> > competitors.
> 
>         You're argument seems specious to me. One could just as easily
>         conclude that it would be better to do as MS does and stick to
>         software and forget about hardware.
 
MS seeem to be doing pretty well at it.

> > One can't expect SCO to give away SCO UNIX, as again they will gain no
> > extra revenue from hardware sales, although they might have done with
> > software sales had Linux never existed.
> 
>         How do you conclude that?
 
I give up with you. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Yuan wrote:

> > > > Why? Do Microsoft do this?
> > > >
> > > 50 UK pounds for Windows 2000 student license,
> >
> > Which isn't exactly free.
> >
> Cheaper than most Unix OS.

	And how much for a Unix license for a student?

> > > My second copy of Windows 2000 was given to me for free by MS UK.
> >
> > That's sounds more like it. So they do do it.
> >
> So? it's a special case.

	Is it? Why didn't you state that when you first mentioned it then?

> > Ah! So that explains his troll post.
> >
> No, it's just a simple fact that second hand HP kits are not as popular as
> Sun kits, hence the seller can only demand less.

	No, its a troll post. "Kirbu" was doing his usual bitching and
	moaning because he can't get it for free.

> > > BUT, you can downloads Solaris ISO from Sun but not HP-UX.
> >
> > That's good. So what?
> >
> You don't like to excersise your brain cells do you?

	At least I make an effort to spell.

> So you can downloads the ISO, use your favourite CD buring software, and put
> a few blank CD-R into your CD-RW, burn from those ISO into bootable CDs as
> installation medias, and install Solaris from those CDs, it works out much
> cheaper than buying the media set from the vendors.

	Which has nothing what so ever to do with the fact that you
	can do it with Sun but not HP. Do try and learn how to reason
	logically.

> > > People get used to the hardwares and softwares they use,
> >
> > So? That doesn't explain how they go from being a home user/student
> > to being in a management position with purchasing power. Unless ...
> > whoever hired them was a complete and utter idiot.
> >
> So you were never a student and you were never a home user?

	Like most everyone else, I was both. What I use nowadays, I
	never used then. Hence I don't see how your (and "Kirbu"'s)
	statement is anything other than specious. If you can't back
	it up with some facts and figures it remains specious.

> > > I can't get over using Outlook Express as my default news client,
> >
> > My condolences.
> >
> My consolences to you too.

	Why? I don't have to concern myself with virii and the other
	plagues OE brings with it. If you're happy with it, good luck
	to you.

> > > > Exactly how many home users go from there immediately to
> > > > management positions with buying power?
> > > >
> > > Not a lot, but guess how MS become so dominent.
> >
> > I'm not here to guess. So don't patronise me with specious
> > statements. Show me facts and figures.
> >
> The same as your non-constructive "so?", "your point being?" comments.

	I'm not the one stating conjecture as fact. If you want to
	do that, you'd better be prepared to defend your stance when
	challenged.

> Students buy cheap MS products ---> students grow up only knowing
> how to use Windows.

	Do they buy such cheap products? Show us the figures. I gather
	you mean "knowing how to use" loosely. Try working at a helpdesk
	sometime where the users use MS day in and day out. After that
	much experience do they "know how to use" it?

> > Linux isn't HP. And quoting someone else's misinformed statement
> > doesn't make it any less misinformed. Show me some proof. How do
> > those who come to prefer HP do so when they don't use it as a
> > student or home user?
> >
> He only uses Linux as a fact to show you how it becomes popular.
> That's the proof you want.

	Try again. His use of it only shows it popular to him. Not to
	anyone else.

> > > lets suppose we all buy from a reseller,
> > > a SS10/20 and U2 can take more than 1 CPU, even
> > > a dual CPU U60 goes for under 1000pounds now.
> >
> > So? How does that constitute an 8 CPU system? At most its only 2.
> >
> Up to 8 CPUs means you can use it on systems with up to 8 CPUs,
> means you can installed it on a dual CPU or quad CPU system, as the
> bew license only allows you to install Solaris on a single CPU machine
> without paying extra.

	Which Sun has probably reasoned is adequate for the majority of
	circumstances. But the only way to know is to see the figures.

> > > Microsoft = software company = sell softwares for $$$.
> > > Sun, HP, IBM, SGI = hardware companies who sell their own softwares =
> > > selling hardwares for living, not softwares.
> >
> > Your point being?
> >
> Are you wasting people's time?

	I'm asking to see some facts. All I'm seeing is baseless conjecture,
	specious arguments and opinions. So who's wasting whose time?

> So Sun, SGI, IBM, HP don't need to rely on software sales for their profit.

	And MS and Linux don't need to rely on hardware sales for their
	profit. Circular argument. So where are we now? Back to square
	one yet?

> Are we short of x86 vendors? are we short of manufacturers producing Intel
> based machines?

	Do we need to be?

> Of course not! Windows runs on x86, with plenty of x86 vendors out there,
> they don't need to worry about having no systems to run Windows on.
> But Irix runs on MIPS, if SGI drops hardware production, and they are the
> main
> manufactorer for MIPS machines, what's the use for developing the OS when
> there is clearly no market for it.

	And you're missing the point. Anything will run on anything if
	ported. You don't need MIPS to have SGI, etc. etc. etc. If SGI's
	OS was worthwhile, it would get ported. No?

> > You haven't presented anything by way of a concrete argument
> > yet. All I'm seeing is conjecture. How about you show me the
> > money?
> >
> http://www.hp.com/hpinfo/investor/financials/annual/2002/graphics2002.pdf
> starting from page 5.

	I haven't got all day to wait for this trash to download.
	Cite the actual figures and then compare to MS's revenues
	for the same period.

> > > I am the only person in the entire department of EE.Eng with a copy of
> SCO
> > > UnixWare
> >
> > My, aren't you lucky. But so what?
> >
> Only supporting "But I don't think any home user or student would
> care less whether or not SCO was given away now." statement using
> the "figures" you asked for.

	Perhaps, perhaps not. We won't know unless they tried.

> 1 out of 400+ EE students, researchers and staffs use SCO UnixWare.
> Without this 1, department will still be running, we will still produce
> research works.

	And would they be able to do it irrespective of what other
	hardware and software they have/use?

> No harm done as it can be replaced by another flavour of OS, mainly Solaris.

	Then the choice doesn't really matter for you?

> > > Those are just my opinions
> >
> > They certainly are. And that's the trouble. Too much baseless
> > opinion and no facts.
> >
> m.....interesting, aren't you all about your own baseless opnions and no
> facts?

	No, I'll all about asking to see some facts for a change.
	I haven't really stated any opinions yet and I've not quoted
	any facts. I'd like to see someone convince me one way or the
	other.

> Don't see you saying anything constructive in this thread at all, only "so
> what?", "your point is?", "How do you conclude that?"

	That's my way of getting you to prove your arguments. If you
	can't back them up, they're not facts. And I'll question every
	point till its proved or disproved.

-am	� 2003

Uli Link wrote:

> The prices for the download versions of Solaris are very moderate.
> For developing apps you the x86 version is a great help.

Anyone trying to install Solaris x86 is likely to have a fair share of
problems. It did not support my Logitech mouse. It supports a small
subset of the hardware Linux does, making it hard to install on an
'average PC'. Not exactly the sort of impression you want to create. 
 
> > HP have a reasonably attractive package that allows non-commerical
> > users to gain the latest copy of the Tru64 operating system and the HP
> > C compiler for Tru64 for $100. Yet HP have no such policy with their
> > other UNIX operating system (HP-UX).
> >
> Use GCC instead. When you want/need portability using the same compiler on
> every platform is a great help.
> The quality/performance of the generated binaries differ over the platforms.

Yes but one can't even get HP-UX (the OS) cheaply, so without the OS,
you have no hope of running gcc.
 
> >
> > One can't expect SCO to give away SCO UNIX, as again they will gain no
> > extra revenue from hardware sales, although they might have done with
> > software sales had Linux never existed. But I don't think any home
> > user or student would care less whether or not SCO was given away now.
> >
> 
> But they did about 5 years ago. And as the gave away a license for free
> educational and non-comercial use the market share raised. And a word to all
> this flame wars vs. SCO
> UnixWare 7 is an excellent product, like the OpenServer5 was 10 years ago.

I was not aware of that. We never installed a copy. 

> I think the problems of *NIX hw vendors is Intel or AMD.
> In the early 90ties you could not buy the processing power on the x86
> platform, not even for really lots of money.

Agreed. PC's were hopeless then and there was no decent OS to run on
them.

> Today a 3GHz Xeon or Athlon blows away Sparc and PowerPC, 

Agreed. That is why my office at work has 2 Ultra 5's and an Ultra 1.
At one time there would have been 10 Suns in there. Now those 7 Suns
have been thrown away, or I've taken them home. 

> if you need more
> processing power you can buy 10 boxes down the road in the supermarket.

That is not quite true - depends on the problem. Writing code for
distrubuted comptuers is a lot harder than writing code for machines
like my Ultra 80 that where all 4 CPUs share the same memory. 

> And Linux and Free/Open/NetBSD has evolved to very solid and serious O/S
> with a little help from the GCC.

Yes, that is true. In the scientific community (which I know well),
Linux is the main OS people use. However, if you need very high
relieability, I don't think PCs are the best solution, although its
outside my area of expertese. I see Sun/HP/IBM etc will still maintain
a market - one they can increase by getting students using their OS. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

"Dr. David Kirkby" wrote:

> > > As someone who works in a University, I am amazed UNIX vendors don't
> > > give away free copies of their operating systems to students and home
> > > users.
> >
> >         Why? Do Microsoft do this? Perhaps you should widen your
> >         job experience.
> 
> Someone else has said they do. I know they encourage the use of MS
> products at universities by giving huge discounts.

	Yes, that's been established now and I'll accept its a fact.
	So we can establish its a common practice from time to time
	from vendor to vendor.

> > > If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> > > their OS releases, home users would be more likely to buy cheap used
> > > HP kit.
> >
> >         How do you conclude that?
> 
> There are a large number of

	Yes?

> > > When those home users go in to industry, they are more likely
> > > to buy **EXPENSIVE** new HP kit.
> >
> >         What makes you think they would be in a position to do so?
> >         Exactly how many home users go from there immediately to
> >         management positions with buying power?
> 
> How on earth do you expect me to know the *exact* number? I think
> anyone who runs Solaris/AIX/HP-UX etc at home is likely to be
> sufficiently interested in computers to take that career path. Or like
> myself, have no professional involvement with computers, yet still
> infulence what kit is bought.

	Because you stated this as fact. I'd like to see what evidence
	there is for it. Its quite possible that they go into an
	environment that already uses the same infrastructure. Or
	they buy into that environment at home so as to improve their
	job prospects. Any argument could be valid. They could all
	well be valid too.

> > > So it's to HP's advantage to give
> > > away HP-UX to students and home users.
> >
> >         I can't see how you can draw that conclusion.
> 
> I stated my logic for that, so I'm unwilling to repeat it.

	Yes, but as a scientist, you should know the value of facts
	and the importance of drawing the correct conclusion from them.
	Given that HP don't do this now, how do you conclude that they
	manage to stay in business? I would conclude, from the stated
	claims, that they can't. But they do, so we have a contradiction.

> > > The same can be said for any UNIX hardware vendor - I strongly suspect
> > > it's in their long term economic interest to encourage people to learn
> > > their operating systems. Once you know one commercial UNIX operating
> > > system, you are likely to want to stick to it.
> >
> >         That sounds more like conjecture. What evidence do you have
> >         to support that claim?
> 
> Becuase I know what products students are confortable with when they
> leave univeristy and what one they don't know.

	Don't you mean for those students that you know?

> >         Why? How many home users/students have 8 CPU systems?
> 
> Probably very few. 4 CPU systems are much more common.

	And 2 and 1 CPU systems even more so?

> I don't see nothing wrong with them changing their old licence conditions
> and changing 8 to 4. But the free Solaris licence that prevented you using
> it if the machine was capable of holding more than one cpu (not even
> having more than one cpu), is silly. That excludes sparcstation 10's
> and 20's, but of which are common amongst home users.

	But it must have been done for a reason. We don't know what
	that reason is so we can only speculate. However, there's
	nothing preventing anyone staying with Solaris 8 if they
	have more than one CPU.

> > > One can't expect Microsoft to give Windoze away for free, as Microsoft
> > > will never gain any extra revenue from hardware sales.
> >
> >         But by your previous arguments, said home users of such
> >         free software would go on to use it commercially.
> 
> Yes, okay, MS should do so. Apprently they do so anyway, although I
> was unaware of it. I know they encourage its use in universities, by
> making it so cheap.

	Yes, that's been established now. However, how prevalent is it
	and how does it compare to its widespread use? Perhaps the
	figures will suggest that giving it away for free isn't that
	great an idea.

> I suspect where I would would switch to StarOffice for students if MS
> products cost a fortune and StarOffice did not. Unfortunatley, even if
> Sun gave us StarOffice,  there would be no inclination to switch to
> it, since Windoze and all the MS products cost us next to nothing.

	The only issue here is that MS Office only runs on MS OSs. StarOffice
	is more OS/platform independent isn't it?

> > > But SGI, HP, IBM, Sun etc. will gain extra sales of hardware, by allowing
> > > home users to learn their operating systems, rather than that of their
> > > competitors.
> >
> >         You're argument seems specious to me. One could just as easily
> >         conclude that it would be better to do as MS does and stick to
> >         software and forget about hardware.
> 
> MS seeem to be doing pretty well at it.

	Yes, they seem to have captured a large slice of the OS pie
	without any hardware base. So what does that tell us really?
	Is hardware important?

> > > One can't expect SCO to give away SCO UNIX, as again they will gain no
> > > extra revenue from hardware sales, although they might have done with
> > > software sales had Linux never existed.
> >
> >         How do you conclude that?
> 
> I give up with you.

	Already? And I was just getting started.

-am	� 2003

Yuan wrote:

> Since Dr. Kirkby told me how fast his HP C3600 is compare to his
> quad 450MHz U80, I have been trying to find a HP machine too.
> For 1000US, you can find an HP C3600 from eBay with faster
> processor than a single 450MHz U60/80 which tend to have higher
> price tag.
> BUT, you can downloads Solaris ISO from Sun but not HP-UX.
> 
> Yuan

I will correct that one. It is a 400 MHz HP C3000 using gcc that is
considerably quicker than a 450 MHz CPU in my Ultra 80 using Sun's
compiler. Overall I'm very impressed with the HP. Very well built,
performs very well, looks nice, ... shame it can't take a few more
CPUs. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

[Yuan]:
>
>   50 UK pounds for Windows 2000 student license, compare to what I
>   paid for HP-UX 11i for 100pounds from eBay + shipping, Irix 6.5
>   base with 6.5.2 and 6.5.13 update for 75pounds + shipping, both
>   WITHOUT license, media kits only.  My second copy of Windows 2000
>   was given to me for free by MS UK.

here's what our students at the Department of Informatics can get for
free:

  1. Windows XP pro
  2. Windows 2000 pro
  3. Windows 2000 Server, Windows 2000 Advanced Server
  4. Windows Server 2003 Web Edition
  5. Windows Server 2003 Std. Ed., Windows Server 2003 Ent. Ed.
  6. MS Visio Professional 2002
  7. MS Project Professional 2002
  8. Visual J# .NET
  9. Visual Studio .NET 2002 (CD5 = ComponentUpdate) (5 CD-er)
 10. Visual Studio .NET 2003 (CD3 = Prerequisites) (3 CD-er)
 11. Visual Studio .NET 2003 Student Tools
 12. MSDN Library .NET 2002 (April 2003) (3 CD-er)
 13. MSDN Library .NET 2003 (for VS.NET 2003) (3 CD-er)

only one licence of each, and they must supply the CD-Rs themselves ;-)

notice that MS Office is not included -- they want to entice system
developers, not secretaries.
-- 
Kjetil T.			|  read and make up your own mind
				|  http://www.cactus48.com/truth.html

Akop Pogosian <akopps+usenet@ocf.berkeley.edu> writes:

> Scott Howard <scott@hunterlink.net.au> wrote:
> > In comp.unix.solaris Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> >> As someone who works in a University, I am amazed UNIX vendors don't
> >> give away free copies of their operating systems to students and home
> >> users. If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> 
> > Sun do.  Sure, it's (legally) limited to one CPU, but that's not generally
> > going to be a big restriction.
> 
> It's legally limited to one CPU _AND_ , in case of SPARC systems, the
> machine must have been sold by either Sun or an authorized
> Sun-reseller as stated at the bottom of this page:
> http://wwws.sun.com/software/solaris/licensing/policies.html
> 
> Moreover, this page states that the right to use Solaris is not
> transferable when the machine is resold with the exceptions of a few
> authorized vendors. This means that nearly all used Sparc systems sold
> on say on ebay today do not come with a valid license to use Solaris
> and you legally CANNOT use the software licensed under the Free Binary
> Solaris License program on those machines. You have to pay for the
> Solaris Upgrade RTU license on such machines, even for a system with
> one CPU, regardless of what you're using it for. The RTU license for a
> single-proc system costs about $100. This is why the Solaris/sparc
> enthusiast community is outraged. In many cases, the price of this
> upgrade RTU license costs a large fraction of or even more than the
> price of many used Sun systems that are resold today. I think it's
> fair to slap this cost on the businesses and other organizations who
> use this software to run their day to day tasks but it's unfair
> towards the hobbyists and enthusiasts who don't derive a commercial
> gain from using those systems and who actively promote the usage of
> Sun hardware and software outside of their homes.
> 


I think "outrage" may overstate it a bit.  I've bought an Ultra 2 and
two 5's, and also bought the $50 "slim media" to go with them- and
used the free binary licence thing to download the cd's and start
installing while the official disks are en-route.  I suspect Sun only
really cares if you start setting up big infrastructure with 2nd-hand
hardware and attempt to get the OS for cheap.

I agree it is sometimes a bit silly- why should the OS have to be
purchased multiple times for a old unit.  OTOH, if the ancient box had
a correspondingly ancient Solaris on it, why should the old licence
let you put Solaris 9 on it for free?  Further, if you're the new
owner of the box and you didn't get the install media, then the prior
owner still has the licence (assuming he/she themselves had it
legally)- so you still need to buy a copy for yourself.

Personally, I'm glad they do offer a cheap media kit so small-time
users can be plausibly legal.  I really can't begrudge them $50/per
for a nice OS that still runs great on ~8 year old hardware.

Gregm

> > The prices for the download versions of Solaris are very moderate.
> > For developing apps you the x86 version is a great help.
>
> Anyone trying to install Solaris x86 is likely to have a fair share of
> problems. It did not support my Logitech mouse.

When you buy a *NIX workstation you will buy most likely a keyboard and
mouse with it.
Even the monitor cables are propietary even if all have that damn 13w3
connectors :-(
I think buying a second hand MS-mouse or better a old Logitech Pilot-3button
mouse is not a problem.

> It supports a small
> subset of the hardware Linux does, making it hard to install on an
> 'average PC'. Not exactly the sort of impression you want to create.

That's true, but it is absolutly no problem getting the components from
Sun's HCL.


> Yes but one can't even get HP-UX (the OS) cheaply, so without the OS,
> you have no hope of running gcc.
>
That's a real disadvantage :-((

> >
> > But they did about 5 years ago. And as the gave away a license for free
> > educational and non-comercial use the market share raised. And a word to
all
> > this flame wars vs. SCO
> > UnixWare 7 is an excellent product, like the OpenServer5 was 10 years
ago.
>
> I was not aware of that. We never installed a copy.

I was lucky enough to get a few personal licenses for UnixWare2, 7 and OSR5.
And the media kits were sold for about 50$ those times.

> > Today a 3GHz Xeon or Athlon blows away Sparc and PowerPC,
>
> Agreed. That is why my office at work has 2 Ultra 5's and an Ultra 1.
> At one time there would have been 10 Suns in there. Now those 7 Suns
> have been thrown away, or I've taken them home.
>
> > if you need more
> > processing power you can buy 10 boxes down the road in the supermarket.
>
> That is not quite true - depends on the problem. Writing code for
> distrubuted comptuers is a lot harder than writing code for machines
> like my Ultra 80 that where all 4 CPUs share the same memory.

Look for a Compaq Proliant, many of them takes up to four PentiumIII.
The problem of writing distributable code is the same for a x86 SMP machine
or a Sparc/PowerPC/Mips architecture. Bringing it to a cluster is a new
dimension of complexity. So far I agree.

>
> > And Linux and Free/Open/NetBSD has evolved to very solid and serious O/S
> > with a little help from the GCC.
And x86 CPUs with enough floatingpoint power.
A 486 was a good CPU for database with lots of integer and I/O those days.

>
> Yes, that is true. In the scientific community (which I know well),
> Linux is the main OS people use. However, if you need very high
> relieability, I don't think PCs are the best solution, although its
> outside my area of expertese. I see Sun/HP/IBM etc will still maintain
> a market - one they can increase by getting students using their OS.

You can buy x86 built for professional use, built like tanks.
Compaq, IBM, Dell. Reliable as the Risc counterparts.


---
Uli

In article <3F23C384.51499CBB@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:

>> If you buy a used Sun, this machine would have been run wich a license before.
>> Even if the seller used a corporate license, the machine definitely did come
>> with a valid license from Sun when bought from Sun.
>> 
>> At least this license could be sold together with the HW. 
>
>Agreed. That is not the case.
>
>> If Sun states differenty, this is illegal and thus void.
>
>I'm no laywer, but I don't think it is. 

IBM did try to do the same before and did lost a case on court.

-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

On Sun, 27 Jul 2003 13:53:53 +0000, Joerg Schilling wrote:

> In article <3F23C384.51499CBB@ntlworld.com>,
> Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> 
>>> If you buy a used Sun, this machine would have been run wich a license before.
>>> Even if the seller used a corporate license, the machine definitely did come
>>> with a valid license from Sun when bought from Sun.
>>> 
>>> At least this license could be sold together with the HW. 
>>
>>Agreed. That is not the case.
>>
>>> If Sun states differenty, this is illegal and thus void.
>>
>>I'm no laywer, but I don't think it is. 
> 
> IBM did try to do the same before and did lost a case on court.

In a -German- court?

IIRC, Microsoft also attempted to void the Windows license on transfer of
ownership of the machine and a German court struck down that restriction.
Unfortunately, those courts' jurisdictions do not extend to the rest of
the world WRT such license restrictions.

In comp.sys.dec Anthony Mandic <o3@hotmail.com> wrote:
>> >
>> You don't like to excersise your brain cells do you?

> 	At least I make an effort to spell.

awai: but you make no effort to think.

(I prefer the former).

-- 
Thomas E. Dickey <dickey@radix.net> <dickey@herndon4.his.com>
http://dickey.his.com
ftp://dickey.his.com

On Sat, 26 Jul 2003, Dave Uhring wrote:

> On Sun, 27 Jul 2003 03:44:04 +0100, Dr. David Kirkby wrote:
> 
> > HP have a reasonably attractive package that allows non-commerical
> > users to gain the latest copy of the Tru64 operating system and the HP
> > C compiler for Tru64 for $100. Yet HP have no such policy with their
> > other UNIX operating system (HP-UX). 
> 
> And the compiler?  As of V5.1b, the compiler has a -separate- license.

I don't believe that this is true. I have several 5.1B systems, and there 
is no separate license for the C compiler.

steve

On Sun, 27 Jul 2003 13:53:53 +0000, Joerg Schilling wrote:

> In article <3F23C384.51499CBB@ntlworld.com>,
> Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> 
>>> If you buy a used Sun, this machine would have been run wich a license before.
>>> Even if the seller used a corporate license, the machine definitely did come
>>> with a valid license from Sun when bought from Sun.
>>> 
>>> At least this license could be sold together with the HW. 
>>
>>Agreed. That is not the case.
>>
>>> If Sun states differenty, this is illegal and thus void.
>>
>>I'm no laywer, but I don't think it is. 
> 
> IBM did try to do the same before and did lost a case on court.

In a -German- court?

IIRC, Microsoft also attempted to void the Windows license on transfer of
ownership of the machine and a German court struck down that restriction.
Unfortunately, those courts' jurisdictions do not extend to the rest of
the world WRT such license restrictions.

In news:3F23C093.548E771C@hotmail.com, Anthony Mandic wrote:

> And how much for a Unix license for a student?

Comming from the SGI group ... there are no student licenses for IRIX.
One can purchase a media set and license for a workstation (IRIX 6.5
AWE) for about $600.  Not good for those people buying up all these
systems on eBay for cheap.

> If SGI's OS was worthwhile, it would get ported. No?

Too costly.  Too many licensed bits.  Too easy to adapt Linux
apparently.

Chris

-- 
#include <disclaimer.h>
cm007i@hotmail.com

On Sun, 27 Jul 2003 19:17:00 +0200, Thomas Dehn wrote:

> The German court rulings go way beyond that.
> M$ have lost so many lawsuits about licenses
> in Germany that they will
> be very reluctant to file another one.

The US laws and courts should be that enlightened.
 
> In Germany, if you have a license for a software,
> you can do with your copy whatever you want, just as if
> you have bought a copy of a book. Make copies for
> yourself, give the CDs to your friends, resell it, whatever.
> The vendor cannot impose any restrictions at
> all after the user has purchased the software,
> especially, license agreements which pop up
> during installation (i.e., when the software has already been
> bought) are completely void.

That freedom does not seem to have degraded their profits or continuing
desire to sell their crap in Germany.  They even undercut SuSE for the
Munich contract and still lost it :))

> (M$ are now trying to prevent that by adding
> technical obstacles such as "recovery CDs")

Those 'recovery CDs' are what MSFT has insisted that OEMs provide rather
than generic installation media.  They have been doing that for quite some
time, too.  Over a year ago I re-installed Win98 on an HP Pavillion
machine for a friend and I had to use the HP supplied CD set.

Hi Doctor,

I think you must try to work in a real company. First, you will win enough
money to buy a HP.
Second, you will understand finally that what's is expensive is to pay
employees not hardware or software even from HP (Do you know the 'tr�s � la
mode' word "consolidation" (=consolidating benefits with less sys admin,
oracle admin,...))
About Itanium, HP is developping for Intel 1000SX chipset which is the glue
to put together many Itanium CPUs. But HP will not leave developpement of
PA-RISC line, just in case of delay to put on the market Itanium and also to
satisfy existing customers.

Alain.

PS: sorry for my english.

"Anthony Mandic" <o3@hotmail.com> wrote in message
news:3F23C093.548E771C@hotmail.com...
> Yuan wrote:
>
> > > > > Why? Do Microsoft do this?
> > > > >
> > > > 50 UK pounds for Windows 2000 student license,
> > >
> > > Which isn't exactly free.
> > >
> > Cheaper than most Unix OS.
>
> And how much for a Unix license for a student?
>
HP-UX, Irix and Solaris have no student licensing program.

> > > > My second copy of Windows 2000 was given to me for free by MS UK.
> > >
> > > That's sounds more like it. So they do do it.
> > >
> > So? it's a special case.
>
> Is it? Why didn't you state that when you first mentioned it then?
>
Because it happens, not just to me, but to the other students as well,
but it's not common for MS to give their softwares away for free.

> > > Ah! So that explains his troll post.
> > >
> > No, it's just a simple fact that second hand HP kits are not as popular
as
> > Sun kits, hence the seller can only demand less.
>
> No, its a troll post. "Kirbu" was doing his usual bitching and
> moaning because he can't get it for free.
>
> > > > BUT, you can downloads Solaris ISO from Sun but not HP-UX.
> > >
> > > That's good. So what?
> > >
> > You don't like to excersise your brain cells do you?
>
> At least I make an effort to spell.
>
?,?????.
It just happens English is not my first language, and I guess a perfect
person like you
never experience something called typo.

> > So you can downloads the ISO, use your favourite CD buring software, and
put
> > a few blank CD-R into your CD-RW, burn from those ISO into bootable CDs
as
> > installation medias, and install Solaris from those CDs, it works out
much
> > cheaper than buying the media set from the vendors.
>
> Which has nothing what so ever to do with the fact that you
> can do it with Sun but not HP. Do try and learn how to reason
> logically.
>
Really, can't see how you can make a valid point here.
Using your own logics to reply on your comment
"I have done enough about logic gates so I don't need to learn any more
on how to reason logically"

> > > > People get used to the hardwares and softwares they use,
> > >
> > > So? That doesn't explain how they go from being a home user/student
> > > to being in a management position with purchasing power. Unless ...
> > > whoever hired them was a complete and utter idiot.
> > >
> > So you were never a student and you were never a home user?
>
> Like most everyone else, I was both. What I use nowadays, I
> never used then. Hence I don't see how your (and "Kirbu"'s)
> statement is anything other than specious. If you can't back
> it up with some facts and figures it remains specious.
>
Again, my condolences to you, for not having the freedom and the early
opportunity to use the softwares you enjoy using.

> > > > I can't get over using Outlook Express as my default news client,
> > >
> > > My condolences.
> > >
> > My consolences to you too.
>
> Why? I don't have to concern myself with virii and the other
> plagues OE brings with it. If you're happy with it, good luck
> to you.
>
Nope, it's for your narrow minded behaviour not being able to accept and
respect other people's preferance, and if the "plagues" can infect in ASCII
form, I should be excited and can't wate to take it into the lab to evaluate
it.

> > > > > Exactly how many home users go from there immediately to
> > > > > management positions with buying power?
> > > > >
> > > > Not a lot, but guess how MS become so dominent.
> > >
> > > I'm not here to guess. So don't patronise me with specious
> > > statements. Show me facts and figures.
> > >
> > The same as your non-constructive "so?", "your point being?" comments.
>
> I'm not the one stating conjecture as fact. If you want to
> do that, you'd better be prepared to defend your stance when
> challenged.
>
If you really put in some time and energy reading and digesting the post,
then
I will, but you are not accepting and absorbing any points we tried to make,
why bother?

> > Students buy cheap MS products ---> students grow up only knowing
> > how to use Windows.
>
> Do they buy such cheap products? Show us the figures. I gather
> you mean "knowing how to use" loosely. Try working at a helpdesk
> sometime where the users use MS day in and day out. After that
> much experience do they "know how to use" it?
>
m...guess you know all about student life than an actual student.
Yes, we love to buy expensive things, we love paying full price for
everything.

> > > Linux isn't HP. And quoting someone else's misinformed statement
> > > doesn't make it any less misinformed. Show me some proof. How do
> > > those who come to prefer HP do so when they don't use it as a
> > > student or home user?
> > >
> > He only uses Linux as a fact to show you how it becomes popular.
> > That's the proof you want.
>
> Try again. His use of it only shows it popular to him. Not to
> anyone else.
>
As much as I don't like the Linux hype, I still have to admit it is a
poprlar OS.
At least I can't find any Unix magazine, there isn't a magazine I know just
for Unix
but quiet a few for Linux.

> > > > lets suppose we all buy from a reseller,
> > > > a SS10/20 and U2 can take more than 1 CPU, even
> > > > a dual CPU U60 goes for under 1000pounds now.
> > >
> > > So? How does that constitute an 8 CPU system? At most its only 2.
> > >
> > Up to 8 CPUs means you can use it on systems with up to 8 CPUs,
> > means you can installed it on a dual CPU or quad CPU system, as the
> > bew license only allows you to install Solaris on a single CPU machine
> > without paying extra.
>
> Which Sun has probably reasoned is adequate for the majority of
> circumstances. But the only way to know is to see the figures.
>
Simple, spend 3 minutes per day for the next 3 months of your life,
collecting
information about how many Sun machines listed on eBay has SMP capability.

> > > > Microsoft = software company = sell softwares for $$$.
> > > > Sun, HP, IBM, SGI = hardware companies who sell their own softwares
=
> > > > selling hardwares for living, not softwares.
> > >
> > > Your point being?
> > >
> > Are you wasting people's time?
>
> I'm asking to see some facts. All I'm seeing is baseless conjecture,
> specious arguments and opinions. So who's wasting whose time?
>
Once again, my point on you not liking to use your brain cells stands,
if you just spend a few seconds to digest the statement, it won't be hard
for
you to figure "the part you ommited" out.

> > So Sun, SGI, IBM, HP don't need to rely on software sales for their
profit.
>
> And MS and Linux don't need to rely on hardware sales for their
> profit. Circular argument. So where are we now? Back to square
> one yet?
>
It's a cycle for MS and Linux with hardwares vendors.
The more PC they sell, the more people use Linux and Windows.
With SGI, they form their own cycle, if they don't sell MIPS workstation and
just
concentrate on the dev of Irix, how is this going to work out?

> > Are we short of x86 vendors? are we short of manufacturers producing
Intel
> > based machines?
>
> Do we need to be?
>
You tell me.

> > Of course not! Windows runs on x86, with plenty of x86 vendors out
there,
> > they don't need to worry about having no systems to run Windows on.
> > But Irix runs on MIPS, if SGI drops hardware production, and they are
the
> > main
> > manufactorer for MIPS machines, what's the use for developing the OS
when
> > there is clearly no market for it.
>
> And you're missing the point. Anything will run on anything if
> ported. You don't need MIPS to have SGI, etc. etc. etc. If SGI's
> OS was worthwhile, it would get ported. No?
>
Yes, it will get ported, but will it run as well as on the original
hardwares?
Can Solaris on a x86 runs as well as on a Sparc machine?

> > > You haven't presented anything by way of a concrete argument
> > > yet. All I'm seeing is conjecture. How about you show me the
> > > money?
> > >
> >
http://www.hp.com/hpinfo/investor/financials/annual/2002/graphics2002.pdf
> > starting from page 5.
>
> I haven't got all day to wait for this trash to download.
> Cite the actual figures and then compare to MS's revenues
> for the same period.
>
As I said, wasting people's time! you asked for the figure, and I show you
the link
to the figure showing in HP's press release, they stated hardwares as their
"KEY PRODUCT"

> > > > I am the only person in the entire department of EE.Eng with a copy
of
> > SCO
> > > > UnixWare
> > >
> > > My, aren't you lucky. But so what?
> > >
> > Only supporting "But I don't think any home user or student would
> > care less whether or not SCO was given away now." statement using
> > the "figures" you asked for.
>
>>

> > 1 out of 400+ EE students, researchers and staffs use SCO UnixWare.
> > Without this 1, department will still be running, we will still produce
> > research works.
>
> And would they be able to do it irrespective of what other
> hardware and software they have/use?
>
 Perhaps, perhaps not. We won't know unless they tried.


> > No harm done as it can be replaced by another flavour of OS, mainly
Solaris.
>
> Then the choice doesn't really matter for you?
>
It does, as I said before, people get used to things, if you ask me to stop
using Solaris
and start using Linux or HP-UX to do everything I do on Solaris, it would
takes me
several weeks just to get used to it.

> > > > Those are just my opinions
> > >
> > > They certainly are. And that's the trouble. Too much baseless
> > > opinion and no facts.
> > >
> > m.....interesting, aren't you all about your own baseless opnions and no
> > facts?
>
> No, I'll all about asking to see some facts for a change.
> I haven't really stated any opinions yet and I've not quoted
> any facts. I'd like to see someone convince me one way or the
> other.
>
Maybe you should try researching it yourself.

> > Don't see you saying anything constructive in this thread at all, only
"so
> > what?", "your point is?", "How do you conclude that?"
>
> That's my way of getting you to prove your arguments. If you
> can't back them up, they're not facts. And I'll question every
> point till its proved or disproved.
>
One thing I learn in school, always do your own research, if you have any
doubt, look it up, you are the only one who can help yourself, those who
are not bothered to are called slackers.

Yuan

Volker Borchert wrote:

> Dual or Quad processor Sparc-10/20 systems are not _that_ uncommon.

I must disagree with that. It costs next to nothing to buy a dual
processor SPARC 20. I bought one with 320 Mb of RAM for �60 (~ $100) 6
months back. I doubt you could buy a dual processor used PC for �60.

I got a quad processor SS20 (4x125 MHz) from work for nothing.  

I also have a quad processor U80 at home, which I admit is a bit
rarer, but now  I could probably buy that U80 from eBay for the cost
of a reasonably decent PC. 

More than 4 CPUs is quite unusual for a home computer. There will
always be the exception (someone here as mentioned he has 6 CPUs), but
I suspect (guessing) the number of home machines with more than 4 CPUs
is less than a few percent of those with 2 or 4 CPUs. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

"Kjetil Torgrim Homme" <kjetilho@yksi.ifi.uio.no> wrote in message
news:1rbrvg876j.fsf@glesvat.ifi.uio.no...
> [Yuan]:
> >
> >   50 UK pounds for Windows 2000 student license, compare to what I
> >   paid for HP-UX 11i for 100pounds from eBay + shipping, Irix 6.5
> >   base with 6.5.2 and 6.5.13 update for 75pounds + shipping, both
> >   WITHOUT license, media kits only.  My second copy of Windows 2000
> >   was given to me for free by MS UK.
>
> here's what our students at the Department of Informatics can get for
> free:
>
>   1. Windows XP pro
99pounds in my school

>   2. Windows 2000 pro
50pounds for it, and so is NT4.0

>   3. Windows 2000 Server, Windows 2000 Advanced Server
>   4. Windows Server 2003 Web Edition
>   5. Windows Server 2003 Std. Ed., Windows Server 2003 Ent. Ed.
MS UK doesn't even offer those to students at discount price.

>   6. MS Visio Professional 2002
>   7. MS Project Professional 2002
>   8. Visual J# .NET
>   9. Visual Studio .NET 2002 (CD5 = ComponentUpdate) (5 CD-er)
>  10. Visual Studio .NET 2003 (CD3 = Prerequisites) (3 CD-er)
>  11. Visual Studio .NET 2003 Student Tools
>  12. MSDN Library .NET 2002 (April 2003) (3 CD-er)
>  13. MSDN Library .NET 2003 (for VS.NET 2003) (3 CD-er)
>
Doh!! We have to pay fot those too! pretty expensive too.

> only one licence of each, and they must supply the CD-Rs themselves ;-)
>
> notice that MS Office is not included -- they want to entice system
> developers, not secretaries.
>
MS Office 2000 = 99pounds, department bought 5 copies for the secretaries.
Not sure about Office XP

Yuan

Yuan wrote:
> "Kjetil Torgrim Homme" <kjetilho@yksi.ifi.uio.no> wrote in message
> news:1rbrvg876j.fsf@glesvat.ifi.uio.no...
> 
>>[Yuan]:
>>
>>>  50 UK pounds for Windows 2000 student license, compare to what I
>>>  paid for HP-UX 11i for 100pounds from eBay + shipping, Irix 6.5
>>>  base with 6.5.2 and 6.5.13 update for 75pounds + shipping, both
>>>  WITHOUT license, media kits only.  My second copy of Windows 2000
>>>  was given to me for free by MS UK.
>>
>>here's what our students at the Department of Informatics can get for
>>free:
>>
>>  1. Windows XP pro
> 
> 99pounds in my school
> 
> 
>>  2. Windows 2000 pro
> 
> 50pounds for it, and so is NT4.0
> 
> 
>>  3. Windows 2000 Server, Windows 2000 Advanced Server
>>  4. Windows Server 2003 Web Edition
>>  5. Windows Server 2003 Std. Ed., Windows Server 2003 Ent. Ed.
> 
> MS UK doesn't even offer those to students at discount price.
> 
> 
>>  6. MS Visio Professional 2002
>>  7. MS Project Professional 2002
>>  8. Visual J# .NET
>>  9. Visual Studio .NET 2002 (CD5 = ComponentUpdate) (5 CD-er)
>> 10. Visual Studio .NET 2003 (CD3 = Prerequisites) (3 CD-er)
>> 11. Visual Studio .NET 2003 Student Tools
>> 12. MSDN Library .NET 2002 (April 2003) (3 CD-er)
>> 13. MSDN Library .NET 2003 (for VS.NET 2003) (3 CD-er)
>>
> 
> Doh!! We have to pay fot those too! pretty expensive too.
> 
> 
>>only one licence of each, and they must supply the CD-Rs themselves ;-)
>>
>>notice that MS Office is not included -- they want to entice system
>>developers, not secretaries.
>>
> 
> MS Office 2000 = 99pounds, department bought 5 copies for the secretaries.
> Not sure about Office XP
> 
> Yuan
> 
> 
HEY GUYS! This is the comp.sys.hp.hpux newsgroup, not the 
debate.about.licensing.and.vendors.behaviour newsgroup. Last time I 
looked we usually discuss and resolve our technical issues here. Please 
go it an IRC or use IM and discuss it. A lot of vendors give their stuff 
away, thats how they get people to use it. If you are working with HP-UX 
then you one way or another have to buy it, usually... Buy a used server 
and the license is included, all you need is media. If it is a 
workstation then make sure it has the OS loaded and then download Ignite 
from HP then you can make a copy of your disks. I don't know about the 
UK in particular, but workstations and servers are not that expensive 
used, and if you are involved in admin work on them it's nice to have a 
sandbox to play with. Sorry but I had to respond when I logged on to the 
newsgroup and saw dozens of replies on this thread.

Anthony Mandic wrote:
> Doug Dingus wrote:
>> Agree here, but you forget Linux.
> 
> Yes, I agree, you can forget Linux.

No. Linux is a Unix-like OS, whether you want to admit it or not. If I was employing a 
student or junior staff member, I would prefer one with Linux experience to one with M$ 
experience. Actually, I'd prefer one with no experience to one with M$ experience.

In fact, IMHO Debian is probably the best Linux distribution and most consistent in many 
ways with Solaris. Furthermore its patching and update programs are better than that of 
most of the commercial Unices. I have a general dislike for Red Hat and its derivatives, 
but that does not mean that they are inherently bad.

At least Linux has not given up the fight for the commercial desktop as the real Unix 
variants have. If you "forget Linux" you are, professionally and stupidly, shooting 
yourself in the foot before putting it firmly in your mouth.

>>Students that learn to use Open Source code will be able to do much
>>of their computing for free on just about any hardware they can find.
> 
> 	How do you figure that?

OK, let's run down the major Unix variants in no partcular order and see what tey run on 
(please correct me if there are any mistakes):

SunOS:	SPARC, SPARC64 and x86 (limited support) (Motorola 68k VME for SunOS 3)
Irix:	big-endian MIPS
HP-UX:	PA-RISC and x86 (IA-64 subset only)
Dynix:	x86 (only Sequent hardware)
AIX:	Power and PowerPC (s390?)
Mac OSX:
	PowerPC
Tru64 (incl Dec Unix, OSF/1, etc):
	Alpha and little-endian MIPS
AUX:	No longer available. We _can_ forget this one.
BSD (incl OpenBSD, NetBSD, FreeBSD):
	Alpha, HP300, PA-RISC, x86 & IA64, Motorola 68k (incl VME, HP300, Amiga & Atari),
	PowerPC (Apple, BeBox), SPARC, SPARC64, VAX, ARM & StrongARM, Amiga (m68k & ppc),
	Hitachi SuperH: SH3 (dreamcast!), SH4, SH5
Linux:	x86 (all from '386 on, incl IA-64, by far the best cover),
	SPARC, SPARC64, Power, PowerPC, Alpha, Motorola 68k,
	ARM, MIPS, PA-RISC, VAX and s390.
	Also, soon to appear: Hitachi SuperH.
	This still leaves all the handhelds which I'm not even going
	to bother with, but most of them now have at least a partial
	Linux port.

So what Doug is saying is that if a student has a computer, he will probably be able to 
run Linux on it. Only the *BSD family compare. But you probably don't consider these to be 
Unix either.

Or maybe your question referred to other considerations like tools, compilers, editors, 
hardware support, etc. I can assure you that all these tools exist for Linux (and *BSD) 
and probably in far greater quantity than for the regular Unixes. In fact, it has been my 
experience that many of the free tools have been of very high quality, better than tools I 
would have had to mortgage my house to buy.

Now that ends my contribution. I am now going to stop reading posts from Anthony Mandic as 
I find them a waste of time. I am a fan of Monty Python, but I have not paid for a 5 
minute argument...

-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GO/! d- s++:+ a+ C++(++++) US++++$ UB++ U*++ P+++ L+++ E--- W+++ N++ w--- O-
M+ V- PS+ PE+ Y+ PGP t+ 5++ X R* tv+ b+ DI++ D G e(*) h++/-- r+++ y?
------END GEEK CODE BLOCK------

-----------------------------------------------------
Bob Hoekstra: APL & Unix Consultant
Telephone:    +44 1483 771028
Mobile:       +44 7710 562345
Email:        Bob.Hoekstra@HoekstraSystems.ltd.uk
-----------------------------------------------------

On Sun, 27 Jul 2003 19:07:30 +0100, Dr. David Kirkby wrote:

> No sure (how do I find out ??), but the name rings a bell.  The 600
> MHz Alpha seems so slow (despite me now using the right compiler
> switches), that I don't tend to use it much.

The Alpha systems are not up right now.  IIRC, the command which displays
your licenses is 'lmf'.

"Dave Uhring" <daveuhring@yahoo.com> wrote:
> On Sun, 27 Jul 2003 19:17:00 +0200, Thomas Dehn wrote:
> 
> > The German court rulings go way beyond that.
> > M$ have lost so many lawsuits about licenses
> > in Germany that they will
> > be very reluctant to file another one.
> 
> The US laws and courts should be that enlightened.

The main problem with the US legal system seems
to be that any powerful corporation or group such as M$ 
can drown its opponents in legal costs, and thus will hardly
lose a lawsuit about a few thousand dollars against Joe Sixpack.
Drowing opponents in legal costs is impossible in Germany.


Thomas

Anthony Mandic <ny@hotmail.com> wrote in message 
(snip)
> 
> > Your best free UNIX is Linux.
> 
> 	Oh? And I suppose you can point us all to Linux's UNIX certification?

Who cares really.  (About the cert.)  Linux is not a UNIX.  It does
act like one though.  From a students point of view, Linux is a very
good OS to learn which is why I wrote what I did.  One could argue the
BSD family is better/worse than Linux.  I would likely agree on a
number of points there, but Linux has the mindshare right now along
with momentum.  Again, from a students point of view, that's a good
thing.

Your point is valid, if hasty.  Maybe I should have written:  "Your
best all-around free UNIX workalike is Linux."  Either way, the intent
is the same.

I did not say Linux was the best UNIX.  My personal commercial UNIX
preference is IRIX :)  I wonder how things will continue to go though.
 Linux has gotten quite good in a short amount of time.  Who knows, it
might end up the best yet.

 
> > Students that learn to use Open Source code will be able to do much
> > of their computing for free on just about any hardware they can find.
> 
> 	How do you figure that?

OSS runs on many different hardware types and vintages.  My use of the
word free is a little loose.  It does cost something to get the code
and it does take time to build / configure it on the target hardware. 
It does not take lots of hard cash and does not have license fees
attached along with their goofy EULA restrictions.  These things make
the majority of hardware useful long after its expected service life.

All you have to do is learn to compile code and understand how the OS
and hardware interact.  From there, many applications are for the
taking.  OSS is perfect for this because you can see down to the core
of what is happening if you want to.  --You can trust the machine to a
very high degree.

Getting things done becomes a matter of choices at this point.  Does
the hardware support the task or not?  Do I run the latest greatest
KDE / Gnome whatever, or do I run some older software instead?  Open
Standards allow communication across hardware and software versions to
a high degree.  These things combine to allow most basic (e-mail, web,
word processing, etc...) tasks to happen on most any machine.

Any student who takes the time to realize this, will be in control of
their computing experience.  They will understand they have choices
beyond those few they are presented and will be in a position to take
advantage of them if they choose.

Given the high costs of secondary education, students should be
learning just a bit more than how to use the latest commercial
offerings.  Educators are fools if they don't present them.  --That's
the other reason I wrote what I did.

"If you can learn how to compile software, you will be set for life." 
About 5 years ago, someone told me this when I was looking for
applications on some older SGI hardware.  I took a chance and started
learning.  Today, I find they were right.

In article <Pine.LNX.4.44.0307271122230.10371-100000@honker.vgersoft.com>,
Steve Thompson  <smt@vgersoft.com> wrote:
>On Sat, 26 Jul 2003, Dave Uhring wrote:
>
>> On Sun, 27 Jul 2003 03:44:04 +0100, Dr. David Kirkby wrote:
>> 
>> > HP have a reasonably attractive package that allows non-commerical
>> > users to gain the latest copy of the Tru64 operating system and the HP
>> > C compiler for Tru64 for $100. Yet HP have no such policy with their
>> > other UNIX operating system (HP-UX). 
>> 
>> And the compiler?  As of V5.1b, the compiler has a -separate- license.
>
>I don't believe that this is true. I have several 5.1B systems, and there 
>is no separate license for the C compiler.

Are you talking about the castrated to K&R compiler?



-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

On Sun, 27 Jul 2003 21:22:31 +0200, Thomas Dehn wrote:

> The main problem with the US legal system seems
> to be that any powerful corporation or group such as M$ 
> can drown its opponents in legal costs, and thus will hardly
> lose a lawsuit about a few thousand dollars against Joe Sixpack.
> Drowing opponents in legal costs is impossible in Germany.

And if a prosecutor with unlimited funds wins a favorable verdict then
politics prevents effective remedy of the original cause of action.

A. Lincoln's ideal of 'government of the people, by the people and for the
people' has taken a serious beating in the last 140 years.

Thomas Dehn wrote:
> "Dave Uhring" <daveuhring@yahoo.com> wrote:
> 
>>On Sun, 27 Jul 2003 13:53:53 +0000, Joerg Schilling wrote:
>>
>>
>>>In article <3F23C384.51499CBB@ntlworld.com>,
>>>Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
>>>
>>>
>>>>>If you buy a used Sun, this machine
>>>>>would have been run wich a license before.
>>>>>Even if the seller used a corporate license,
>>>>>the machine definitely did come
>>>>>with a valid license from Sun when bought from Sun.
>>>>>
>>>>>At least this license could be sold together with the HW.
> 
> 
> A license which probably is not for Solaris 8 or 9 on any
> of these old boxes which are available for a nickel
> and a dime on eBay.

Also, a lot of these licenses were university or corporate subscription 
or site licenses and have since expired. The university or corporation 
has no right to sell the software to you unless that's in the agreement 
or they reach some formal arrangement with Sun, because you're not one 
of their license members.

On Sun, 27 Jul 2003, it was written:

> Even the monitor cables are propietary even if all have that damn 13w3
> connectors :-(

The reason for those connectors is that they're much better
quality than the commodity crap that people use with the HD15
connectors.

Also, how are 13W3 connectors any more proprietory than HD15?

-- 
Rich Teer, SCNA, SCSA

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-online.net

Yuan wrote:
> As much as I don't like the Linux hype, I still have to admit it is a
> poprlar OS.
> At least I can't find any Unix magazine, there isn't a magazine I know just
> for Unix
> but quiet a few for Linux.

Try Sys Admin (http://www.sysadminmag.com/) and become a member of Sage 
(http://www.sage.org/) to receive ;login:

You are right, the rest have all disappeared. However, much of what you read in Linux mags 
is applicable to Unix. Pity that so much is at a very simple level :-(

>>And you're missing the point. Anything will run on anything if
>>ported. You don't need MIPS to have SGI, etc. etc. etc. If SGI's
>>OS was worthwhile, it would get ported. No?
>>
> 
> Yes, it will get ported, but will it run as well as on the original
> hardwares?
> Can Solaris on a x86 runs as well as on a Sparc machine?

Nobody has the source for Irix outside of SGI. Why should SGI port it? They have gone very 
heavily into Linux support for x86 machines, which they also produce. I suspect that in 
the long run Irix/MIPS will become an expensive niche market, eventually to die out. If 
SGI survive it will be as a shipper of cool, mainly x86, hardware and perhaps some 
high-end servers.


-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GO/! d- s++:+ a+ C++(++++) US++++$ UB++ U*++ P+++ L+++ E--- W+++ N++ w--- O-
M+ V- PS+ PE+ Y+ PGP t+ 5++ X R* tv+ b+ DI++ D G e(*) h++/-- r+++ y?
------END GEEK CODE BLOCK------

-----------------------------------------------------
Bob Hoekstra: APL & Unix Consultant
Telephone:    +44 1483 771028
Mobile:       +44 7710 562345
Email:        Bob.Hoekstra@HoekstraSystems.ltd.uk
-----------------------------------------------------

On Sun, 27 Jul 2003, Dave Uhring wrote:

> On Sun, 27 Jul 2003 19:17:00 +0200, Thomas Dehn wrote:
>
> > The German court rulings go way beyond that.
> > M$ have lost so many lawsuits about licenses
> > in Germany that they will
> > be very reluctant to file another one.
>
> The US laws and courts should be that enlightened.

<Rant>
I have no faith in a legal system that lets idiots sue companies
like McDonalds - AND WIN! - because they burned themselves by
spilling the hot coffee in their cup into their lap.  "But there
was no warning on the cup, saying the coffee is hot and I shouldn't
put it into my lap, therefore McDonalds is negligent, therefore I
want my big cash payment".

Give me a break.  If you buy a cup of coffee, you expect it
to be hot.  And if you're stupid enough to put a cup of hot
liquid between your legs while you're driving, then frankly
you get what you deserve.  If people took as much effort into
takeing responsibilty for their own actions as looking for a
way to sue someone, we'd all be better off.
</Rant>

Sorry about that; needless litigation is one of my pet peeves!

-- 
Rich Teer, SCNA, SCSA

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-online.net

Bob Hoekstra wrote:

> OK, let's run down the major Unix variants in no partcular order and see what tey run on
> (please correct me if there are any mistakes):
> 
> SunOS:  SPARC, SPARC64 and x86 (limited support) (Motorola 68k VME for SunOS 3)
> Irix:   big-endian MIPS
> HP-UX:  PA-RISC and x86 (IA-64 subset only)
> Dynix:  x86 (only Sequent hardware)
> AIX:    Power and PowerPC (s390?)
> Mac OSX:
>         PowerPC
> Tru64 (incl Dec Unix, OSF/1, etc):
>         Alpha and little-endian MIPS
> AUX:    No longer available. We _can_ forget this one.
> BSD (incl OpenBSD, NetBSD, FreeBSD):
>         Alpha, HP300, PA-RISC, x86 & IA64, Motorola 68k (incl VME, HP300, Amiga & Atari),
>         PowerPC (Apple, BeBox), SPARC, SPARC64, VAX, ARM & StrongARM, Amiga (m68k & ppc),
>         Hitachi SuperH: SH3 (dreamcast!), SH4, SH5

Don't forget the Sony Playstation 2, and all the other systems that
NetBSD runs on!

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

On 27 Jul 2003 09:41:29 GMT
js@cs.tu-berlin.de (Joerg Schilling) wrote:

> Sun's statements are illegal and thus void!
> 
> Sun is not allowed to prohibit people to sell a machine _with_ a valid
> license.

I think you're right in stating that (at least in Germany, and
probably in Europe) Sun cannot prohibit owners from selling their
Sun systems with the OS licenses they came with. That doesn't
mean of course that Sun must supply the new owners with a 
Solaris 9 license for free. 

-- 
Stefaan
-- 
"What is stated clearly conceives easily."  -- Inspired sales droid

>
> > Even the monitor cables are propietary even if all have that damn 13w3
> > connectors :-(
>
> The reason for those connectors is that they're much better
> quality than the commodity crap that people use with the HD15
> connectors.
>
> Also, how are 13W3 connectors any more proprietory than HD15?
>

It's not the connector itself:

The pinout of the 13w3 is very different for each of Sun/SGI/IBM.
The HD15 is well accepted standard.
This requires workstation users buying a monitor cable for horror prices.

---
Uli

On Sun, 27 Jul 2003 22:19:51 +0000, Rich Teer wrote:

> I have no faith in a legal system that lets idiots sue companies
> like McDonalds - AND WIN! - because they burned themselves by
> spilling the hot coffee in their cup into their lap.  "But there
> was no warning on the cup, saying the coffee is hot and I shouldn't
> put it into my lap, therefore McDonalds is negligent, therefore I
> want my big cash payment".

One of the primary reasons for such failures in our legal system is that
there simply are too many lawyers and half of them passed their bar exams
in the lower 50 percentile.  Smart, successful lawyers remain in private
practice where they can earn obscenely high incomes.  The remainder become
public prosecutors, defenders, judges and members of the legislative
bodies.

There is no reason to expect sanity in laws and court decisions in this
environment.
 
> Sorry about that; needless litigation is one of my pet peeves!

No litigation is needless.  All the lawyers involved -need- to make their
mortgage, BMW and boat payments.  They also -need- to support their
mistresses and pay the college expenses for their offspring to bring them
to the same level of mercenary competence as their parents.

Dr. David Kirkby wrote:
> Bob Hoekstra wrote:
> 
> 
>>OK, let's run down the major Unix variants in no partcular order and see what tey run on
>>(please correct me if there are any mistakes):
>>
>>SunOS:  SPARC, SPARC64 and x86 (limited support) (Motorola 68k VME for SunOS 3)
>>Irix:   big-endian MIPS
>>HP-UX:  PA-RISC and x86 (IA-64 subset only)
>>Dynix:  x86 (only Sequent hardware)
>>AIX:    Power and PowerPC (s390?)
>>Mac OSX:
>>        PowerPC
>>Tru64 (incl Dec Unix, OSF/1, etc):
>>        Alpha and little-endian MIPS
>>AUX:    No longer available. We _can_ forget this one.
>>BSD (incl OpenBSD, NetBSD, FreeBSD):
>>        Alpha, HP300, PA-RISC, x86 & IA64, Motorola 68k (incl VME, HP300, Amiga & Atari),
>>        PowerPC (Apple, BeBox), SPARC, SPARC64, VAX, ARM & StrongARM, Amiga (m68k & ppc),
>>        Hitachi SuperH: SH3 (dreamcast!), SH4, SH5
> 
> 
> Don't forget the Sony Playstation 2, and all the other systems that
> NetBSD runs on!

I didn't forget them David - the above list is by processor.  The Sony PS2 is in fact a 
MIPS processor IIRC. If I'd listed by architecture/board/manufacturer I'd still be typing. 
For the fanatics http://www.netbsd.org/Ports/ makes impressive reading. Personally, I 
prefer http://www.openbsd.org/plat.html as I think OpenBSD has some benefits not provided 
by the others. But we are all entitled to our opinions.

-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GO/! d- s++:+ a+ C++(++++) US++++$ UB++ U*++ P+++ L+++ E--- W+++ N++ w--- O-
M+ V- PS+ PE+ Y+ PGP t+ 5++ X R* tv+ b+ DI++ D G e(*) h++/-- r+++ y?
------END GEEK CODE BLOCK------

-----------------------------------------------------
Bob Hoekstra: APL & Unix Consultant
Telephone:    +44 1483 771028
Mobile:       +44 7710 562345
Email:        Bob.Hoekstra@HoekstraSystems.ltd.uk
-----------------------------------------------------

Bob Hoekstra wrote:
> 
> Dr. David Kirkby wrote:
> > Bob Hoekstra wrote:
> >
> >
> >>OK, let's run down the major Unix variants in no partcular order and see what tey run on
> >>(please correct me if there are any mistakes):
> >>
> >>SunOS:  SPARC, SPARC64 and x86 (limited support) (Motorola 68k VME for SunOS 3)
> >>Irix:   big-endian MIPS
> >>HP-UX:  PA-RISC and x86 (IA-64 subset only)
> >>Dynix:  x86 (only Sequent hardware)
> >>AIX:    Power and PowerPC (s390?)
> >>Mac OSX:
> >>        PowerPC
> >>Tru64 (incl Dec Unix, OSF/1, etc):
> >>        Alpha and little-endian MIPS
> >>AUX:    No longer available. We _can_ forget this one.
> >>BSD (incl OpenBSD, NetBSD, FreeBSD):
> >>        Alpha, HP300, PA-RISC, x86 & IA64, Motorola 68k (incl VME, HP300, Amiga & Atari),
> >>        PowerPC (Apple, BeBox), SPARC, SPARC64, VAX, ARM & StrongARM, Amiga (m68k & ppc),
> >>        Hitachi SuperH: SH3 (dreamcast!), SH4, SH5
> >
> >
> > Don't forget the Sony Playstation 2, and all the other systems that
> > NetBSD runs on!
> 
> I didn't forget them David - the above list is by processor.  The Sony PS2 is in fact a
> MIPS processor IIRC. If I'd listed by architecture/board/manufacturer I'd still be typing.
> For the fanatics http://www.netbsd.org/Ports/ makes impressive reading. Personally, I
> prefer http://www.openbsd.org/plat.html as I think OpenBSD has some benefits not provided
> by the others. But we are all entitled to our opinions.

I appologise. 

As you say, the list of ports for NetBSD is very impressive - probably
more than for Linux even.

On SPARC at least, there is some support for multiple CPUs on NetBSD.
As far as I'm aware, there is no such support in the OpenBSD or
FreeBSD, and nor is there likely to be in the near future. 

I also had a real hard time installing OpenBSD on SPARC, which rather
put me off of it. But it seems to run okay once installed. I've not
used it much, but it does seem quite sable and functional. 


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Oscar del Rio wrote:
> 
> "Dr. David Kirkby" <drkirkby@ntlworld.com> wrote
> > As someone who works in a University, I am amazed UNIX vendors don't
> > give away free copies of their operating systems to students and home
> > users.
> 
> Have you checked the Sun EduSoft Portfolio for Individial Users?
> http://www.sun.com/products-n-solutions/edu/promotions/edusoft/individual/
> 
> Students and faculty can download Solaris and many other Sun products,
> including compilers, sparc AND x86, for academic and research use,
> for free.

Last time I looked at that link, I was convinced it would cost $495,
which is hardly free. But I must admit, it does appear
students/educators can now download many Sun products for free,
although I'd need to read the licence conditions more carefully to be
sure. 

*IF* this is so, that is a very positive step by Sun. 

It would be nice if Sun extended that to home users too. I don't
suppose there are than many home users buying Sun compilers at full
commercial prices, so I doubt it would cause a significant loss of
software sales revenue to Sun. It would however convince such users of
the superiority of the Sun tools (compared to gcc/gdb), so that they
are more likely to buy such tools if they use Suns professionally, or
might do so at a later date. 

Although others might disagree, I don't think there is a great deal of
point running Linux, FreeBSD, NetBSD or OpenBSD on UNIX workstations
from Sun, HP, SGI, IBM, Dec, .... etc. I personally feel it is better
run the operating systems supplied by the manufacturers of the
hardware. 

Unless students and home users can obtain Solaris, HP-UX, IRIX, AIX,
Tru64 or whatever free (or at very reasonable costs), they won't buy
the hardware. Despite what a certain idiot on here seems to think, I
believe a significant fraction of people buying such hardware are
likely to be in a position to influence the buying of UNIX hardware,
either now, or at a later date. They are more likely to suggest/appove
systems they know are good. 
 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Doug Dingus wrote:

> > > Your best free UNIX is Linux.
> >
> >       Oh? And I suppose you can point us all to Linux's UNIX certification?
> 
> Who cares really.

	It matters to those who are concerned with these things.

> (About the cert.)  Linux is not a UNIX.

	Correct, its not.

> It does act like one though.

	It does? But then why doesn't it call itself one?

> From a students point of view, Linux is a very
> good OS to learn which is why I wrote what I did.

	Why?

> One could argue the BSD family is better/worse than Linux.

	One could but one would need to provide facts.

> I would likely agree on a number of points there, but Linux has the
> mindshare right now along with momentum.

	And exactly how much mindshare and momentum does it have?

> Again, from a students point of view, that's a good thing.

	Is it? How?

> Your point is valid, if hasty.  Maybe I should have written:  "Your
> best all-around free UNIX workalike is Linux."  Either way, the intent
> is the same.

	That would yet need to be determined. What other free UNIX or
	UNIX-like choices are available and how do they compare?

> I did not say Linux was the best UNIX.

	Since its not UNIX, I don't think you could.

> My personal commercial UNIX preference is IRIX :)  I wonder how things
> will continue to go though.

	Based on the some posts I see from time to time, I think you
	are not the only one who is concerned about the current state
	of things in the industry.

>  Linux has gotten quite good in a short amount of time.  Who knows, it
> might end up the best yet.

	When you say "quite good in a short amount of time" do you imply
	that it wasn't good before then? How long has it been around now
	and when did it become "good"? How do you determine this goodness?
	From its UNIX compatibility? Stability? Or something else?

> > > Students that learn to use Open Source code will be able to do much
> > > of their computing for free on just about any hardware they can find.
> >
> >       How do you figure that?
> 
> OSS runs on many different hardware types and vintages.  My use of the
> word free is a little loose.  It does cost something to get the code
> and it does take time to build / configure it on the target hardware.
> It does not take lots of hard cash and does not have license fees
> attached along with their goofy EULA restrictions.  These things make
> the majority of hardware useful long after its expected service life.

	Being a little loose with words and their meaning doesn't help.
	I'd like to know what real advantage Open Source code imparts.
	As far as I can tell, its possession is not useful for the vast
	majority of students.

> All you have to do is learn to compile code and understand how the OS
> and hardware interact.

	And how do rate this learning? Easy or hard?

> From there, many applications are for the
> taking.  OSS is perfect for this because you can see down to the core
> of what is happening if you want to.  --You can trust the machine to a
> very high degree.

	The machine or the code? And how do you develop this understanding?

> Getting things done becomes a matter of choices at this point.  Does
> the hardware support the task or not?  Do I run the latest greatest
> KDE / Gnome whatever, or do I run some older software instead?  Open
> Standards allow communication across hardware and software versions to
> a high degree.  These things combine to allow most basic (e-mail, web,
> word processing, etc...) tasks to happen on most any machine.

	And what happens when everyone changes their own version of the
	Open Source code to suit their own particular needs? What
	happens to the standards?

> Any student who takes the time to realize this, will be in control of
> their computing experience.  They will understand they have choices
> beyond those few they are presented and will be in a position to take
> advantage of them if they choose.

	And do they choose to do so? Why or why not?

> Given the high costs of secondary education, students should be
> learning just a bit more than how to use the latest commercial
> offerings.  Educators are fools if they don't present them.  --That's
> the other reason I wrote what I did.

	Secondary or tertiary education? And in exactly what fields of
	education? How does it help someone learning dramatic art for
	example?

> "If you can learn how to compile software, you will be set for life."
> About 5 years ago, someone told me this when I was looking for
> applications on some older SGI hardware.  I took a chance and started
> learning.  Today, I find they were right.

	I would argue that if I win the lottery I would be set for life.
	I learnt to compile once and it hasn't set me for life. I know
	plenty of people who are set for life and know nothing of
	compiling. I could just as easily argue that if I learn to tie
	shoelaces I would be set for life. But this is just an empty
	statement that carries no substance without any hard proof to
	support it.

-am	� 2003

Frankly, I've seen your name on other posts to this list. If i remember
right, most of the time it wasn't to offer any help or info, but to ask for
free this or that or to carp why there wasn't enough free available to
university students and faculty.
I get paid for what i do, and just assume that everyone else deserves to get
paid for what they do - monetary, in-kind, or psychic reward, their pick.
The beggar has the right to ask, but no right to insist. It would be a
courtesy, i think, to accept graciously that which is free, and just shut up
about what rightfully is not. My $.02.
And, btw, the big scholastic giveaway did not  appear to pan out as per
Apple and Digital.
And, lastly, my profound apologies to the massive cross-post - it is part of
the original poster's reply-to.

"Dr. David Kirkby" <drkirkby@ntlworld.com> wrote in message
news:3F233C74.4F4C3BAB@ntlworld.com...
> As someone who works in a University, I am amazed UNIX vendors don't
> give away free copies of their operating systems to students and home
> users. If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> their OS releases, home users would be more likely to buy cheap used
> HP kit. When those home users go in to industry, they are more likely
> to buy **EXPENSIVE** new HP kit. So it's to HP's advantage to give
> away HP-UX to students and home users.
>
> The same can be said for any UNIX hardware vendor - I strongly suspect
> it's in their long term economic interest to encourage people to learn
> their operating systems. Once you know one commercial UNIX operating
> system, you are likely to want to stick to it.
>
> Why on earth Sun don't change the licence conditions of their
> so-called free licence for Solaris on SPARC, I don't know. Solaris can
> be freely downloaded, yet its illegal to use it on virtually all used
> Sun kit sold. Their previous licence, which allowed for its use on
> machines of up to 8 CPUs seemed more sensible.
>
> HP have a reasonably attractive package that allows non-commerical
> users to gain the latest copy of the Tru64 operating system and the HP
> C compiler for Tru64 for $100. Yet HP have no such policy with their
> other UNIX operating system (HP-UX).
>
> I'd like to know HP's logic of the differing policies of Tru64 and
> HP-UX. It seems very odd to sell Tru64 and compiler for $99, yet give
> no discounts on the HP-UX.
>
> One can't expect Microsoft to give Windoze away for free, as Microsoft
> will never gain any extra revenue from hardware sales. But SGI, HP,
> IBM, Sun etc. will gain extra sales of hardware, by allowing home
> users to learn their operating systems, rather than that of their
> competitors.
>
> One can't expect SCO to give away SCO UNIX, as again they will gain no
> extra revenue from hardware sales, although they might have done with
> software sales had Linux never existed. But I don't think any home
> user or student would care less whether or not SCO was given away now.
>
>
>
> --
> Dr. David Kirkby,
> Senior Research Fellow,
> Department of Medical Physics,
> University College London,
> 11-20 Capper St, London, WC1E 6JA.
> Tel: 020 7679 6408 Fax: 020 7679 6269
> Internal telephone: ext 46408
> e-mail davek@medphys.ucl.ac.uk

Bob Hoekstra wrote:

> No. Linux is a Unix-like OS, whether you want to admit it or not.

	And where is it admitted on the Linux sites?

> If I was employing a student or junior staff member, I would prefer one
> with Linux experience to one with M$ experience. Actually, I'd prefer one
> with no experience to one with M$ experience.

	That may be your preference but is that the norm?

> At least Linux has not given up the fight for the commercial desktop as the real Unix
> variants have. If you "forget Linux" you are, professionally and stupidly, shooting
> yourself in the foot before putting it firmly in your mouth.

	So its in a fight for the commercial desktop. And what percentage
	of that market does it command?

> >>Students that learn to use Open Source code will be able to do much
> >>of their computing for free on just about any hardware they can find.
> >
> >       How do you figure that?
> 
> OK, let's run down the major Unix variants in no partcular order

	Why? What does that tell us about the advantage of Open Source code
	to students?

> So what Doug is saying is that if a student has a computer, he will probably be able to
> run Linux on it.

	Assuming that said student knows how to port if required.
	How many know how to do that? Does Open Source code teach
	them that?

> Only the *BSD family compare. But you probably don't consider these to be
> Unix either.

	Are they certified as such?

> Or maybe your question referred to other considerations like tools, compilers, editors,
> hardware support, etc.

	No my question is what advantage does Open Source code offers students.

> Now that ends my contribution. I am now going to stop reading posts from Anthony Mandic as
> I find them a waste of time. I am a fan of Monty Python, but I have not paid for a 5
> minute argument...

	Another quitter.

-am	� 2003

In article <bg1iu4$2d7$1$8300dec7@news.demon.co.uk>,
Bob Hoekstra  <Bob.Hoekstra@HoekstraSystems.ltd.uk> wrote:
>Yuan wrote:
>> As much as I don't like the Linux hype, I still have to admit it is a
>> poprlar OS.
>> At least I can't find any Unix magazine, there isn't a magazine I know just
>> for Unix
>> but quiet a few for Linux.
>
>Try Sys Admin (http://www.sysadminmag.com/) and become a member of Sage 
>(http://www.sage.org/) to receive ;login:
>
>You are right, the rest have all disappeared. However, much of what you
>read in Linux mags 
>is applicable to Unix. Pity that so much is at a very simple level :-(
			.
			.
			.
It does indeed hurt that all the printed Unix magazines,
apart from *SA* and *;login*, have indeed folded.  These
two certainly reward attention.  Online, UnixReview.com
joins them.
-- 

Cameron Laird <Cameron@Lairds.com>
Business:  http://www.Phaseit.net
Personal:  http://phaseit.net/claird/home.html

GT Jeff Cherer wrote:
> 
> Frankly, I've seen your name on other posts to this list. If i remember
> right, most of the time it wasn't to offer any help or info, but to ask for
> free this or that or to carp why there wasn't enough free available to
> university students and faculty.

I don't know what "this list" is, since I did cross-post to several
lists, which were relevant. One to a Sun list, one to an SGI list, one
to an AIX list ..etc.
Despite the fact there are numerous Sun lists, I posted to only one. 

In the case of the non-Sun lists, then I would admit I've never given
much help, since I know very little about AIX, HP-UX, IRIX or Tru64.
There is no point on me offering advice on a subject I know very
little about. 

Despite the fact I am not a professional system administrator, I have
however given considerable help to others on lists related to Suns. 

I just did a search for my name on Google and found I'd made 441 posts
to comp.sys.sun.solaris, of which very few were asking for anything
free. The vast majority either seeking help on a Solaris issue, or
offering advice where I thought I was able to. Perhaps you would be as
kind as to count the number of posts asking for free items, and tell
us what fraction that is. 

I tend to read comp.sys.sun.admin more than comp.sys.solaris, so there
you will find 2410 posts. Perhaps you would be as kind as to count the
number of posts asking for free items, and tell us what fraction that
is.

I've made 1400 posts to comp.sys.sun.hardware. Perhaps you would be as
kind as to count the number of posts asking for free items, and tell
us what fraction that is. 

Do a search on my name and the word free in comp.sys.sun.wanted and
you will find 0 posts from me asking for free items, but a couple
offering unwanted items for no charge.


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring wrote:
> 
> On Sun, 27 Jul 2003 19:07:30 +0100, Dr. David Kirkby wrote:
> 
> > No sure (how do I find out ??), but the name rings a bell.  The 600
> > MHz Alpha seems so slow (despite me now using the right compiler
> > switches), that I don't tend to use it much.
> 
> The Alpha systems are not up right now.  IIRC, the command which displays
> your licenses is 'lmf'.

You do seem to have remembered correctly, as lmf is the right command. 

It seems I do indeed have the OSF-DEV license. Whether or not that
gives me a working C++ compiler I don't know, but the C compiler
functions okay. 

lmf> list
Product                   Status                     Users: Total     
Active

OSF-DEV                   active                            unlimited
OSF-SVR                   active                            unlimited
OSF-USR                   active                            unlimited
OSF-BASE                  active                            unlimited


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring <daveuhring@yahoo.com> writes:

>IIRC, Microsoft also attempted to void the Windows license on transfer of
>ownership of the machine and a German court struck down that restriction.
>Unfortunately, those courts' jurisdictions do not extend to the rest of
>the world WRT such license restrictions.

But the German court in this case was following European law, I believe.
(IANAL).

Casper
-- 
Expressed in this posting are my opinions.  They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

"Dr. David Kirkby" <drkirkby@ntlworld.com> writes:

>> >>SunOS:  SPARC, SPARC64 and x86 (limited support) (Motorola 68k VME for SunOS 3)

There was one release for PowerPC systems (2.5.1) as well.
Aside: SPARC64 is not "64 bit SPARC" but rather a Fujitsu-Siemens trademark
for their family of SPARCv9 processors.

Casper
-- 
Expressed in this posting are my opinions.  They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

Nico Kadel-Garcia <nkadel@verizon.net> writes:

>Her case had a lot of similar complaints documented in the court 
>records, and the coffee was in fact hotter than that at most coffee 
>vendors. (It tastes better that way and you get more coffee per pound of 
>beans.)

>McDonald's thereafter lowered the temperate of the coffee, I think. It's 
>certainly seems cooler than it was before that case.

To which we reply "So what".  I don't think the coffee is any colder;
it now carries a warning label in the US.  Something like "This coffee
is hotter than ordinary coffee" or some such.

That's about as silly as "Objects are closer than they appear" on
the right wing mirror.



		    GENERIC WARNING SIGN
		STATING THE BLEEDING OBVIOUS
	    (but our lawyer made us put it here)

Casper
-- 
Expressed in this posting are my opinions.  They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

> It's not like I'll be phoning HP for tech support either so it costs HP
> nothing.
> 

Although I agree with your philosophy, I would just like to point out
that "it also doesn't cost games manufacturers if someone copies their
games".  My point is that I'm sure HP would rather sell everyone a
copy than let them make their own (copies) and lose potential revenue.
 But I'm all for student discounts/freebies and hobby licenses.  There
should be a way of registering yourself as a hobby/student user imo
and be able to have the software legally for non-commercial use.
I think this should also apply to all business software vendors,
except maybe those that already have a home market like Microsoft and
Apple, but at the same time older OS's should be free (like stuff
which is 2-3 years old or more because let's face it they're not
likely to sell those in vast quantities anymore).

As a final point, SGI make little or no effort to hunt down students
who use their old versions of IRIX, as has probably already been
mentioned in the sgi newsgroups elsewhere, even when it has been
blatently made available free online by non-sgi retailers.  I like
this policy, it lets the little people use the big software ;)

Anthony Mandic <ny@hotmail.com> writes:

> Doug Dingus wrote:

>> Agree here, but you forget Linux.

> 	Yes, I agree, you can forget Linux.

>> Your best free UNIX is Linux.

> 	Oh? And I suppose you can point us all to Linux's UNIX
> 	certification?

You mean the one that VMS got first, followed by MVS?

see figure 2...

-- 
Paul Repacholi                               1 Crescent Rd.,
+61 (08) 9257-1001                           Kalamunda.
                                             West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.

Kjetil Torgrim Homme <kjetilho@yksi.ifi.uio.no> writes:

> here's what our students at the Department of Informatics can get for
> free:

>   1. Windows XP pro
>   2. Windows 2000 pro
>   3. Windows 2000 Server, Windows 2000 Advanced Server
>   4. Windows Server 2003 Web Edition
>   5. Windows Server 2003 Std. Ed., Windows Server 2003 Ent. Ed.
>   6. MS Visio Professional 2002
>   7. MS Project Professional 2002
>   8. Visual J# .NET
>   9. Visual Studio .NET 2002 (CD5 = ComponentUpdate) (5 CD-er)
>  10. Visual Studio .NET 2003 (CD3 = Prerequisites) (3 CD-er)
>  11. Visual Studio .NET 2003 Student Tools
>  12. MSDN Library .NET 2002 (April 2003) (3 CD-er)
>  13. MSDN Library .NET 2003 (for VS.NET 2003) (3 CD-er)

> only one licence of each, and they must supply the CD-Rs themselves ;-)

How much does it cost the institution? M$ charge at least one local
uni ~$3M/yr, plus a head cost for EVERY x86, running billycrap or not.
(or did a few years ago. Plus their admin seems to have been totally
BDSed as well)

-- 
Paul Repacholi                               1 Crescent Rd.,
+61 (08) 9257-1001                           Kalamunda.
                                             West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.

"James Wilson" <aopc98@dsl.pipex.com> writes:

....

> From my own view point its a fantastic way to learn HP-UX which
> surely is in HP's best interest? If I hadn't saved this monster it
> surely would have got scrapped and whats the point of that? The fact
> is that the machine would have originally had a HP-UX licence with
> it so I have no problem with "illegally" running HP-UX on it
> now. What possibe use is HP-UX without the HP hardware? nothing at
> all. Are there any HP hardware clones? I don't think so!

See if you can find the original paperwork and license. You *may* only
have to inform HP to be 100% legal.

> It's not like I'll be phoning HP for tech support either so it costs
> HP nothing.

Lucky you :) 

-- 
Paul Repacholi                               1 Crescent Rd.,
+61 (08) 9257-1001                           Kalamunda.
                                             West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.

On Mon, 28 Jul 2003 09:12:26 +0100, Dr. David Kirkby wrote:

> Dave Uhring wrote:
>> 
>> On Sun, 27 Jul 2003 19:07:30 +0100, Dr. David Kirkby wrote:

> It seems I do indeed have the OSF-DEV license. Whether or not that
> gives me a working C++ compiler I don't know, but the C compiler
> functions okay. 
> 
> lmf> list
> Product                   Status                     Users: Total     
> Active
> 
> OSF-DEV                   active                            unlimited
> OSF-SVR                   active                            unlimited
> OSF-USR                   active                            unlimited
> OSF-BASE                  active                            unlimited

I'm pretty sure that the C++ compiler requires CXXOSF.  No free licenses
seem to be available except for a 60-day trial version.

On Mon, 28 Jul 2003 09:40:23 +0000, Casper H. S. Dik wrote:

> Dave Uhring <daveuhring@yahoo.com> writes:
> 
>>IIRC, Microsoft also attempted to void the Windows license on transfer of
>>ownership of the machine and a German court struck down that restriction.
>>Unfortunately, those courts' jurisdictions do not extend to the rest of
>>the world WRT such license restrictions.
> 
> But the German court in this case was following European law, I believe.
> (IANAL).

If true that extends sanity regarding license issues to a broader area.

Jonathan wrote:
> 
> > It's not like I'll be phoning HP for tech support either so it costs HP
> > nothing.
> >
> 
> Although I agree with your philosophy, I would just like to point out
> that "it also doesn't cost games manufacturers if someone copies their
> games".  My point is that I'm sure HP would rather sell everyone a
> copy than let them make their own (copies) and lose potential revenue.

I think the analogy between games and UNIX software is flawed for a
number of reasons:

1) In the case of games manufacturers, they would indeed loose revenue
if they allowed everyone to copy their games. I don't think this is so
with UNIX operating systems and compilers. There are not than many
home users going to pay  more than $1000 on a compiler. They will
likely use gcc instead and have to accept the poorer performance. 

2) Someone copying a Sony Playstation 2 game, is unlikely to be in a
position of spending lots of money on new Sony games hardware or
software when in industry. So Sony would never gain anything in the
future. In constrast someone knowing Solaris/HP-UX/AIX etc. as a
student or home user, is likely to be in a position of influencing the
purchase of UNIX hardware when they work. 

3) If anyone copies a Sony Playstation 2 game, they unlikely to think
when next buying a games console "I will buy a Sony product, as I know
how to use that and don't want to incur a steep learning curve to
learn how to use a new games console." However, given a choice of UNIX
vendors, they are likely to pick the ones they know. 

Hence Sony (I use them as an example as it's the only manufacturer of
games I know of), would *loose* in every way if they did not try to
stop piracy. The commercial UNIX vendors would I feel *gain* from
allowing home/student users to obtain their software for zero or
minimal cost.  


>  But I'm all for student discounts/freebies and hobby licenses.  There
> should be a way of registering yourself as a hobby/student user imo
> and be able to have the software legally for non-commercial use.

Yes agreed. 

> As a final point, SGI make little or no effort to hunt down students
> who use their old versions of IRIX, as has probably already been
> mentioned in the sgi newsgroups elsewhere, even when it has been
> blatently made available free online by non-sgi retailers.  I like
> this policy, it lets the little people use the big software ;)

It is not nice to have to do something illegal though. If SGI don't
mind people doing this, they should make it legal. It would be
relatively easy for them to make the software unattractive to
commerical pirates - an issue that I don't think is a big problem. 

To be fair to the UNIX vendors, I have managed to obtain quite legally
many items at reduced or zero cost. IBM in particular have been very
helpful. These have usually been under academic agreements, and so are
not available to the average home user. 

(To save Anthony Mandic the bother of replying, I'll tell you now I
have no intension of answering *any* responses from you. Your
stupidity is in a class of its own.)

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Casper H.S. Dik <Casper.Dik@Sun.COM> writes...
}Nico Kadel-Garcia <nkadel@verizon.net> writes:
}>Her case had a lot of similar complaints documented in the court 
}>records, and the coffee was in fact hotter than that at most coffee 
}>vendors. (It tastes better that way and you get more coffee per pound of 
}>beans.)
} 
}>McDonald's thereafter lowered the temperate of the coffee, I think. It's 
}>certainly seems cooler than it was before that case.
} 
}To which we reply "So what".  I don't think the coffee is any colder;
}it now carries a warning label in the US.  Something like "This coffee
}is hotter than ordinary coffee" or some such.
} 
}That's about as silly as "Objects are closer than they appear" on
}the right wing mirror.
} 
} 
} 
}		    GENERIC WARNING SIGN
}		STATING THE BLEEDING OBVIOUS
}	    (but our lawyer made us put it here)
} 
}Casper

Congratulations on being an ignoramus.
Before blathering you should do research.

Yes, there are a lot of frivilous lawsuits. Yes some of them result
in the addition of yet another moronic warning label to something.
But this was not one of those.

The woman in question was hospitalized for week due to 3rd degree burns.
Not just 1st degree burns, or even 2nd degree burns, but 3rd degree burns.
She required skin grafts.

Have you ever spilled coffee on yourself? Did you require the dead tissue
from where it touched you to be surgically removed and replaced via skin
grafts? I suspect not. This wasn't just plain hot, it was stupid hot.

The temperature they served the coffee at at that time was 185 degrees,
give or take 5, as required by their operations manual. That is 20 or
more degrees higher than anyplace else served it. Since then the temperature
of the coffee served at that location has been measured to be lower by
something like 25 degrees (measured by a reporter a day after the verdict). 

Besides which, there had been over 700 complaints, some of which also
involved 3rd degree burns. They had already paid out some unkown amount
of money to settle some of these - something over half a million dollars
in total.

The lawyer she hired had been involved in a case a few years earlier
for pretty much exactly the same thing. For that case he had questioned
a McDonald's quality assurance manager who said he was aware of the risk
and had no plans to reduce the temperature. In that earlier case they
settled for $27,500. In another earlier case, a McDonalds employee spilled
coffee while passing it out to the customer causing 2nd and 3rd degree
burns - they settled that case for $230,000. That's right. They had already
paid over half a million dollars to people for getting burns from their
product and were aware of the risk of serious burns, and yet they had no
plans to change it. That's where the willfull negligence comes in.
The same quality assurance manager testified in this trial. He reported
that McDonalds was aware that it's coffee sometimes caused serious burns,
but that they had never consulted a burn expert. He also said that
McDonalds had decided not to warn their customers about the possibility
of severe burns even though they would not expect it to be possible.
He also said they didn't plan to change they way they served coffee,
and gave the excuse that there are more serious dangers than that in
a restaurant.

She only asked for $20,000. This was pretty much the amount of her medical
bills. McDonalds instead offered all of $800. Do you think you could pay
a $20,000 medical bill with $800? I bet you couldn't. So it went to court.
Why is not clear - they had settled similar cases before for much more
than the requested $20,000. So McDonalds lost, and rightfully so. The
ammount that the jury awarded her in compensatory damages was $160,000.
The amount that was awarded as punative damages was $2.7 million - an
amount equal to the revenue McDonalds gets in roughly 2 days from its
coffee sales. (A judge later reduced it from $2.7 million to $480,000,
or triple the compensatory damages - well under what McDonalds makes in
a single day from sales of just coffee.) Then, after the fact, there was
a secret settlement. Therefore, nobody by the parties involved know what
the actual final outcome was but it was probably for less than the full
amount.

For some other facts that people often seem to get wrong: She was not
driving the car, she was a passenger. Also, the car was not moving at
the time of the coffee spill - the driver had stopped before pulling
onto the street so that she could add cream to the coffee and it was
spilled while trying to remove the plastic lid from the styrofoam cup.

All this info, and more, is available on the web.

--- Carl

Lincoln, actually, was the one who put the final nail in the
coffin of the Republic, and reaffirmed the nation's true character
to be Empire. 'For the people, etc' was rhetoric. Jefferson was the
true champion of the idealistic vision of america which for a while
was so firmly believed be so many people that it almost came into
existence. However, once matters got down to it.. all that was pushed
to the side.
ok, we're getting really offtopic now
isildur


On Sun, 27 Jul 2003, Dave Uhring wrote:

> On Sun, 27 Jul 2003 21:22:31 +0200, Thomas Dehn wrote:
>
> > The main problem with the US legal system seems
> > to be that any powerful corporation or group such as M$
> > can drown its opponents in legal costs, and thus will hardly
> > lose a lawsuit about a few thousand dollars against Joe Sixpack.
> > Drowing opponents in legal costs is impossible in Germany.
>
> And if a prosecutor with unlimited funds wins a favorable verdict then
> politics prevents effective remedy of the original cause of action.
>
> A. Lincoln's ideal of 'government of the people, by the people and for the
> people' has taken a serious beating in the last 140 years.
>
>

you have this ongoing obsession with certification. What's the deal with
that?

> > Only the *BSD family compare. But you probably don't consider these to be
> > Unix either.
>
> 	Are they certified as such?
>
> > Or maybe your question referred to other considerations like tools, compilers, editors,
> > hardware support, etc.
>
> 	No my question is what advantage does Open Source code offers students.

umm.. let's see.
freedom (not just monetary, but license-wise, legal, IP, etc type freedom also),
the means both technical and authoritative to modify the code to your heart's
content, the rich, active development community, need I continue?
I'm not sure what dungeon you live in, surely one that is certified for
something, but it would behoove you to occasionally let the sun and fresh
air in.

"Uli Link" <Ulrich--nO--(dot)-sPAM--Link@Epost.de> writes:

> > 1) In the case of games manufacturers, they would indeed loose revenue
> > if they allowed everyone to copy their games. I don't think this is so
> > with UNIX operating systems and compilers. There are not than many
> > home users going to pay  more than $1000 on a compiler. They will
> > likely use gcc instead and have to accept the poorer performance.
> 
> Don't agree: the optimization of the GCC has become much better:
> in 9 out of 10 times the GCC produces faster binaries than vendor compilers
> Perhaps not, if you will tune every single object, and put a lot of #pragma
> in your source to keep the vendor compilers from producing invalid code.
> On AIX IBM's xlc build faster common-mode bins, but if you mind peak
> performance, you should try the GCC 3.x and the cpu-specific optimizations.
> Never got working binaries from any GNU code with xlc and "-O5".
> GCC's results were up to 10% faster on PowerPC, with greater difference for
> float compared to integer.
> Same for Solaris x86, SCO and NCR's MP-RAS.
> A great time-saver of the GCC is the fix of bugs in the O/S include headers.

I have never yet seen gcc 3.2 or previous generate faster code than
Sun's C compiler.  I've seen improvements up to 10* by switching to
Sun's cc, depending on the code.  Thats with specific cpu/architecture
selection and high optimization selected for both gcc and cc,
compiling plain C code.

> 
> Another question is the productivity of an integrated dev environment and
> the turn-around-times.
> The tools from Sun are much more comfortable but tend to bring elder
> machines to the knees.

Please explain how a GUI will make me faster than a commandline, and
what a "turn around time" is.

Gregm

>the means both technical and authoritative to modify the code to your heart's
>content, the rich, active development community, need I continue?
>I'm not sure what dungeon you live in, surely one that is certified for
>something, but it would behoove you to occasionally let the sun and fresh
>air in.

The same dungeon that you live in, "The Real World".  That would be the world
where money grubbing corporations have killed any hope of the HAL9000 being a
reality in 2001 or 2010.  One in which bad software is released as product and
then we are forced to pay for the patches, which make things worse.  Also, the
same world in which similar large corporation run by grey hair management want
computers made by other big corporations that are considered stable because
the average dolt can NOT make changes as they see fit.  Until that very same
grey haired management dies off you should not expect to see anything other
than big monolithic approaches to massive central compute farms run on closed
off operating systems.  The alternative is small unstable computers run by
wide-open source operating systems that can't come close to the power and long
term stability of those monoliths.

Oh well, at least HP is shipping the DEC AlphaServer line again.  Perhaps we
will have super-computer power at the desktop cheaply on that.  Or not. It is
just nice to see the Digital label is still alive somewhere else other than my
DEC VT320 terminal.

Dennis

See www.blastwave.org for the home of open source development and software on
Solaris.  Both Intel and Sparc fully supported.

ps: yes I know that I am ranting

In article <3f24ef87$0$49104$e4fe514c@news.xs4all.nl>,
Casper H.S. Dik  <Casper.Dik@Sun.COM> wrote:
>Dave Uhring <daveuhring@yahoo.com> writes:
>
>>IIRC, Microsoft also attempted to void the Windows license on transfer of
>>ownership of the machine and a German court struck down that restriction.
>>Unfortunately, those courts' jurisdictions do not extend to the rest of
>>the world WRT such license restrictions.
>
>But the German court in this case was following European law, I believe.
>(IANAL).

I also believe so.

A big problem in international newsgroups seems to be that US citizens 
often assume that the USA is "the world". 


-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

In article <3F2368E5.ED5037A0@hotmail.com>,
	Anthony Mandic <nu@hotmail.com> writes:
> "Dr. David Kirkby" wrote:
>> 
>> As someone who works in a University, I am amazed UNIX vendors don't
>> give away free copies of their operating systems to students and home
>> users.
> 
> 	Why? Do Microsoft do this? Perhaps you should widen your
> 	job experience.

Actually, yes, they do.  We (the CS Department) subscribe to a service
with MS that allows us to give copies and licenses to our students for
OSes and development tools and they can take them with them when they
leave school.

bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill@cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>   

[Paul Repacholi]:
>
>   Kjetil Torgrim Homme <kjetilho@yksi.ifi.uio.no> writes:
>   
>   > here's what our students at the Department of Informatics can get for
>   > free:
>   >   1. Windows XP pro
>   >  [...]
>   >  13. MSDN Library .NET 2003 (for VS.NET 2003) (3 CD-er)
>   > only one licence of each, and they must supply the CD-Rs themselves ;-)
>   
>   How much does it cost the institution?

I believe it is free.

>   M$ charge at least one local uni ~$3M/yr, plus a head cost for
>   EVERY x86, running billycrap or not.  (or did a few years
>   ago. Plus their admin seems to have been totally BDSed as well)

we do pay licence fees to Microsoft for every PC, whether it's running
Windows or Linux.  as long as a Microsoft OS was bought with the PC,
we're allowed to upgrade it to whatever version we'd like to run on
it.  this is a university-wide deal, and the rest of the university is
mostly running Windows on the desktop.  our 1000 Linux PC's drown in
the 30000 total hosts, so the savings in simplified licence management
easily make up for it.

-- 
Kjetil T.			|  read and make up your own mind
				|  http://www.cactus48.com/truth.html

I know I'm going to regret this, but I just can't let this dipstick get away without a 
response :-(

Anthony Mandic wrote:
> Bob Hoekstra wrote:
> 
> 
>>No. Linux is a Unix-like OS, whether you want to admit it or not.
> 
> 
> 	And where is it admitted on the Linux sites?

 From http://www.kernel.org/:

	"What is Linux?

	Linux is a clone of the operating system Unix,
	written from scratch by Linus Torvalds with assistance
	from a loosely-knit team of hackers across the Net.
	It aims towards POSIX and Single UNIX Specification
	compliance.

	It has all the features you would expect in a modern
	fully-fledged Unix, including true multitasking, virtual
	memory, shared libraries, demand loading, shared
	copy-on-write executables, proper memory management,
	and TCP/IP networking.

	Linux was first developed for 32-bit x86-based PCs (386
	or higher). These days it also runs on (at least) the
	Compaq Alpha AXP, Sun SPARC and UltraSPARC, Motorola
	68000, PowerPC, PowerPC64, ARM, Hitachi SuperH, IBM S/390,
	MIPS, HP PA-RISC, Intel IA-64, DEC VAX, AMD x86-64 and
	CRIS architectures.

	Linux is easily portable to most general-purpose 32- or
	64-bit architectures as long as they have a paged memory
	management unit (PMMU) and a port of the GNU C compiler
	(gcc)..."

Is that clear enough? Even a half-wit like you should grasp that ... or maybe you need 
twice the brain cells before you reach that status?

>>If I was employing a student or junior staff member, I would prefer one
>>with Linux experience to one with M$ experience. Actually, I'd prefer one
>>with no experience to one with M$ experience.
> 
> 	That may be your preference but is that the norm?

Why do I care? I was talking for myself, about myself, and in the first person. You say 
you take care with your spelling. Pity about your reading skills.

>>At least Linux has not given up the fight for the commercial desktop as the real Unix
>>variants have. If you "forget Linux" you are, professionally and stupidly, shooting
>>yourself in the foot before putting it firmly in your mouth.
> 
> 	So its in a fight for the commercial desktop. And what percentage
> 	of that market does it command?

Small but growing. What percentage does your favourite Unix have?

>>>>Students that learn to use Open Source code will be able to do much
>>>>of their computing for free on just about any hardware they can find.
>>>
>>>      How do you figure that?
>>
>>OK, let's run down the major Unix variants in no partcular order
> 
> 
> 	Why? What does that tell us about the advantage of Open Source code
> 	to students?

This is as simple as I know how to put it.
Please do the following:
1. Revisit to my previous post.
2. Very carefully, and giving it your full (albeit obviously limited) attention, read the 
parts that you have decided to cull from this posting.
3. Then read Doug Dingus's statement, your question and my response again.
4. If this doesn't answer your question, repeat step 3.
5. Once you have been through steps 3 & 4 a few time, and if it still doesn't make sense, 
telephone me. For a nominal fee, my wife will provide some remedial reading (and numeracy 
if you need that as well) teaching (this is what she does). I'm sure that within a month 
or two she can get you to understand short stories written in plain English adequately.

>>So what Doug is saying is that if a student has a computer, he will probably be able to
>>run Linux on it.
> 
> 	Assuming that said student knows how to port if required.
> 	How many know how to do that? Does Open Source code teach
> 	them that?

What???? Did you actually read any of the previous post? Is there some blockage in what 
passes for your brain that actually block and/or corrupts all information coming from the 
outside?

>>Only the *BSD family compare. But you probably don't consider these to be
>>Unix either.
> 
> 	Are they certified as such?

Why should they be?

>>Or maybe your question referred to other considerations like tools, compilers, editors,
>>hardware support, etc.
> 
> 	No my question is what advantage does Open Source code offers students.

That is your current question. I was dealing with your previous one: "How do you figure that?"

But sometimes it's convenient to change the question if you don't like the answer, isn't it?

But in answer to the new question, Open Source code offers to the students exactly the 
same as any other code does. Oh, but wait, they don't have the other code, do they? In 
short, it offers them the ability to debug, correct, improve or screw it up according to 
their ability. However, that has never been my point. The OS movement has provided us with 
a mound of high quality software for all types of usage, much of which compares favourably 
with what is available commercially. But why repeat myself ... you're not listening anyway.

>>Now that ends my contribution. I am now going to stop reading posts from Anthony Mandic as
>>I find them a waste of time. I am a fan of Monty Python, but I have not paid for a 5
>>minute argument...
> 
> 	Another quitter.

You may be a plonker, but you really knows how to get up someone's nose. In my defense, I 
am certainly not a quitter, although I obviously lack the good sense to abide by my 
earlier decision.


-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GO/! d- s++:+ a+ C++(++++) US++++$ UB++ U*++ P+++ L+++ E--- W+++ N++ w--- O-
M+ V- PS+ PE+ Y+ PGP t+ 5++ X R* tv+ b+ DI++ D G e(*) h++/-- r+++ y?
------END GEEK CODE BLOCK------

-----------------------------------------------------
Bob Hoekstra: APL & Unix Consultant
Telephone:    +44 1483 771028
Mobile:       +44 7710 562345
Email:        Bob.Hoekstra@HoekstraSystems.ltd.uk
-----------------------------------------------------

"Bill Gunshannon" <bill@cs.uofs.edu> wrote:
> In article <3F2368E5.ED5037A0@hotmail.com>,
> Anthony Mandic <nu@hotmail.com> writes:
> > "Dr. David Kirkby" wrote:
> >> 
> >> As someone who works in a University, I am amazed UNIX vendors don't
> >> give away free copies of their operating systems to students and home
> >> users.
> > 
> > Why? Do Microsoft do this? Perhaps you should widen your
> > job experience.
> 
> Actually, yes, they do.  We (the CS Department) subscribe to a service
> with MS that allows us to give copies and licenses to our students for
> OSes and development tools and they can take them with them when they
> leave school.

And how much do "you" (the CS department, resp. the whole
university) pay to M$ per year in total? Not just for that
"service", but for everything.


Thomas

"Dennis Clarke" <dclarke@blastwave.org> wrote in message
news:Pine.GSO.4.53.0307281226110.7975@blastwave...

[snip]

> Oh well, at least HP is shipping the DEC AlphaServer line again.  Perhaps
we
> will have super-computer power at the desktop cheaply on that.  Or not. It
is
> just nice to see the Digital label is still alive somewhere else other
than my
> DEC VT320 terminal.

They never stopped shipping Alphas.  You can get them dirt cheap on Ebay or
from dealers like Island Computer, http://www.islandco.com.  Island has a
cabinet with 36 DS10L's for $20K!  I dare say you CAN'T find that much
horsepower anywhere for near that price.  In addition, Tru64 UNIX hobbyist
license can be bought from HP for $99 or just install Redhat 7.2.  No more
Alpha versions after 7.2.  :(

BTW, those DS10L's have a Compaq logo painted on them instead of DEC.  :(
I'm sure Island would gladly scrape the logos off.  :)

--
Kenneth Farmer  <><

Drop in some time...
http://www.EnterpriseUnix.org  |  http://www.Tru64.org
http://www.OpenVMS.org  |  http://dcl.OpenVMS.org
http://www.EnterpriseLinux.org  |  http://www.LinuxHPC.org

> > Don't agree: the optimization of the GCC has become much better:
> > in 9 out of 10 times the GCC produces faster binaries than vendor
compilers
> > Perhaps not, if you will tune every single object, and put a lot of
#pragma
> > in your source to keep the vendor compilers from producing invalid code.
> > On AIX IBM's xlc build faster common-mode bins, but if you mind peak
> > performance, you should try the GCC 3.x and the cpu-specific
optimizations.
> > Never got working binaries from any GNU code with xlc and "-O5".
> > GCC's results were up to 10% faster on PowerPC, with greater difference
for
> > float compared to integer.
> > Same for Solaris x86, SCO and NCR's MP-RAS.
> > A great time-saver of the GCC is the fix of bugs in the O/S include
headers.
>
> I have never yet seen gcc 3.2 or previous generate faster code than
> Sun's C compiler.  I've seen improvements up to 10* by switching to
> Sun's cc, depending on the code.  Thats with specific cpu/architecture
> selection and high optimization selected for both gcc and cc,
> compiling plain C code.
>
Don't have written that this is valid for Solaris on Sparc, but I've never
seen a factor of ten on the E450 I've compared GCC 2.95.2 and Sun's Workshop
4 a few years ago.
It is maybe??? possible to write code leading to such extreme results, but
you could never expect such factor in real life. Not even with the code for
SPECint and SPECfp which no commercial vendor will ignore.
If you can build the Apache/SSL/PHP4 running 10times as fast as the GCC
build I will do the packaging for you.
But Sun will make the next version slower, since no one buys new hw anymore
;-)

> >
> > Another question is the productivity of an integrated dev environment
and
> > the turn-around-times.
> > The tools from Sun are much more comfortable but tend to bring elder
> > machines to the knees.
>
> Please explain how a GUI will make me faster than a commandline, and
> what a "turn around time" is.

Debugging with multiple source windows is a great help. Syntax colouring
too.
Navigation in big projects with hundreds of objects a visual source browser
is no luxury.
Agree that editing with vi is faster than anything else. But most of the
time programmers are thinking or searching, not typing. Just my 2ct. I'm
more an admin for programmers than a programmer.

Turn around is the time from starting the preprocessor till finish of the
linker or a fatal error.
This is the time the programmer waits for the results.

---
Uli

> I mean give people SOME credit.  Who on earth thinks that if they
> buy a <insert brand> car, they'll be able to drive like a stunt
> car driver?

  errr .. I got my Vette up past 207Kph just the other day.

  I guess that qualifies me as a guy the needs a warning label on
  the dash, something along the lines of "You have exceeded the
  speed limit by a factor of two.  Slow down you DOLT!"

  Maybe something less subtle would help  ;)

Dennis

Carl Perkins wrote:

> Yes, there are a lot of frivilous lawsuits. 

This is getting off the topic somewhat, so I'm a bit reluctant to
continue this much further. However, the following points need making.

A friend of mine, born in the UK, studying English and History in the
UK , did one year of her degree in the USA around 1982. While in the
USA she saw someone injured who had fallen down some stairs. Nobody
would offer any first aid, for fear of being sued if something went
wrong. Coming from the UK she was very surprised by this. (Whether or
not she offered any aid I don't recall). 

Although the situation is not as bad in the UK as the USA, I know of
medical doctors in the UK, who will tend to turn a 'blind eye' to
those needing medical assistance in the street, airline etc, for fear
of the hassle it can cause them. (Just for the record, I have a Ph.D,
not a medical degree of any kind). 

There have been a number of companies advertising on UK television,
offering to take on trivial claims on a 'no-win no-fee' basis.
Apparently those cases that were won, the people involved often got
very little. I'm glad to see several such companies have gone bust. 

In this country at least, Law and Medicine studies take some of the
best students. It is considerably more difficult to get into
university in the UK to study Law or Medicine, that it is to study
engineering, computer science, maths etc. I wish there were were fewer
lawyers and more medical doctors. This country would be a lot better
place to live in. 


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

On Mon, 28 Jul 2003 19:31:32 -0400, Dennis Clarke wrote:

>   errr .. I got my Vette up past 207Kph just the other day.

Pretty tame.  Back in 1970 I drove across Nevada at 145 mph.  But the
speed limit then was "reasonable and proper".

Bob Hoekstra wrote:
> Anthony Mandic wrote:


> I know I'm going to regret this, but I just can't let this dipstick get away without a
> response :-(

I think we should all ignore Anthony 'dipstick' Mandic. If nobody
responded to anything he said, he would leave these newsgroups alone. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

> > > Don't agree: the optimization of the GCC has become much better:
> > > in 9 out of 10 times the GCC produces faster binaries than vendor
> compilers
> > > Perhaps not, if you will tune every single object, and put a lot of
> #pragma
> > > in your source to keep the vendor compilers from producing invalid code.
> > > On AIX IBM's xlc build faster common-mode bins, but if you mind peak
> > > performance, you should try the GCC 3.x and the cpu-specific
> optimizations.
> > > Never got working binaries from any GNU code with xlc and "-O5".
> > > GCC's results were up to 10% faster on PowerPC, with greater difference
> for
> > > float compared to integer.
> > > Same for Solaris x86, SCO and NCR's MP-RAS.
> > > A great time-saver of the GCC is the fix of bugs in the O/S include
> headers.
> >
> > I have never yet seen gcc 3.2 or previous generate faster code than
> > Sun's C compiler.  I've seen improvements up to 10* by switching to
> > Sun's cc, depending on the code.  Thats with specific cpu/architecture
> > selection and high optimization selected for both gcc and cc,
> > compiling plain C code.
> >
> Don't have written that this is valid for Solaris on Sparc, but I've never
> seen a factor of ten on the E450 I've compared GCC 2.95.2 and Sun's Workshop
> 4 a few years ago.
> It is maybe??? possible to write code leading to such extreme results, but
> you could never expect such factor in real life. Not even with the code for
> SPECint and SPECfp which no commercial vendor will ignore.
> If you can build the Apache/SSL/PHP4 running 10times as fast as the GCC
> build I will do the packaging for you.
> But Sun will make the next version slower, since no one buys new hw anymore
> ;-)

Regarding the GCC issue, I would like to state that I have seen a performance
increase with SGI's MIPSpro compilers for IRIX - most noteworthy example would
be a xmame build I did once and it performed much better with MIPSpro.  It has
also been my experience that sometimes it just depends on the program itself.

"Uli Link" <Ulrich--nO--(dot)-sPAM--Link@Epost.de> writes:
> > I have never yet seen gcc 3.2 or previous generate faster code than
> > Sun's C compiler.  I've seen improvements up to 10* by switching to
> > Sun's cc, depending on the code.  Thats with specific cpu/architecture
> > selection and high optimization selected for both gcc and cc,
> > compiling plain C code.
> >
> Don't have written that this is valid for Solaris on Sparc, but I've never
> seen a factor of ten on the E450 I've compared GCC 2.95.2 and Sun's Workshop
> 4 a few years ago.
> It is maybe??? possible to write code leading to such extreme results, but
> you could never expect such factor in real life. Not even with the code for
> SPECint and SPECfp which no commercial vendor will ignore.
> If you can build the Apache/SSL/PHP4 running 10times as fast as the GCC
> build I will do the packaging for you.
> But Sun will make the next version slower, since no one buys new hw anymore
> ;-)

They are not manufactured examples.  The 10* improvement I saw was for
a whetstone benchmark run.  I've seen other smaller factors when
compiling gzip & bzip with Sun cc for example- but I have never seen
or measured anything compiled with Sun's cc run more slowly than code
compiled by gcc.  Sometimes the speedup isn't all that apparent,
sometimes it is.

This is not to say gcc sucks- I think its a great compiler and I value
it for its ubiquity and because it does generate pretty good code.
Sometimes compiler portability wins over maximum code performance,
sometimes not.


> > >
> > > Another question is the productivity of an integrated dev environment
> and
> > > the turn-around-times.
> > > The tools from Sun are much more comfortable but tend to bring elder
> > > machines to the knees.
> >
> > Please explain how a GUI will make me faster than a commandline, and
> > what a "turn around time" is.
> 
> Debugging with multiple source windows is a great help. Syntax colouring
> too.
> Navigation in big projects with hundreds of objects a visual source browser
> is no luxury.
> Agree that editing with vi is faster than anything else. But most of the
> time programmers are thinking or searching, not typing. Just my 2ct. I'm
> more an admin for programmers than a programmer.

I use Emacs for multiple windows and alt-tab to flip to an xterm to
make and ddd to debug.  Its not an IDE and there is no mouse fiddling
and no half-baked editor to make things difficult.  I would probably
die if I had to use vi... ;)


> Turn around is the time from starting the preprocessor till finish of the
> linker or a fatal error.
> This is the time the programmer waits for the results.

So an IDE speeds this up in what way?

Gregm

Uli Link wrote:

> When you buy a *NIX workstation you will buy most likely a keyboard and
> mouse with it.
> Even the monitor cables are propietary even if all have that damn 13w3
> connectors :-(

True, but its very easy to look on eBay and find a Sun auctioned with
a Sun mouse, a Sun keyboard and a Sun monitor. Often if there is no
monitor, sellers have 13W3->VGA adapters available. 

I think it would be a lot harder to pick up a PC on eBay with all
components on the Solaris x86 HCL. 

> > It supports a small
> > subset of the hardware Linux does, making it hard to install on an
> > 'average PC'. Not exactly the sort of impression you want to create.
> 
> That's true, but it is absolutly no problem getting the components from
> Sun's HCL.

That's not true entirely true. I recently went into several PC shops
with a list of network cards on the Solaris HCL. I did not care what
card it was, as long as it was on the Solaris HCL. I could not find a
single network card in any shop that was on the Solaris x86 HCL. (I
know you could quite rightly argue I'd have a hard time picking up a
Sun keyboard/mouse locally.)

In contrast to Solaris x86, Linux seems to support a lot more hardware
- by no means everything, but a lot more. 

Despite the fact I've bought around a dozen used Suns over the years
(from an IPC to an Ultra 80), I've had far less hassle getting them up
and running than any PC with either Linux or Solaris x86. I've had
very few hardware compatibility problems with Suns running Solaris -
everything just seems to work, despite the fact much of the RAM, all
the disks, the tape drive, the modem and the scanner are all third
party. 

Given the task of buying a dozen used computers and getting them up
and running in the shortest possible time, I'd much prefer Suns to
PCs. I can't comment on HP/IBM/SGI/Dec hardware, as I have far less
experience of them. 
-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

On Tue, 29 Jul 2003 02:42:55 +0100, Dr. David Kirkby wrote:

> That's not true entirely true. I recently went into several PC shops
> with a list of network cards on the Solaris HCL. I did not care what
> card it was, as long as it was on the Solaris HCL. I could not find a
> single network card in any shop that was on the Solaris x86 HCL. (I
> know you could quite rightly argue I'd have a hard time picking up a
> Sun keyboard/mouse locally.)

Are you saying that there are no shops selling off-lease business machines
in a metro area as large as London?

The Intel NICs with 82557 or 82558 chips are fully supported by Solaris
x86 and are inexpensive, ~$10-15 USD in the St. Louis, MO, area.  They are
even more inexpensive on ebay.com, selling for as little as 5 of them for
$20 USD, including shipping within the US.

Of course, YMMV, being in Europe.  But those NICs are ubiquitous and
should be readily available to anyone with the good sense to -look- for
them.

Alan Johnson wrote:

> HEY GUYS! This is the comp.sys.hp.hpux newsgroup, not the
> debate.about.licensing.and.vendors.behaviour newsgroup. Last time I
> looked we usually discuss and resolve our technical issues here.
> Please
> go it an IRC or use IM and discuss it. A lot of vendors give their
> stuff away, thats how they get people to use it. If you are working
> with HP-UX then you one way or another have to buy it, usually... Buy
> a used server and the license is included,

Let beside that this thread also affects HP-UX in some way You're wrong,
Servers _don't_ come with an HP-UX license. Only workstations carry at least
an 2-user license of HP-UX 10.20 or above, Server licensing is different,
and if a server get's sold the license is _not_ included when selling this
thing...

Benjamin

Dr. David Kirkby wrote:

> As someone who works in a University, I am amazed UNIX vendors don't
> give away free copies of their operating systems to students and home
> users. If HP gave away HP-UX, whereas Sun, SGI, IBM etc. did not give
> their OS releases, home users would be more likely to buy cheap used
> HP kit. When those home users go in to industry, they are more likely
> to buy **EXPENSIVE** new HP kit. So it's to HP's advantage to give
> away HP-UX to students and home users.

No, it's not. You simply estimate that students playing around with <insert
any commercial unix here> automatically recommend that specific platform
when getting a job that makes them responsible for buying decisions. That
simply is wrong. Someone that makes a platform decision for some critical
stuff for it's employee mostly because he has experience with <insert any
commercial unix here> would do that exactly one time before he probably
get's fired. Every somewhat responsible and knowledgeable person simply has
to evaluate every possible solution for their certain computing problems, no
matter if he privately plays around with Sun, HP9000, SGI or whatever.
People being responsible for IT decisions usually have to know a bit more
than what they gain from playing around with an old Unix machine, and
especially they have to have at least some basic knowledge about more than
one platform. For real big investments You normally get evaluation systems
for a certain period for free from the different Unix vendors to help You to
evaluate if a platform is suitable for Your needs. Hobbyist experience
provides exactly nothing to platform decisions, especially not the bigger
ones which include a bit more than one or two workstations...

> The same can be said for any UNIX hardware vendor - I strongly suspect
> it's in their long term economic interest to encourage people to learn
> their operating systems. Once you know one commercial UNIX operating
> system, you are likely to want to stick to it.

Which is the best way to catapult itself out of the job flow. If You have to
handle and administer Unix systems the worst thing You can do is sticking to
a single Unix and/or platform. By doing that You might experience a very
high probability to get replaced by someone which is a bit more flexible. In
the business today You can't just concentrate on one thing and nothing else,
You have to be flexible and have knowledge about more than one platform or
You'll be lost. We had bunches of such guys, and we got rid of them...

> Why on earth Sun don't change the licence conditions of their
> so-called free licence for Solaris on SPARC, I don't know. Solaris can
> be freely downloaded, yet its illegal to use it on virtually all used
> Sun kit sold. Their previous licence, which allowed for its use on
> machines of up to 8 CPUs seemed more sensible.

Simply because even Sun realized that a Hobbyist program brings around zero
real $$$ renevue for them. The Solaris download program is now more aimed at
customers buying a used/refurbished system from a Sun Partner, thus making
it more attractive for smaller businesses to invest in used/refurbished Sun
hardware instead of going the PC way....

> HP have a reasonably attractive package that allows non-commerical
> users to gain the latest copy of the Tru64 operating system and the HP
> C compiler for Tru64 for $100. Yet HP have no such policy with their
> other UNIX operating system (HP-UX).

Right, simply because the Tru64 program is a relict from Compaq. It will
certainly be terminated when HP-UX and Tru64 merge as a single Unix for the
IA64 platform...

> One can't expect Microsoft to give Windoze away for free, as Microsoft
> will never gain any extra revenue from hardware sales. But SGI, HP,
> IBM, Sun etc. will gain extra sales of hardware, by allowing home
> users to learn their operating systems, rather than that of their
> competitors.

Which competitors? Except the Tru64 and OpenVMS programs there is no real
Hobbyist program any more simply because the companies give a sh*t on
Hobbyists. Why the hell should they? Don't You think that Sun isn't aware
that bunches of hobbyists download and install Solaris on their old machines
even when the licensing rules forbid that? Do You really believe HP, IBM and
SGI don't know that hobbyists copy and trade the OS CD's or buy/sell them on
ebay and install the OS on machines that are not licensed for these OS
versions? Except SGI Germany which shows some (illegal) efforts in
attempting to press people selling IRIX media kits on ebay.de to remove
their auctions the Unix vendors don't give a sh*t on what hobbyists do, as
long as they don't make money with the illegal OS installations. On the
other hand, they don't provide hobbyist programs since such programs only
cause costs without bringing any noticeable renevue. And for sure, no
Hobbyist will ever be going to spend a lot of money on a used/refurbished
system from a certified Sun reseller when he get's the same system for a lot
less on ebay and Co. And forget about a Hobbyist buying a new system...

I'm myself a Hobbyist, and I got my HP-UX 11i license for my K-Class cheaper
because we have very good connections to HP when being one of their biggest
customer. But still, I paid much more than I paid for the WindowsXP running
on my PC at home. From a hobbyist side I understand Your point of view, but
from the business side I have to deal with every day I can tell You that
mostly an hobbyist program is just lost money, no matter if it's HP, Sun,
IBM or SGI. That sound sad, but it's the truth.

Benjamin

Joerg Schilling wrote:

> If you buy a used Sun, this machine would have been run wich a
> license before. Even if the seller used a corporate license, the
> machine definitely did come with a valid license from Sun when bought
> from Sun.

Not necessarily. We use corporate licensing, and when buying Sun equipment
we buy machines without any OS license and have them integrated in the
corporate licensing scheme...

> At least this license could be sold together with the HW. If Sun
> states differenty, this is illegal and thus void.

No, it's not. You probably are thinking about the
Windows-sale-only-allowed-with-certain-hardware-case we had some time ago in
Germany. But that's totally different. The point was simply that Microsoft
can't forbid a person and/or a dealer to sell an OEM version of Windows
separately from the hardware that it came with. The problem for Microsoft
was that the consumer, when buying the PC that comes with Windows, has made
a contract with the dealer he bought the system from, and not with
Microsoft, so Microsoft has no influence on the contract conditions. The
same is valid for most smaller and midsize dealers that don't buy their
Windows from Microsoft directly but from bigger distributors, thus also not
having a direct contract with Microsoft...

In the Unix sector it's different. Most Unix customers like my employee has
a contract directly with the Unix vendor, and of course such contracts allow
the coupling of a certain software to a certain hardware, which is _not_
illegal, even not in Germany. Such contracts are indiviual contracts and
fully conform to current law.

Benjamin

"Dr. David Kirkby" wrote:
> 
> Carl Perkins wrote:
> 
> > Yes, there are a lot of frivilous lawsuits.
> 
> This is getting off the topic somewhat, so I'm a bit reluctant to
> continue this much further. However, the following points need making.
> 
> A friend of mine, born in the UK, studying English and History in the
> UK , did one year of her degree in the USA around 1982. While in the
> USA she saw someone injured who had fallen down some stairs. Nobody
> would offer any first aid, for fear of being sued if something went
> wrong. Coming from the UK she was very surprised by this. (Whether or
> not she offered any aid I don't recall).
> 
That is why many states in the U.S. have a "good samaritan's
law". This type of law gives anyone immunity for any damage
they do...if they stop and help someone who is hurt in an
accident.


--
+----------------------------------------------------------------+
|   Charles and Francis Richmond     richmond at plano dot net   |
+----------------------------------------------------------------+

In article <bg38em$4lh$00$1@news.t-online.com>,
	"Uli Link" <Ulrich--nO--(dot)-sPAM--Link@Epost.de> writes:
> 
>> 1) In the case of games manufacturers, they would indeed loose revenue
>> if they allowed everyone to copy their games. I don't think this is so
>> with UNIX operating systems and compilers. There are not than many
>> home users going to pay  more than $1000 on a compiler. They will
>> likely use gcc instead and have to accept the poorer performance.
> 
> Don't agree: the optimization of the GCC has become much better:
> in 9 out of 10 times the GCC produces faster binaries than vendor compilers

I benchmarked it against Sun's compilers probably every couple of
years since 1995, and Sun's compilers have always produced faster code,
typically by 10% - 15% on both sparc and x86 in program runtime.
Obviously will depend on the nature of the program -- something i/o
bound isn't going to notice much, whereas something CPU bound might
see even bigger differences than I did.

I also benchmarked it against DEC's OSF/1 compiler many years ago,
and GCC lost out big time to that, but that was probably around 1995/6
and I haven't done that test again since then.

My attempts to benchmark on 88k failed because the GCC optimiser was
simply too broken to produce usable code on 88k. In any case, I had
nothing to compare with as the two 88k system vendors I knew, Motorola
and DG, both shipped (broken) GCC as their native compiler anyway.
(Maybe no one ever wrote any other C compiler for the 88k?)
Again, this is going back to, probably, 1998.

-- 
Andrew Gabriel
Consultant Software Engineer

>
> So an IDE speeds this up in what way?
>

None at all,
at least for the time the machine works, but makes you feel comfortable.

The programmer's debugging  _can_  be faster with a better overview.
But I've seen some programmers (coming from the W$dooze side)
spending more time with debugging than thinking before make.
You could not blame the IDE for such behaviour.

And to keep the flame cool: Eight Megabytes Always Continous Swapping has
lost much of it's horror on machines with gigs of RAM. *NIX offers a free
choice of editors to everyones personal flavour.

---
Uli

On Tue, 29 Jul 2003 00:54:16 +0100, Dr. David Kirkby <drkirkby@ntlworld.com>
 wrote in <3F25B7A8.AAF43DB6@ntlworld.com>:

> There have been a number of companies advertising on UK television,
> offering to take on trivial claims on a 'no-win no-fee' basis.
> Apparently those cases that were won, the people involved often got
> very little. I'm glad to see several such companies have gone bust. 

	That was because of a change in the rules for legal aid for
private lawsuits though, wasn't it?  The Government (before my time, so I
don't know which one) allowed the concept by allowing insurance companies
to take the risk of the judgement outcome: lose the case and the insurance
pays costs; win the case and the cost of insurance is added to the
damages.  I think it's the insurance companies which have gone bust,
dragging the ambulance-chasers down with them.

-- 
Ivan Reid, Electronic & Computer Engineering,     ___     CMS  Collaboration,
Brunel University.     Ivan.Reid@brunel.ac.uk             Room 40-1-B12, CERN
        KotPT -- "for stupidity above and beyond the call of duty".

Dave Uhring wrote:
> 
> On Tue, 29 Jul 2003 02:42:55 +0100, Dr. David Kirkby wrote:
> 
> > That's not true entirely true. I recently went into several PC shops
> > with a list of network cards on the Solaris HCL. I did not care what
> > card it was, as long as it was on the Solaris HCL. I could not find a
> > single network card in any shop that was on the Solaris x86 HCL. (I
> > know you could quite rightly argue I'd have a hard time picking up a
> > Sun keyboard/mouse locally.)
> 
> Are you saying that there are no shops selling off-lease business machines
> in a metro area as large as London?


I live in Essex and visited several local PC shops selling new
components to find a network card on the HCL. This was only necessary
because the ethernet connection on the motherboard was not seen by
Solaris. So one might conclude the motherboard was not on the HCL. 

After the PC shops local to me in Essex failed to have a suitable
network card, I visited a very large (supermarket sized) PC shop (PC
World), near where I work. Again I was unable to obtain a network card
on the Sun HCL there.

There are shops in London selling used PCs, but I admit I did not got
to them looking for a network card. With hindsight I should have done.
There is one close to where I work, but realistically had they not
have one (which they might well have done), I did not want to trek
half-way around London to find one.

I eventually ordered one online, which did the trick.

It seems the Solaris x86 HCL contains mainly the big-make, more
expensive components. I don't blame Sun, for that, because they are
more likely to be available, whereas manufacturers of the cheaper
cards seem to be forever changing. The result is that components on
the HCL cost more than the cheapest components available. The PC
market is very dominated by price, so PC shops don't tend to stock
cards from 3COM, Intel etc, but instead have the cheaper, often
unknown (to me anyway) manufacturers. 

The HCL contains complete computers too. Again, these are not the
cheapest PCs available. Those cheap PCs manufacturers will use the
cheapest motherboard/network/hard-disk/ram ... etc they can get away
with, so Sun can't put such machines on the HCL.

My own personal view, on the difficulty of assembling PCs and getting
them working with their respective operating systems is:

Solaris (x86) > Linux > Windoze. 

If I was to include Sun workstations in addition to PCs, I'd be
tempted to say:

Solaris (x86) > Linux > Windoze > Solaris (SPARC).

others might well disagree of course. I've only ever installed Solaris
x86 a couple of times, so this might not a true reflection in
practice. 
-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

On Tue, 29 Jul 2003, Alexis Cousein wrote:

> Andrew Gabriel wrote:

> > I benchmarked it against Sun's compilers probably every couple of
> > years since 1995,

> A couple of years is a lot in recent gcc history ;).

Yes Alexis, it is, but MIPSpro and SUNWspro still produce smaller and
faster binaries on Irix and Solaris for me than GCC 3.x on either.

Jamie Bowden
-- 
"It was half way to Rivendell when the drugs began to take hold"
Hunter S Tolkien "Fear and Loathing in Barad Dur"
Iain Bowen <alaric@alaric.org.uk>

In article <3f258626_1@news.arcor-ip.de>,
	"Thomas Dehn" <thomas-usenet@arcor.de> writes:
> 
> "Bill Gunshannon" <bill@cs.uofs.edu> wrote:
>> In article <3F2368E5.ED5037A0@hotmail.com>,
>> Anthony Mandic <nu@hotmail.com> writes:
>> > "Dr. David Kirkby" wrote:
>> >> 
>> >> As someone who works in a University, I am amazed UNIX vendors don't
>> >> give away free copies of their operating systems to students and home
>> >> users.
>> > 
>> > Why? Do Microsoft do this? Perhaps you should widen your
>> > job experience.
>> 
>> Actually, yes, they do.  We (the CS Department) subscribe to a service
>> with MS that allows us to give copies and licenses to our students for
>> OSes and development tools and they can take them with them when they
>> leave school.
> 
> And how much do "you" (the CS department, resp. the whole
> university) pay to M$ per year in total? Not just for that
> "service", but for everything.
> 

Other than this service, we (the CS Department) also pay to license
MS Office for our labs and Faculty.  I don't know the amount, as I
don't sign the checks, but it is standard educational pricing.  What
the rest of the University does is not my problem and I can't say
how much they pay.  But, after a short evaluation, the cost of the
MS educationl program for OSes & Development Software was found to
be not only resonable but also within our budget constraints which
are very tight.  When you figure that we build most of our own machines
and so they don't come with any OS (and thus no MS tax like on a DELL
or Gateway, regardless of the OS you run on it) this is even more of
a bargain.

bill

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
bill@cs.scranton.edu     |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>   

On Mon, 28 Jul 2003 15:28:54 +0200, Uli Link
<Ulrich--nO---sPAM--Link@Epost.de> wrote:
>Don't agree: the optimization of the GCC has become much better:
>in 9 out of 10 times the GCC produces faster binaries than vendor compilers

gcc only does well with the current Linux flavor-of-the-week favorite
processor. Its Alpha code generator sucks rocks.

Benjamin Gawert wrote:

> Don't You think that Sun isn't aware
> that bunches of hobbyists download and install Solaris on their old machines
> even when the licensing rules forbid that? Do You really believe HP, IBM and
> SGI don't know that hobbyists copy and trade the OS CD's or buy/sell them on
> ebay and install the OS on machines that are not licensed for these OS
> versions? Except SGI Germany which shows some (illegal) efforts in
> attempting to press people selling IRIX media kits on ebay.de to remove
> their auctions the Unix vendors don't give a sh*t on what hobbyists do, as
> long as they don't make money with the illegal OS installations. On the
> other hand, they don't provide hobbyist programs since such programs only
> cause costs without bringing any noticeable renevue.

I don't believe the costs involved in allowing Solaris to be
***legally*** used by hobbyyists would be high. It can be downloaded
for no charge now, so what extra would it cost to change the licence
to allow its legal use? Do you think the costs involved would be more
or less than the benifit Sun have gained from tools written by others
and made freely available, such as sendmail, openssh etc? 



-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

It is never a bargain to give a monopolist more mindshare and
leverage in the marketplace.  Ultimately you, your students and
the rest of us ALWAYS wind up paying for Mr Bill's broken
warez.

It is appropriate to teach students about MS... as a negative
example.  It is also important for them to understand the guts
of an OS.  They aren't supposed to use "Wizards" they are
expected to BE Wizards.

BJ

Bill Gunshannon wrote:
> In article <3f258626_1@news.arcor-ip.de>,
> 	"Thomas Dehn" <thomas-usenet@arcor.de> writes:
> 
>>"Bill Gunshannon" <bill@cs.uofs.edu> wrote:
>>
>>>In article <3F2368E5.ED5037A0@hotmail.com>,
>>>Anthony Mandic <nu@hotmail.com> writes:
>>>
>>>>"Dr. David Kirkby" wrote:
>>>>
>>>>>As someone who works in a University, I am amazed UNIX vendors don't
>>>>>give away free copies of their operating systems to students and home
>>>>>users.
>>>>
>>>>Why? Do Microsoft do this? Perhaps you should widen your
>>>>job experience.
>>>
>>>Actually, yes, they do.  We (the CS Department) subscribe to a service
>>>with MS that allows us to give copies and licenses to our students for
>>>OSes and development tools and they can take them with them when they
>>>leave school.
>>
>>And how much do "you" (the CS department, resp. the whole
>>university) pay to M$ per year in total? Not just for that
>>"service", but for everything.
>>
> 
> 
> Other than this service, we (the CS Department) also pay to license
> MS Office for our labs and Faculty.  I don't know the amount, as I
> don't sign the checks, but it is standard educational pricing.  What
> the rest of the University does is not my problem and I can't say
> how much they pay.  But, after a short evaluation, the cost of the
> MS educationl program for OSes & Development Software was found to
> be not only resonable but also within our budget constraints which
> are very tight.  When you figure that we build most of our own machines
> and so they don't come with any OS (and thus no MS tax like on a DELL
> or Gateway, regardless of the OS you run on it) this is even more of
> a bargain.
> 
> bill
> 

In article <3F246BDF.815C3B88@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
>suppose there are than many home users buying Sun compilers at full

I suspect your wrong. I see SCCIS-800-T9UM for $400 @ SunStore.

I'm not sure how that price compares to the C bundles from IBM, HPQ
or SGI.
 
John
groenveld@acm.org

John D Groenveld wrote:
> 
> In article <3F246BDF.815C3B88@ntlworld.com>,
> Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> >suppose there are than many home users buying Sun compilers at full
> 
> I suspect your wrong. I see SCCIS-800-T9UM for $400 @ SunStore.
> 
> I'm not sure how that price compares to the C bundles from IBM, HPQ
> or SGI.
> 
> John
> groenveld@acm.org

The C++ compiler for AIX costs between $2200 and $2500 (round numbers)
for a single user licence, depending if downloaded or bought on CD. 

http://www-132.ibm.com/webapp/wcs/stores/servlet/CategoryDisplay?lang=en_US&catalogId=-840&categoryId=1826260&cgmenbr=1&dualCurrId=73&langId=-1&cgrfnbr=1826260&y=11&x=10&cntrfnbr=1&storeId=1&cntry=840&scrfnbr=73

(just C, without C++ is cheaper, but I don't think it includes all the
tools of the C++ version).

I think you would have to agree that is just not acceptable to most
home users. Add that to the cost of the AIX operating system, and you
are looking at serious $$$$. The hardware is quite cheap - I put
together a machine with 4x332 MHz CPUs and 1 Gb of RAM for about �1000
($1600) including shipping it over from the USA to the UK. Due to its
weight, I had to use Fedex (USPS would not handle it), so shipping
charges were a fair chunk of that $1600. 

I've not played with them much, but the AIX development tools look
very impressive. The level of warnings given by the compiler is
excellent, with them clearly separated into the trivial (failing to
use the value returned by prinf) to the more serious. From my very
limited experience of it (used for about an hour), the debugger looks
very good too. (I only used it for about an hour, since I'd found the
two bugs with it during that time). 

IBM have an excellent university scheme:
http://www-3.ibm.com/software/info/university/
where university staff can obtain AIX, compilers, as well as a lot of
other stuff for zero cost - for non commercial use of course. This
includes free technical support. 

It would be nice if that was extended to home users (without technical
support of course). 

Clearly IBM see the benefits of working with universities - its a
shame HP don't with HP-UX and the tools for development under HP-UX.
I'm really impressed with the HP hardware, and its performance with
gcc is excellent. I'd expect it would be better with HP's compiler,
but with no economical way of obtaining the compiler, I doubt I'll
ever know.

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Hi 

This is what I'm paying for Operating Systems...


HPUX - they sent me a free cdr copy off 11. The program was called
easy-setup

Solaris x86 9 was in the low $20 range
Solaris Sparc 9 was a free download

Aix 4.2.1 was about $6.00 off of ebay

irix53.iso       - google (enough said)

OpenVMS 7.1 was about $25 I think (hobbyist)

Windows - msdnaa $0 (bought what its worth too)

Linux $0 (best deal ever)

SCO Unixware 7.11 was about $35

On Tue, 29 Jul 2003 14:27:41 -0400, David Michaels <david.michaels@uc.edu>
wrote:
>HPUX - they sent me a free cdr copy off 11. The program was called
>easy-setup

Oh, really? I've got a C110 that's sitting idle because I can't get 10.20 to
install on it (keeps asking for the attached keyboard type, even when
running on the serial console, and then hangs), and would like to get a
legitimate copy of something that will install.

In article <bg06es$o27$1@news.cs.tu-berlin.de>,
Joerg Schilling <js@cs.tu-berlin.de> wrote:
>Which is the real problem.

Microsoft has a better marketing team than Sun or at least has
an upper management willing to devote more resources to the
task.

Microsoft's Academic Alliances <URL:http://www.msdnaa.net/> gets
heavily promoted, but Sun's EduSoft and other initiatives gets
word-of-mouth. :(

John
groenveld@acm.org

In article <9MUUa.69854$Pn5.19467@news-lhr.blueyonder.co.uk>,
Yuan <NoSpam@nospam.com> wrote:
>HP-UX, Irix and Solaris have no student licensing program.

Sun provides discounts for Solaris as well several other software products
students to "free as in free beer."
<URL:http://www.sun.com/products-n-solutions/edu/promotions/edusoft/>

>Can Solaris on a x86 runs as well as on a Sparc machine?

Yes.

John
groenveld@acm.org

In article <bg5pkl$kurhq$3@ID-135708.news.uni-berlin.de>,
Bill Gunshannon <bill@cs.uofs.edu> wrote:
>MS Office for our labs and Faculty.  I don't know the amount, as I
>don't sign the checks, but it is standard educational pricing.  What
>the rest of the University does is not my problem and I can't say
>how much they pay.  But, after a short evaluation, the cost of the

If you can track down the PO, it might reference one or more of
these programs.

Microsoft Campus Agreement
<URL:http://www.microsoft.com/education/?ID=CampusAgreement>

Microsoft Student Select
<URL:http://www.microsoft.com/education/default.asp?ID=StudentSelect>

MSDN Academic Alliance
<URL:http://www.msdnaa.net/>

John
groenveld@acm.org

In article <3F23BE8A.C11571FD@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
>Sun gave us StarOffice,  there would be no inclination to switch to
>it, since Windoze and all the MS products cost us next to nothing.  

How much is "next to nothing" for both end-user and the institution?
Here Office XP licenses for department use is $68.82 and for student personal
use is $28.04 plus whatever the institution has to pony up behind the
scenes.

StarOffice is free AFAICT.

John
groenveld@acm.org

In article <slrnbicrn6.cbl.jmaynard@thebrain.conmicro.cx>,
Jay Maynard <jmaynard@conmicro.cx> wrote:
>On Mon, 28 Jul 2003 15:28:54 +0200, Uli Link
><Ulrich--nO---sPAM--Link@Epost.de> wrote:
>>Don't agree: the optimization of the GCC has become much better:
>>in 9 out of 10 times the GCC produces faster binaries than vendor compilers
>
>gcc only does well with the current Linux flavor-of-the-week favorite
>processor. Its Alpha code generator sucks rocks.


Not true for Sun's compiler (sparc _and_ x86) and also not true for the SCO
Intel C-compiler.

What compilers are you talking about? Or do you just wrote somtehing like:

'Oh Linux does not have this, in the future Linux will have it because somebody 
will write it...'.

-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

John D Groenveld wrote:
> 
> In article <3F23BE8A.C11571FD@ntlworld.com>,
> Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> >Sun gave us StarOffice,  there would be no inclination to switch to
> >it, since Windoze and all the MS products cost us next to nothing.
> 
> How much is "next to nothing" for both end-user and the institution?
> Here Office XP licenses for department use is $68.82 and for student personal
> use is $28.04 plus whatever the institution has to pony up behind the
> scenes.
> 
> StarOffice is free AFAICT.
> 
> John
> groenveld@acm.org

I was wrong to use the term 'next to nothing'. The truth is I don't
know what it costs. I've never heard anyone moan about the cost and
consider using StarOffice instead of Office 2000. But I don't work in
the IT department, or purchasing, so I'm not informed about these
matters. 

I don't think MS are stupid enough to do what Wolfram Research did,
which was to hick the price of a university site licence so much, that
99% of the college switched to a competitive product. Clearly (well to
me, but not everyone it seems), it's in Microsoft's interest for all
our graduates to leave knowing Windoze+Office 2000, rather than
StarOffice or OpenOffice. Just like I feel it's in Sun's interest for
our students to know Solaris, and HP's interest for our students to
know HP-UX etc etc. No matter what happens, may will leave knowing
Linux. 


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dr. David Kirkby wrote:

> I don't believe the costs involved in allowing Solaris to be
> ***legally*** used by hobbyyists would be high. It can be downloaded
> for no charge now, so what extra would it cost to change the licence
> to allow its legal use? Do you think the costs involved would be more
> or less than the benifit Sun have gained from tools written by others
> and made freely available, such as sendmail, openssh etc?

Yes, the costs would be much higher than the benefit. For You it's of course
just a change of a few words of the license text, but for real there are
much more efforts necessary which lead to a lot of costs. Not to forget the
stuff within Solaris that Sun doesn't own but has just licensed by the
copyright holders. Such license contracts are quite strict and don't allow
Sun to make the licensed parts available. If they want to do that You can be
sure that most of the partys Sun is licensing stuff from don't change their
licensing scheme for free...

It's not always as simple as it might seem on the first look...

Benjamin

Lord Isildur <isildur@andrew.cmu.edu> wrote in message news:<Pine.GSO.4.55L-032.0307281057490.1636@unix1.andrew.cmu.edu>...
> Not to start a flamewar, but... don't ignore the BSD's. 'best' is a very
> hard to define term.. personally i consider NetBSD or OpenBSD to be 'best'
> and Linux to be very suboptimal.. 

I agree about the BSD's.  They should not be ignored.  There are a lot
of things to recommend about them.  (High Security, stability are
among the top ones for me.)

I just like the Linux side of things right now because of the
mindshare and development.  There is lots going on.  From a sysadmin
point of view, I like BSD.  Its very easy to just build a minimal
system and know it will do the job.

Your opinion about 'best' is valid.  Damn messy value judgements have
such high potential for trouble.

How about:

A solid free UNIX would be Linux ;)

Benjamin Gawert wrote:
> Dr. David Kirkby wrote:
> 
> 
>>I don't believe the costs involved in allowing Solaris to be
>>***legally*** used by hobbyyists would be high. It can be downloaded
>>for no charge now, so what extra would it cost to change the licence
>>to allow its legal use? Do you think the costs involved would be more
>>or less than the benifit Sun have gained from tools written by others
>>and made freely available, such as sendmail, openssh etc?
> 
> 
> Yes, the costs would be much higher than the benefit. For You it's of course
> just a change of a few words of the license text, but for real there are
> much more efforts necessary which lead to a lot of costs. Not to forget the
> stuff within Solaris that Sun doesn't own but has just licensed by the
> copyright holders. Such license contracts are quite strict and don't allow
> Sun to make the licensed parts available. If they want to do that You can be
> sure that most of the partys Sun is licensing stuff from don't change their
> licensing scheme for free...
> 
> It's not always as simple as it might seem on the first look...

But you could argue this in the opposite direction as well. Historically there has been a 
symbiotic relationship between Unix and Universities (SUN = Stanford University Network) 
and every Unix is distributed with piles of free, open source and public domain software. 
In fact this trend still continues: I don't think there is a Unix variant today that does 
not include perl as part of the distribution, and most have come to rely on it (i.e. it is 
not just a freebie, but an integral part of the environment that diferentiates their 
product from their competition).

I have sympathy for both sides of the argument. There is a cost to allow free/cheap use of 
Unix at home, but also a benefit. On balance, I would like to see more done to encourage 
use of all Unix variants at home.

There definitely is an advantage to the Unix producers in this. Speaking for myself (and 
I'm not unique), I have been instrumental in several purchases of various sizes, mainly 
from Sun, that might have gone to M$ if it had not been for my involvement. This 
represents additional income many times the value of licenses that I use.

Then there are the companies (I know of one that did this) who threw Unix out the door 
because there was nobody on site with adequate knowledge to solve some recurring problems. 
This represents business (and potentially a continuing future stream of business) that has 
been lost and difficult/expensive/impossible to recover. This might have been prevented if 
there was more familiarity with Unix amongst the staff.


-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GO/! d- s++:+ a+ C++(++++) US++++$ UB++ U*++ P+++ L+++ E--- W+++ N++ w--- O-
M+ V- PS+ PE+ Y+ PGP t+ 5++ X R* tv+ b+ DI++ D G e(*) h++/-- r+++ y?
------END GEEK CODE BLOCK------

-----------------------------------------------------
Bob Hoekstra: APL & Unix Consultant
Telephone:    +44 1483 771028
Mobile:       +44 7710 562345
Email:        Bob.Hoekstra@HoekstraSystems.ltd.uk
-----------------------------------------------------

Gilles Gravier wrote:
> 
> Free Solaris?
> 
> If you are a developer, have a look at : http://sdc.sun.com/solaris/
> 
> If you are a university member (student, teacher or whatever), have a
> look at : http://wwws.sun.com/software/solaris/binaries/
> 
> Free Solaris. Sure. Very sure!

It's nice to see a response from someone at Sun on this matter - thank
you Gilles. 

To quote from 
http://wwws.sun.com/software/solaris/binaries/

"For only the cost of media plus shipping or download related costs,
you can use the software for non-commercial usage on single processor
systems supplied to you by Sun or its authorized distributors or based
on the x86 architecture."

Note those three words **supplied to you** 

So if I wish to use a SPARC based computer legally, that machine must
have been **supplied to me** by Sun or a Sun authorised distributor.
Most home users are going to get Suns from skips, free from work,
cheap on eBay or wherever. They are *not* going to have the computer
supplied to them by Sun or a Sun authorised distributor. Commerical
users buy from such places, not home users. 

> Of course, we don't want that to be used on very high end scalable
> machines... for these systems, you'll usually want support, and a full
> license.

Sun made several very low-end (by the standards of today) scalable
multi-processor computers. The SPARCstation 10 and 20 fall into this
category. There is no way in the year 2003 that one could consider the
SPARCstation 10 or 20 "very high end scalable machines", but they are
still excluded from this license, since they can take more than one
CPU. 

I have 6 Suns at home - five SPARCstation 20's (fitted with one or two
CPUs each) and an Ultra 80, fitted with 4 CPUs. Every machine I have
is excluded from this free license for two reasons.

a) None were supplied to me by Sun or a Sun authorised distributor.
b) Each is capable of holding more than one CPU.

I have previously owned a number of single processor machines (IPC,
SPARC 4, SPARC 5, Ultra 30), but each would be excluded since I did
not get them from Sun or a Sun authorised distributor. 

Over the years I have owned about a dozen Sun worksations and used
them all for non-commerical use at home. Not a single one would fall
within the terms of the current 'free' license. 

> But to evaluate, to play around with, to use at home, or at school, or
> to develop on, this is very sufficient.
> 
> Gilles.

It is *not* sufficient - it is totally useless.

That 'free' licence has two conditions that make it useless for
non-commercial, home, student use. It needs changing in two main ways.

1) The restriction that the machine is supplied to the individual by
Sun or a Sun authorised reseller needs removing. 

2) The restriction about a machine being capable of holding only 1 CPU
needs removing. The Solaris 8 licence allowed up to 8 CPUs, which I
personally would have thought acceptable to all. Even 4 would be
acceptable to 99.9% of home users. There is one home-user here
mentioning he had 6 CPUs. But more than 4 is very rare. 


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

In article <1r8yqgynej.fsf@glesvat.ifi.uio.no>,
Kjetil Torgrim Homme  <kjetilho@yksi.ifi.uio.no> wrote:
>hmm, I hope they didn't give the students the impression it was free,
>that's just dishonest.

Yes, unfortunately, perception is reality.
<URL:http://www.collegian.psu.edu/archive/2002/07/07-15-02tdc/07-15-02dnews-06.asp>
And since the program covered departments and faculty/staff, they all
operated under misimpression that is was free as in free beer.

Microsoft's marketing here has been exceptional.

John
groenveld@acm.org

In article <3F251918.B861101C@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> 1) In the case of games manufacturers, they would indeed loose revenue
> if they allowed everyone to copy their games. I don't think this is so
> with UNIX operating systems and compilers. There are not than many
> home users going to pay  more than $1000 on a compiler. They will
> likely use gcc instead and have to accept the poorer performance. 

And write software that only compiles on GCC, so that the commercial
vendors own users end up having to install GCC just to build and run
open-source software.

GCC is a fine compiler, but it's becoming a monoculture for software
developers... it's almost like GCC and a collection of related tools
are turning into a "linux emulator" for UNIX systems. So... where the
cry used to be "all the world's a VAX", it's now "all the world's Red
Hat".

I guess that's OK, if you like a monoculture.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

In article <PtBK/oHpv+sa092yn@visi.com>,
Richard Steiner <rsteiner@visi.com> wrote:
>If one's product is preloaded everywhere by default, one doesn't need a
>marketing team at all.

From where I sit, the big win for Microsoft with including the then
current Office 2000 in the "free" program was to encourage upgrades
from 97 and migrations from WordPerfect.

By including W2K, they encouraged Win98/ME and NT 4.0 users to upgrade.
I upgraded my laptop running WinME (and Solaris) to W2K under the program.  

The program encouraged my unit to centralize and more accurately
inventory its systems, which I guess will lead to greater adoptions
of Windows and Office 2003.

Just because one or more of Microsoft's products is preinstalled on
lots of systems, doesn't mean that their marketing team can emulate Sun's
performance.
Even maintaining the monopoly requires investment in marketing. 

John
groenveld@acm.org

> one platform. For real big investments You normally get evaluation systems
> for a certain period for free from the different Unix vendors to help You to
> evaluate if a platform is suitable for Your needs. Hobbyist experience
> provides exactly nothing to platform decisions, especially not the bigger
> ones which include a bit more than one or two workstations...

1. You obviously don't work in a university.

2. Hobbyist experience can do a hell of a lot for smaller installations,
as I've seen firsthand.  These add up after a while.  How the hell do you
suppose Linux got so big?  It's the ultimate hobbyist OS.

3. I see personal preference come into play all the time.  This isn't
necessarily a good thing (although I make no apologies for arguing against
Windows under any circumstances), but it's tolerated quite a bit.

4. Everyone has to start somewhere.  I worked a steady succession of jobs
starting as a low-grade technician and (literally) box mover, and spent
enough time tinkering around that I had a real job waiting for me after
graduating.

> Which is the best way to catapult itself out of the job flow. If You have to
> handle and administer Unix systems the worst thing You can do is sticking to
> a single Unix and/or platform.

That doesn't seem to stop many people.  I know plenty of people who buy

1. SGIs only because that's what they're used to and they don't realize
that Linux will probably (in their case) do a better job

2. Linux only, the reverse of (1)

3. Solaris only, see (1) and (2)

I once had an application I was working on moved from a Linux machine to a
Sun because the sysadmins liked Solaris.  The Sun was slower and caused
several things to break - we were all much happier with the Linux box.
I like Sun, but they've done very well getting people to buy Sun boxes
just because they're "standard".

One need look no further than the sickening dominance of Windows in IT to
see that homogeneity is embraced, experience in other platforms is
often useless, and technical considerations are not thought out very well.

Nat Echols wrote:

> 1. You obviously don't work in a university.

No, I'm working for a not really small company with several international
subsidiaries and a very good relationship to Sun, HP and IBM. We regulary
spend a very nice amount of $$$ for hardware which helps us to maintain our
buisness and otherwise brings money to the hardware manufacturer we're buing
from. And I can assure You, that not one of our IT decision was even barely
influenced by some hobbyist experience...

> 2. Hobbyist experience can do a hell of a lot for smaller
> installations,
> as I've seen firsthand.

For smaller installations maybe. But what renevue do You think such "smaller
installations" bring to Sun, HP and Co:? Nothing...

> These add up after a while.  How the hell do
> you suppose Linux got so big?  It's the ultimate hobbyist OS.

Linux grows simply because it's free and runs on comodity hardware which
performance-wise kicks ass even much more expensive proprietary systems from
all the vendors offering proprietary unix platforms around. Linux is a
competitor to the commercial Unixes, and due to it's origins (freeware based
development, GPL licensing scheme) can be shared with everyone who wants it
at no licensing costs. Wanting the same from a commercial Unix is stupid
since unlike Linux commercial Unixes regularly not only contain some really
big investments from side of the unix vendor to develop and maintain the
platform but also contains software that the unix vendors have to license
itself from 3rd parties. All this are quite huge inital and even running
expenses that have to get paid to keep the unix vendor on the market. And
the money necessary to keep them alive comes from big installations and
customer that not only buy software and maybe the necessary hardware but
from customers buying service contracts.

> 3. I see personal preference come into play all the time.  This isn't
> necessarily a good thing (although I make no apologies for arguing
> against Windows under any circumstances), but it's tolerated quite a
> bit.

Maybe tolerated, but when it comes to important things platform decision for
a medium size or really big company there is not much room for personal
preferences. The only thing what counts are cost- and risk-calculations
including several ways to solve a computing problem, mostly including much
more than just one or two platforms. If You're going to spend several
million dollars on IT equipment You rather avoid on listening on what people
are playng at home with, in the business segment it's just numbers that
count.

> 4. Everyone has to start somewhere.  I worked a steady succession of
> jobs starting as a low-grade technician and (literally) box mover,
> and spent enough time tinkering around that I had a real job waiting
> for me after graduating.

Sure, and for learning at home a unix box is a perfect way. But if You
manage to climb the ladder up to a position where You are responsible for IT
investments of some real amount (not just a few PC's) You certainly will
have learned that You won't do Yourself a favour on buying <insert
commercial unix vendor here> because You liked it when playing with this
platform at home, You certainly will also settle on cost/risk analysis to
evaluate the best solution for Your emplyee's computing needs, otherwiese
You risk to make the wrong decision and maybe make room for someone else
that cares more for facts than for feelings when spending other people's
money...

>> Which is the best way to catapult itself out of the job flow. If You
>> have to handle and administer Unix systems the worst thing You can
>> do is sticking to a single Unix and/or platform.
>
> That doesn't seem to stop many people.

That's right, but most of them that do it that way do it only once, and
after that they can look around for a new job, preferrably nothing that has
something to do with computers...

> One need look no further than the sickening dominance of Windows in
> IT to
> see that homogeneity is embraced, experience in other platforms is
> often useless, and technical considerations are not thought out very
> well.

That might certainly be true for smaller businesses, but who cares as this
isn't the market that the traditional unix vendors aim at. They usually go
only for either large installations with a lot of workstations and bigger
servers or at the HPC area. That's where the valuable Solaris, HP-UX, IRIX
and AIX customers are, and that's where most of the renevue comes from. And
really, that's where no-one with at least a minimum interest on keeping his
job would make any decisions based on the experience he had with an old Unix
workstation at home...

I don't understand where the problem is. Except SGI Germany no-one of the
unix vendors cares about what we hobbyists do with their machines and OSes,
as long as You don't make money with it and just play around or learn they
give a sh*t on that. Hobbyists programs would require investments, and
especially at the current oekonomical situation every company thinks twice
about where to invest money. Up til now the Hobbyists didn't prove that they
are a benefit for the vendor, even when Solaris was free for everyone Sun
didn't gain anything from it, so what the hell are You expecting? With
today's ways of getting software it should be possible for every hobbyist to
get an OS for his "new" old box, and the Unix vendors don't care what
version of their Unix You're running on Your old box...

Some time ago after a free-IRIX-poll in comp.sys.sgi.* I talked to an IBM
executive about the hobbyist thing. He said, even if someone would send them
a list of all hobbyists using AIX unlicensed on their old RS/6000 they
wouldn't do anything, simply because it wouldn't be worth the effort. If
people play with these systems fine, but they should not request any support
from IBM since they are not worth the effort. HP btw. seems to think similar
regarding HP-UX and Hobbyists, and the Tru64 program is just a relict from
Compaq...

Benjamin

Greg Menke wrote:
> 
> Akop Pogosian <akopps+usenet@ocf.berkeley.edu> writes:
> > It's legally limited to one CPU _AND_ , in case of SPARC systems, the
> > machine must have been sold by either Sun or an authorized
> > Sun-reseller as stated at the bottom of this page:
> > http://wwws.sun.com/software/solaris/licensing/policies.html
> >
> > This is why the Solaris/sparc
> > enthusiast community is outraged. 
> 
> I think "outrage" may overstate it a bit.  

> Personally, I'm glad they do offer a cheap media kit so small-time
> users can be plausibly legal.  I really can't begrudge them $50/per
> for a nice OS that still runs great on ~8 year old hardware.
> 
> Gregm

Gregm,
	You might not be outraged by the cost of buying a $50 media kit for
Solaris 9, but you are breaking the law if you instal Solaris 9 from
it. Check out 
http://store.sun.com/catalog/doc/BrowsePage.jhtml?cid=34895

and see the bits about "licenses for pre-owned Sun systems purchased
from third parties"

You can't legally use the free binary license, assuming you did not
buy those machines from Sun authorised distributors, but from eBay or
similar. It's doubtful your U5's or U2 ever legally had Solaris 9 on
them, since they were I believe all discontinued before Solaris 9 was
shipped. 

You need to visit:

http://store.sun.com/catalog/doc/BrowsePage.jhtml?cid=91371&parentId=82426

and sort out your licenses. If I'm not mistaken, in addition to the
$50 media kit you have already bought, you need to buy:

Description:  Solaris 9 Multiprocessor Desktop Upgrade RTU License, 
4 CPU Maximum, SPARC Platform Edition
Part Number:  SOLIS-090-G9U9 
$199

to licence your Ultra 2 - at even higher cost if the U2 is considered
a server. 

and you need to buy two 

Description:  Solaris 9 Desktop Upgrade
RTU License, 1 CPU
Maximum                                                        Part
Number:  SOLIS-090-D9U9 
$99 

to license your two Ultra 5's. Your $50 media kit does not (legally)
allow you to install the software on those machines. You don't need to
buy multiple media kits, but you do need to buy one. 

So to licence the machines, you now need to spend another $397, in
addition to that you have already paid out. So your media kit and
licences will cost you just under $450!! Do you mind that ? Would that
outrage you ?

I have five SPARCstation 20's here. Most (if not all) have 2 CPUs. If
I wanted to run Solaris 9 on any of them (which I don't), it would
cost me $199 for each machine. The last SPARCstation 20 I bought, with
2 CPUs and 320 Mb of RAM cost me sixty UK pounds, or about $100.
Another, with 4 CPUs (I've since split the CPUs up), cost me nothing.
So $199 for a license seems a bit OTT to me. 

The quad processor Ultra 80 I have was sold as a workstation by Sun,
not as a server. But I do share its disks with other machines, so I
guess you could say it's acting as a server. Whether or not I would
need a $199 4 CPU desktop license or a $999 4 CPU server licence to
legally run Solaris 9, I am not sure.

That's why people are outraged. Few would object to buying a $50 media
kit and being able to copy the OS to machines for non-commercial use.
But spending more on a license than what the hardware is costing (for
some machines at least), seems a bit much. 

If you, or anyone else, can point me to a way to legally licence a
home machine for Solaris 9, used non-commercially, at a sensible cost,
please let me know. 

Gilles, who works for Sun, has said on the newsgroup that
non-commerial use is covered by the free Solaris binary licence. But
my reading of the licenses would suggest this is not the case. Please
correct me if I am wrong. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Alain wrote:
> 
> Hi Doctor,
> 
> I think you must try to work in a real company. First, you will win enough
> money to buy a HP.

You don't need to be insulting. Given your attitude, I refuse to reply
any further. 

> Alain.
> 
> PS: sorry for my english.

Your English I don't mind, but your attitude I do. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

In article <3F28A32C.3BB87B91@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
>Gilles, who works for Sun, has said on the newsgroup that
>non-commerial use is covered by the free Solaris binary licence. But
>my reading of the licenses would suggest this is not the case. Please

Sun may intend to allow non-commercial home use on multi-cpu capable
systems with the Free Solaris Binary License, but their legalise does
not.

I don't know why Sun cannot specifically allow home use of multi-cpu
capable machines while protecting stockowners best interest by
preventing grey market procurred SF15Ks being used in production
with same.

My guess is that the Solaris product group is budgeted only
a very limited time with Sun's landsharks who could add a definition
for "noncommercial, home use" to the license and exempt it.

John
groenveld@acm.org

On Thu, 31 Jul 2003 06:03:40 +0100
"Dr. David Kirkby" <drkirkby@ntlworld.com> wrote:

> So to licence the machines, you now need to spend another $397, in
> addition to that you have already paid out. So your media kit and
> licences will cost you just under $450!! Do you mind that ?

It's not an awful lot of money, but one would have to
include it in the price of the system. It might not be
worth upgrading from a legally obtained Solaris 8
license, for example. It might also cause people not
to purchase older Sun systems, or motivate them to
run Linux on them instead of Solaris.
If the cost of the license tips the scales in favour
of a Sun Blade 150 rather than a used U60 on eBay, then
Sun scores (they're a hardware company, mainly). If,
on the other hand, the potential Sun user decides to
buy a PC or an Apple, Sun has lost nothing (mindshare
and goodwill don't mean anything to beancounters; 
the idea that people may be close to passionate about
hardware and software is utterly foreign to them).

> Would that outrage you ?

Most certainly not. I might decide (as a home user or
a student) that Sun isn't for me. I might decide to
run Solaris 9 illegally - after all, people ignore the
speed limits all the time.
As a professional software developer, the $199 license
for a workstation a lot. If I need Solaris 9 for my work,
I'd buy the license (like I buy Windows licenses, and
development systems licenses, and other software I need). 

There seems to be no place for computer companies run
by enthusiasts, like 25 years ago. Sun is one of the
few that shows some feeble remains of the spirit of
those days. Look at HP for a truly sad case of a company
once driven by technical excellence becoming a purely
marketing-driven re-badgetting behemoth. 

-- 
Stefaan
-- 
"What is stated clearly conceives easily."  -- Inspired sales droid

Benjamin Gawert wrote:

> Nat Echols wrote:
>
> > 1. You obviously don't work in a university.
>
> No, I'm working for a not really small company with several international
> subsidiaries and a very good relationship to Sun, HP and IBM. We regulary
> spend a very nice amount of $$$ for hardware which helps us to maintain our
> buisness and otherwise brings money to the hardware manufacturer we're buing
> from. And I can assure You, that not one of our IT decision was even barely
> influenced by some hobbyist experience...

I would be very surprised if none of your companies purchasing decisions were
influenced by by peoples personal preferences.

Initial purchase price is easy for accountants to look at on their own, but it
wouldn't be the first time a major system purchase wasn't influenced more
by what the FD (CFO)'s nephew just started at college's personal preference
is than anything the technical IT staff in the company felt.

Running cost of system are much more difficult to gauge. If the guy costing
how much time admin'ing the system will be, prefers brand X to brand Y they
will probably feel brand X costs less to look after, coz for them it does...
they
know it better... every thing is quicker.

>
>
> > 2. Hobbyist experience can do a hell of a lot for smaller
> > installations,
> > as I've seen firsthand.
>
> For smaller installations maybe. But what renevue do You think such "smaller
> installations" bring to Sun, HP and Co:? Nothing...
>

hat depends on how good the dealer chain is. Lots of small systems are worth
more than a few big ones. Hence Compaq could buy DEC.

>
> > These add up after a while.  How the hell do
> > you suppose Linux got so big?  It's the ultimate hobbyist OS.
>
> Linux grows simply because it's free and runs on comodity hardware which
> performance-wise kicks ass even much more expensive proprietary systems from
> all the vendors offering proprietary unix platforms around. Linux is a
> competitor to the commercial Unixes, and due to it's origins (freeware based
> development, GPL licensing scheme) can be shared with everyone who wants it
> at no licensing costs. Wanting the same from a commercial Unix is stupid
> since unlike Linux commercial Unixes regularly not only contain some really
> big investments from side of the unix vendor to develop and maintain the
> platform but also contains software that the unix vendors have to license
> itself from 3rd parties. All this are quite huge inital and even running
> expenses that have to get paid to keep the unix vendor on the market. And
> the money necessary to keep them alive comes from big installations and
> customer that not only buy software and maybe the necessary hardware but
> from customers buying service contracts.
>

Linux originally found it's way into lots of companies through the back door.
In many/most cases the management didn't even know it was there. I've
come across cases where an administrator has got fed up with re-installing
Windows for the n'th time over a weekend trying to solve a problem. Then
at 6 PM on Sunday evening made a decisions to move the box to Linux so
that they can just make it work and go home to have a life. The first time
management knew that the system wasn't running Windows was when it's
lack of down time stuck out like a saw thumb. And no this wasn't a small site
this was a major international company.

>
> > 3. I see personal preference come into play all the time.  This isn't
> > necessarily a good thing (although I make no apologies for arguing
> > against Windows under any circumstances), but it's tolerated quite a
> > bit.
>
> Maybe tolerated, but when it comes to important things platform decision for
> a medium size or really big company there is not much room for personal
> preferences. The only thing what counts are cost- and risk-calculations
> including several ways to solve a computing problem, mostly including much
> more than just one or two platforms. If You're going to spend several
> million dollars on IT equipment You rather avoid on listening on what people
> are playng at home with, in the business segment it's just numbers that
> count.
>
> > 4. Everyone has to start somewhere.  I worked a steady succession of
> > jobs starting as a low-grade technician and (literally) box mover,
> > and spent enough time tinkering around that I had a real job waiting
> > for me after graduating.
>
> Sure, and for learning at home a unix box is a perfect way. But if You
> manage to climb the ladder up to a position where You are responsible for IT
> investments of some real amount (not just a few PC's) You certainly will
> have learned that You won't do Yourself a favour on buying <insert
> commercial unix vendor here> because You liked it when playing with this
> platform at home, You certainly will also settle on cost/risk analysis to
> evaluate the best solution for Your emplyee's computing needs, otherwiese
> You risk to make the wrong decision and maybe make room for someone else
> that cares more for facts than for feelings when spending other people's
> money...
>
> >> Which is the best way to catapult itself out of the job flow. If You
> >> have to handle and administer Unix systems the worst thing You can
> >> do is sticking to a single Unix and/or platform.
> >
> > That doesn't seem to stop many people.
>
> That's right, but most of them that do it that way do it only once, and
> after that they can look around for a new job, preferrably nothing that has
> something to do with computers...
>
> > One need look no further than the sickening dominance of Windows in
> > IT to
> > see that homogeneity is embraced, experience in other platforms is
> > often useless, and technical considerations are not thought out very
> > well.
>
> That might certainly be true for smaller businesses, but who cares as this
> isn't the market that the traditional unix vendors aim at. They usually go
> only for either large installations with a lot of workstations and bigger
> servers or at the HPC area. That's where the valuable Solaris, HP-UX, IRIX
> and AIX customers are, and that's where most of the renevue comes from. And
> really, that's where no-one with at least a minimum interest on keeping his
> job would make any decisions based on the experience he had with an old Unix
> workstation at home...
>

If that experience teaches them that some things are easier on one box than
another, then they would be foolish in the extreme not to use that experience
at work as well.

>
> I don't understand where the problem is. Except SGI Germany no-one of the
> unix vendors cares about what we hobbyists do with their machines and OSes,
> as long as You don't make money with it and just play around or learn they
> give a sh*t on that. Hobbyists programs would require investments, and
> especially at the current oekonomical situation every company thinks twice
> about where to invest money. Up til now the Hobbyists didn't prove that they
> are a benefit for the vendor, even when Solaris was free for everyone Sun
> didn't gain anything from it, so what the hell are You expecting? With
> today's ways of getting software it should be possible for every hobbyist to
> get an OS for his "new" old box, and the Unix vendors don't care what
> version of their Unix You're running on Your old box...
>

In the case of HP, when the system was bought, an HP-UX license was bought
for it. All PA-Risc based systems from HP come with an HP-UX license, even
MPE systems have a Unix license because of some of the engineering diagnostics.

So as far as they are concerned someone paid for it at some point.

With IA64 based systems this might cease to be the case, I don't know.

But trying to sue end hobbyist would only be a case of throwing good money
after bad.

>
> Some time ago after a free-IRIX-poll in comp.sys.sgi.* I talked to an IBM
> executive about the hobbyist thing. He said, even if someone would send them
> a list of all hobbyists using AIX unlicensed on their old RS/6000 they
> wouldn't do anything, simply because it wouldn't be worth the effort. If
> people play with these systems fine, but they should not request any support
> from IBM since they are not worth the effort. HP btw. seems to think similar
> regarding HP-UX and Hobbyists, and the Tru64 program is just a relict from
> Compaq...
>
> Benjamin

Peter da Silva wrote:
> 
> In article <3F251918.B861101C@ntlworld.com>,
> Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> > 1) In the case of games manufacturers, they would indeed loose revenue
> > if they allowed everyone to copy their games. I don't think this is so
> > with UNIX operating systems and compilers. There are not than many
> > home users going to pay  more than $1000 on a compiler. They will
> > likely use gcc instead and have to accept the poorer performance.
> 
> And write software that only compiles on GCC, so that the commercial
> vendors own users end up having to install GCC just to build and run
> open-source software.
> 
> GCC is a fine compiler, but it's becoming a monoculture for software
> developers... it's almost like GCC and a collection of related tools
> are turning into a "linux emulator" for UNIX systems. So... where the
> cry used to be "all the world's a VAX", it's now "all the world's Red
> Hat".
> 
> I guess that's OK, if you like a monoculture.

I agree with you entirely there. As far as 90% of open-source
developers are concerned, if it compiles on gcc, it is okay. Few seem
to make any effort to check it with other compilers. 

Like it or not, gcc is now a standard. I don't think its a very good
standard in many ways, but it is no doubt a standard - at least
according to a dictionary definition of a standard
http://www.m-w.com/cgi-bin/dictionary?va=standard
before anyone jumps up and tells me gcc does not conform to this or
that ISO standard. At least one compiler (IBM's for AIX I think), has
switches to enable gcc extensions.

I think the code that gcc allows to pass without issuing a warning,
needs tightening up. -Wall is next to useless. A lot more could be
done to ensure that code is portable, before the -Wall option gives no
warnings. I think if the compiler warnings issued were much more
comprehensive, then people would tend to remove possible portability
problems, making the code more portable. As it is, people see that
their code compiles okay on gcc with no warnings and think it is good
code. In reality, gcc -Wall is far too lax. (You post has made me
decide to run gccbug and report this as a bug). 

I have here a number of test machines I use for checking my software.
I've put quite a bit of effort in to checking the portability of 
http://atlc.sourceforge.net/

Whenever possible I have tried to obtain the manufacturers compilers.
I have legal licences for Solaris, AIX, IRIX and Tru64 - all obtained
at zero or minimal cost under one scheme or the other. 

I've yet to get a copy of HP's compiler for HP-UX, so have no choice
but to install gcc on that. It's a shame really, as the HP PA-RISC box
is the best performing machine I have, so I expect it would do even
better with an HP compiler. I can't seem to find one on eBay. 


Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring wrote:

> Pretty tame.  Back in 1970 I drove across Nevada at 145 mph.  But the
> speed limit then was "reasonable and proper".

	There used to be no speed limit in the Northern Territory until
	someone wiped themselves out on a straight road in a Ferrari
	during a Cannonball Run event. A lot of people got peeved about
	that.

-am	� 2003

"Dr. David Kirkby" wrote:

> This is getting off the topic somewhat, so I'm a bit reluctant to
> continue this much further. However, the following points need making.

	Somewhat? You started this stupid off-topic thread. Now you'll
	have to live with it.

-am	� 2003

Lord Isildur wrote:
> 
> Lincoln, actually, was the one who put the final nail in the
> coffin of the Republic, and reaffirmed the nation's true character
> to be Empire. 'For the people, etc' was rhetoric. Jefferson was the
> true champion of the idealistic vision of america which for a while
> was so firmly believed be so many people that it almost came into
> existence. However, once matters got down to it.. all that was pushed
> to the side.

	Hmmm ... didn't Jefferson own slaves and Lincoln free them?

> ok, we're getting really offtopic now

	Its been that way since the start.

-am	� 2003

Anthony Mandic <pk@hotmail.com> writes:

> Dave Uhring wrote:
> 
> > Pretty tame.  Back in 1970 I drove across Nevada at 145 mph.  But the
> > speed limit then was "reasonable and proper".
> 
> 	There used to be no speed limit in the Northern Territory until
> 	someone wiped themselves out on a straight road in a Ferrari
> 	during a Cannonball Run event. A lot of people got peeved about
> 	that.

Got peeved about what? Crashed Ferrari, or speed limits?

Bye, Dragan

-- 
Dragan Cvetkovic, 

To be or not to be is true. G. Boole      No it isn't.  L. E. J. Brouwer

!!! Sender/From address is bogus. Use reply-to one !!!

Dragan Cvetkovic wrote:

> Got peeved about what? Crashed Ferrari, or speed limits?

	The subsequent introduction of speed limits. Only the Japanese
	millionaire who crashed would have gotten peeved about his car
	but he wasn't in any condition to be peeved about anything ever
	again. I think the moral there is that you don't need a speed
	limit to tell you to use common sense.

-am	� 2003

Dave Uhring <daveuhring@yahoo.com> writes:

> It's a shame about that Ferrari.  But imposing speed limits does not
> necessarily lead to fewer highway fatalities.  The US interstate highway
> system has 30% higher fatalities/mile_of_travel than the German Autobahn
> and the Autobahn has no speed limit at all outside of urban areas.

While I lived in Germany (1992-1997), most Autobahns I was using there did
have a speed limit (be it 100kmh or 130kmh). Even outside of urban
areas. Or if there was no speed limit it would be only for a short while
most of the time.

On the other hand, it is hard to find "outside of urban areas" in Germany
with the country being so dense populated :-)

Bye, Dragan

-- 
Dragan Cvetkovic, 

To be or not to be is true. G. Boole      No it isn't.  L. E. J. Brouwer

!!! Sender/From address is bogus. Use reply-to one !!!

Dave Uhring wrote:

> It's a shame about that Ferrari.  But imposing speed limits does not
> necessarily lead to fewer highway fatalities.  The US interstate highway
> system has 30% higher fatalities/mile_of_travel than the German Autobahn
> and the Autobahn has no speed limit at all outside of urban areas.

	Which tells us what? The standard of driving between the countries
	varies or the quality of vehicles varies or the road conditions are
	different? Anything else?

-am	� 2003

On Thu, 31 Jul 2003 11:54:32 -0400, Dragan Cvetkovic wrote:

> While I lived in Germany (1992-1997), most Autobahns I was using there did
> have a speed limit (be it 100kmh or 130kmh). Even outside of urban
> areas. Or if there was no speed limit it would be only for a short while
> most of the time.

Maybe that TV show about the Autobahn I saw last week was in error?  They
filmed a 525hp Porsche running at 212 mph on one of the roads and a Merc
passing a police van at 120 mph, not kph.
 
> On the other hand, it is hard to find "outside of urban areas" in Germany
> with the country being so dense populated :-)

That film showed a lot of rural countryside.  Perhaps they simply showed
only the one part of the highway with no speed limits.

On Fri, 01 Aug 2003 01:56:41 +1000, Anthony Mandic wrote:

> 	Which tells us what? The standard of driving between the countries
> 	varies or the quality of vehicles varies or the road conditions are
> 	different? Anything else?

Probably a combination of factors.  Most American drivers fail to pay heed
to the traffic conditions around them.  The Autobahn is probably better
maintained than most US highways and the German automobiles are designed
to be run at speed.  Certainly the Ferrari was designed to be driven at
high speed.

What happened to the Ferrari?  Did it hit a 'roo or wallaby, or perhaps a
pothole in the road?  We have a situation here with deer.  I've hit 2 of
them while driving in the wee hours of the morning.

Rich Teer wrote:

>> Not that this ever happens anywhere, mind you. Surely every company
>> has made a real cost/risk analysis on Windows as the general-purpose
>> OS too
>
> I doubt it.  If they did, they'd realise how expensive and risky
> running Windoze is, and start looking for an alternative.

Such general statements are nothing but false. Windows isn't always a worse
choice over some Unix operating system, it simply depends on the task and
the configuration...

Benjamin

On Thu, 31 Jul 2003 10:31:52 -0700, Benjamin Gawert wrote:

> Rich Teer wrote:

>> I doubt it.  If they did, they'd realise how expensive and risky
>> running Windoze is, and start looking for an alternative.
> 
> Such general statements are nothing but false. Windows isn't always a worse
> choice over some Unix operating system, it simply depends on the task and
> the configuration...

Granted that one cannot run such secure and capable apps as MSSQL Server,
IIS and MS Exchange Server on UNIX, please tell us just which other tasks
make Windows a better choice.

Bob Hoekstra <Bob.Hoekstra@HoekstraSystems.ltd.uk> writes:

> But you could argue this in the opposite direction as
> well. Historically there has been a symbiotic relationship between
> Unix and Universities (SUN = Stanford University Network) and every
> Unix is distributed with piles of free, open source and public
> domain software. In fact this trend still continues: I don't think

Having made the same mistake myself long ago;

SUN \=\ Sun The Sun of Sun Microcomputers is NOT in any way
related to Stanford Uni Networks.

-- 
Paul Repacholi                               1 Crescent Rd.,
+61 (08) 9257-1001                           Kalamunda.
                                             West Australia 6076
comp.os.vms,- The Older, Grumpier Slashdot
Raw, Cooked or Well-done, it's all half baked.
EPIC, The Architecture of the future, always has been, always will be.

On Thu, 31 Jul 2003 10:46:44 -0500, Dave Uhring <daveuhring@yahoo.com>
wrote:

>On Fri, 01 Aug 2003 01:19:51 +1000, Anthony Mandic wrote:
>
>> Dave Uhring wrote:
>> 
>>> Pretty tame.  Back in 1970 I drove across Nevada at 145 mph.  But the
>>> speed limit then was "reasonable and proper".
>> 
>> 	There used to be no speed limit in the Northern Territory until
>> 	someone wiped themselves out on a straight road in a Ferrari
>> 	during a Cannonball Run event. A lot of people got peeved about
>> 	that.
>
>It's a shame about that Ferrari.  But imposing speed limits does not
>necessarily lead to fewer highway fatalities.  The US interstate highway
>system has 30% higher fatalities/mile_of_travel than the German Autobahn
>and the Autobahn has no speed limit at all outside of urban areas.

Saw the numbers on a "Modern Marvels" episode recently and the numbers
are almost identical despite the speed difference.  30% is an extreme
exaggeration.  

The one thing that the Autobahn has that the Interstate system is
lacking is enforcement of the slower traffic keep right laws.



-- 
gburnore at DataBasix dot Com 
---------------------------------------------------------------------------
                    How you look depends on where you go.
---------------------------------------------------------------------------
Gary L. Burnore                       |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
                                      |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
Official .sig, Accept no substitutes. |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
                                      |  �� 0 1 7 2 3 / ݳ� 3 7 4 9 3 0 ۳
Black Helicopter Repair Services, Ltd.|     Official Proof of Purchase
===========================================================================

On Thu, 31 Jul 2003 11:22:16 -0500, Dave Uhring <daveuhring@yahoo.com>
wrote:

>On Thu, 31 Jul 2003 11:54:32 -0400, Dragan Cvetkovic wrote:
>
>> While I lived in Germany (1992-1997), most Autobahns I was using there did
>> have a speed limit (be it 100kmh or 130kmh). Even outside of urban
>> areas. Or if there was no speed limit it would be only for a short while
>> most of the time.
>
>Maybe that TV show about the Autobahn I saw last week was in error? 


If you were able to get 30% higher out of the statistics they
provided, it was likely you. :)

> They
>filmed a 525hp Porsche running at 212 mph on one of the roads and a Merc
>passing a police van at 120 mph, not kph.
>

That part was true.
 
>> On the other hand, it is hard to find "outside of urban areas" in Germany
>> with the country being so dense populated :-)
>
>That film showed a lot of rural countryside.  Perhaps they simply showed
>only the one part of the highway with no speed limits.

They didn't show any of the congested areas that's for sure. Nothing
near Darmstadt for example.
-- 
gburnore at DataBasix dot Com 
---------------------------------------------------------------------------
                    How you look depends on where you go.
---------------------------------------------------------------------------
Gary L. Burnore                       |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
                                      |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
Official .sig, Accept no substitutes. |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
                                      |  �� 0 1 7 2 3 / ݳ� 3 7 4 9 3 0 ۳
Black Helicopter Repair Services, Ltd.|     Official Proof of Purchase
===========================================================================

On Thu, 31 Jul 2003 14:04:02 -0400, Gary L. Burnore wrote:

> If you were able to get 30% higher out of the statistics they
> provided, it was likely you. :)

IIRC, the number for the Autobahn was 0.32 fatalities/100,000 miles while
the number for the US interstate was 0.42.

Perhaps you taped the show and can provide the exact numbers.

Scott Howard wrote:

> "Virtually all used Sun kit"?  Last I looked the vast majority of 2nd
> hand Solaris machines which are out there (and especially ones which are
> priced in the range which "students and home users" are likely to be
> buying) DO fit in the free license criteria. Sure, there are some
> dual-proc (or bigger) machines out there, but the majority are single
> proc machines.
> 
>   Scott

Scott,
	I think you are wrong to say that the majority of machines being sold
that are priced in the range of students and home users are single
processor machines. I think the majority are symmetical
multi-processor (SMP) machines. 

Of course, what an individal can afford is very much dependant on
their personal circumstances, but for me at least I can justify to me
(and my wife) buying any of the following Sun workstations. I'll
forget servers, since they are not typically what home users want. 

SPARC 4 and older - All too old to bother considering even for the
poorest student, so I'll forget all them. 
 
SPARC 5  
SPARC 10 (SMP)
SPARC 20 (SMP)
Ultra 1 
Ultra 2 (SMP)
Ultra 5
Ultra 10 
Ultra 30 
Ultra 60 (SMP)
Ultra 80 (SMP)
Blade 100 
Blade 150 

Out of those 12 models, 5 are multi-processor capable. These are
fairly evenly distributed in price range.

Of the dealers I know of, most seem to have more SS20's than anything
else, suggesting that it was the most popular machine. There's one on
eBay at his minute with a starting price of $9.95 and no bids with
under and hour to go. They are dirt-cheap. 

Nobody who knows is likely to publically state how many of each of Sun
boxes  were sold, but I suspect the SMP capable SS20 sold the most. 

I might not be typical home-user, but here is the data on the machines
I've *personally* own(ed). 

Of the 6 Suns I currently own, (5 x SS20 and 1 x U80), all are SMP. 

Of the 4 other UNIX boxes I currently own (HP C3000, Dec Alpha, IBM
RS/6000, SGI Octane), 2 are SMP.

So of 10 UNIX boxes I currenly own, 8 are SMP. 

If you own a SMP machine, the chances are you are going to put more
than one CPU in it. (The majority of SMP capable machines I own have
more than one CPU, although one or two of the SS20's don't). 

Of the 12 Suns I have owned at one time or another (1 x IPC, 1 x SPARC
4, 2 x SPARC 5, 5 x SPARC 20, 1 x U30, 1 x U60, 1 x U80), 7 (just over
50%) are SMP machines. 

(Perhaps any dealers that sell the above machines would like to
comment on the fraction of Sun workstations they have that are SMP
capable.)

Hence if Sun did ever considering allowing free home use of Solaris
again, that use would really need to include SMP machines as before.
People owning Sun SMP machines are probably not in a minority, but the
majority - at least that's what I think anyway! I may be wrong. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

more than just that, the roads themselves are engineered a lot better. Having
been on both, I have to say, most of the interstates in the US that i've driven on (this experience is admittedly restricted to the north-eastern quarter
of the country, i've not driven further west than chicago and michigan, nor
further south than a bit below DC) have terrible design. No banking of curves,
no attention paid to smooth merges on and off, and so on and so forth,
compared to the german roads, which are designed much more intelligently.

At the same time, I have to point out that statistics that link speed with
accidents are unavoidably contaminated by the prevailing speed limit on
the road being studied. The most dangerous factor in highway traffic, aside
from a freak event such as a mechanical failure or an insane or drunk driver,
comes from traffic not flowing at a uniform (or close to it) speed. If the
'natural' speed (what people drive irrespective of legal limits) on a road is,
say, 70 mph, but the limit is 55, then you get some mix of speeds. When these
interact, you get conditions much worse than if everyone was driving 70, or
if everyone was driving 55. On a road where the limit is 55, naturally
the people doing 'research' will conclude that driving above 55 will cause
problems.  so would people driving 40 on the same road. When issues beyond
efficiency of traffic flow and safety of the road come into play, e.g.
politics, revenue pressures from fines and penalties, pressures to spend
money on construction or repair of roads frequently, etc, the true
safety and efficiency of the road suffers greatly. 'reasonable and safe'
is probably the best way to go, overall. However, it makes it difficult to
generate steady, predictable revenue from penalties, and reduces the
amount of secondary control that governments or law enforcement have, and the
overall tendency is toward the increase of such control over time..

Another thing to keep in mind is that vehicle inspections are much more
stringent in Germany than in the US. I am not going to get into the pros and
cons of safety inspections, but i will point out that most vehicles which are
on the road in the US would not be street legal in Germany. Here, the common
practice seems to be that as long as the lights work, the brakes are good,
and the tires aren't bald, the car is fine. In germany, the inspections are
much more intensive and the standard much stricter. It's very expensive to
keep  a car on the road there, and people are forced to trade them in much
sooner or face high maintenance costs to keep the car within spec. The
situation is similar in Japan. Any rust on the vehicle, for example, will
likely fail it on an inspection right off the bat.
(now, one reason they can do this because they have a market in eastern europe
for their used vehicles on a large scale with low transportation cost,
among other things.. but that's another matter entirely)

Isildur


On Thu, 31 Jul 2003, Gary L. Burnore wrote:

> On Thu, 31 Jul 2003 10:46:44 -0500, Dave Uhring <daveuhring@yahoo.com>
> wrote:
>
> >On Fri, 01 Aug 2003 01:19:51 +1000, Anthony Mandic wrote:
> >
> >> Dave Uhring wrote:
> >>
> >>> Pretty tame.  Back in 1970 I drove across Nevada at 145 mph.  But the
> >>> speed limit then was "reasonable and proper".
> >>
> >> 	There used to be no speed limit in the Northern Territory until
> >> 	someone wiped themselves out on a straight road in a Ferrari
> >> 	during a Cannonball Run event. A lot of people got peeved about
> >> 	that.
> >
> >It's a shame about that Ferrari.  But imposing speed limits does not
> >necessarily lead to fewer highway fatalities.  The US interstate highway
> >system has 30% higher fatalities/mile_of_travel than the German Autobahn
> >and the Autobahn has no speed limit at all outside of urban areas.
>
> Saw the numbers on a "Modern Marvels" episode recently and the numbers
> are almost identical despite the speed difference.  30% is an extreme
> exaggeration.
>
> The one thing that the Autobahn has that the Interstate system is
> lacking is enforcement of the slower traffic keep right laws.
>
>
>
> --
> gburnore at DataBasix dot Com
> ---------------------------------------------------------------------------
>                     How you look depends on where you go.
> ---------------------------------------------------------------------------
> Gary L. Burnore                       |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
>                                       |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
> Official .sig, Accept no substitutes. |  �۳�ݳ޳�ݳ��ۺݳ޳�ݳݳ޳�ݳ��۳
>                                       |  �� 0 1 7 2 3 / ݳ� 3 7 4 9 3 0 ۳
> Black Helicopter Repair Services, Ltd.|     Official Proof of Purchase
> ===========================================================================
>

On Thu, 31 Jul 2003, Dr. David Kirkby wrote:

> Each time a company gets a temp in, one might currently reasonably
> expect him/her to know Windoze. The amount of time that would be spent
> on staff training for UNIX is currently too high I feel to make it an
> economic switch for the desktop.

For certain, perhaps many, things, some user training is
required.  But assuming said temp is running StarOffice,
I don't think much training would be required.  (You click
on the butterfly icon instead of the one that says "Office
2000").

When you start getting into sysadminy kind of things, yes,
the difference is huge.  But I wouldn't expect a temp to
do that sort of thing anyway (unless I was specifically
hiring them to do that).

> The cheapest PCs are now more than capable of running Windoze for
> typical office applications. So you can't really say you need more

For desktop apps, I'll concede that point.  How fast does
one's computer need to idle anyway?  :-)

> Linux and Solaris x86 need similarly priced hardware, but training
> costs are higher. I'm sure this will change over time as Linux is
> becoming more popular.

I think part of the problem there is that in UNIX-land,
there's a clear distinction between "end user" and "system
administrator".  But in Windoze-land, the end user is
quite often the sysadmin - even of they don't know it.

-- 
Rich Teer, SCNA, SCSA

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-online.net

Dave Uhring wrote:
> You might want to read this:
> 
> http://www.sun.com/2000-0509/edu/

Dave,
	That page has another interesting quote:

"Making sure that all students in the U.S. and around the globe have
access to the Internet -- regardless of where they live or how much
money they have -- is one of the key issues we face as we move into
the twenty-first century." But note the 'Authorized Education
Auctions' on that page, where one must have a .edu e-mail address in
the US to participate!! 

If Sun wants to make a difference to how all students around the world
participate in the internet, it would be helpful if they allowed
others from outside the USA to bid on the Sun auctions. 

Sun do give educational discounts in the UK, but finding information
about these on the http://www.sun.co.uk web site is often impossible.
I bought a couple of Sun compilers under the ScholarPAC scheme, which
allowed academics to buy the software for personal (non-commerical)
use at reduced prices. However, information on this could not be found
on the Sun UK web site. Nowhere is it mentioned.

I just did a search on "ScholarPAC" on http://www.sun.co.uk and found
one PDF file introducing the Solaris 8 10/01 operating environment,
which must someone in there have the word "ScholarPAC". 

In contrast, do the same search on www.sun.com and you find 21 links.

Essentially, it is very difficult for those outside the USA to find
out about Sun educational offers. Some, like the ScholarPAC scheme are
available in the UK but are not widely publicised. Others, like the
'Authorized Education Auctions' are not available to those of us
outside the USA, despite the fact they are mentioned on a page
describing Sun's vision of the whole world being connected to the
internet. 

Sun, like many large companies, thinks the world stops at the end of
the USA. 

I hope that does not offend "GT Jeff Cherer" <gcherer@texas.net> who
said this about me:

"If i remember right, most of the time it wasn't to offer any help or
info, but to ask for free this or that or to carp why there wasn't
enough free available to university students and faculty."

You now have a bit more ammunition to throw at me Geoff. Have a nice
day. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dr. David Kirkby wrote:
>...
> Of the dealers I know of, most seem to have more SS20's than anything
> else, suggesting that it was the most popular machine...
> Nobody who knows is likely to publically state how many of each of Sun
> boxes  were sold, but I suspect the SMP capable SS20 sold the most. 

Quite right. The SS10 was an amazing box when it first appeared, but the SS20 was in many 
respects a good step (not quite a giant leap) beyond it.

Like David, I have had a number of Suns, starting with a wonderful 3/160. Most have been 
"retired" now, but I'm still running 4 SS20s and 3 Ultra 5s. While the U5 is much faster, 
the SS20 is fantastically reliable and just keeps going, year after year. I wish I could 
say the same for the SS10 :-(

-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GO/! d- s++:+ a+ C++(++++) US++++$ UB++ U*++ P+++ L+++ E--- W+++ N++ w--- O-
M+ V- PS+ PE+ Y+ PGP t+ 5++ X R* tv+ b+ DI++ D G e(*) h++/-- r+++ y?
------END GEEK CODE BLOCK------

-----------------------------------------------------
Bob Hoekstra: APL & Unix Consultant
Telephone:    +44 1483 771028
Mobile:       +44 7710 562345
Email:        Bob.Hoekstra@HoekstraSystems.ltd.uk
-----------------------------------------------------

Rich Teer wrote:
> 
> On Thu, 31 Jul 2003, Dr. David Kirkby wrote:
> 
> > Each time a company gets a temp in, one might currently reasonably
> > expect him/her to know Windoze. The amount of time that would be spent
> > on staff training for UNIX is currently too high I feel to make it an
> > economic switch for the desktop.
> 
> For certain, perhaps many, things, some user training is
> required.  But assuming said temp is running StarOffice,
> I don't think much training would be required.  (You click
> on the butterfly icon instead of the one that says "Office
> 2000").

And when they loose a file? And to write a CD to put a powerpoint
presentation on it? But then there is no powerpoint on Solaris, so if
I was going to give a presentation on someone else's computer, I'd
want to test the software running Powerpoint anyway, despite the fact
StarOffice has its own version. Sorry, I just can't agree that UNIX is
ready for the desktop of the average person. 

I took the decisions to remove Microsoft Office from my PC at work and
put StarOffice on instead. Then I needed to print off someone's PhD
thesis to proof-read. Frightened StarOffice would not be 100% accurate
in reading a large thesis with loads of diagrams, tables, fonts etc, I
decided to put Microsoft Office back on it. 

I'd love to see the back of Windoze, and I'm sure that will come, but
I don't personally think it is there just yet. 
 
> When you start getting into sysadminy kind of things, yes,
> the difference is huge.  But I wouldn't expect a temp to
> do that sort of thing anyway (unless I was specifically
> hiring them to do that).

Can you connect to a printer on the network easily as a normal user on
UNIX? It's not unusual for a printer to break down, so someone has to
print to a printer they have never done before

 
> > The cheapest PCs are now more than capable of running Windoze for
> > typical office applications. So you can't really say you need more
> 
> For desktop apps, I'll concede that point.  How fast does
> one's computer need to idle anyway?  :-)

Well it needs to be fast enough to run the application at an
acceptable speed, even if the application is only used 1% of the
uptime of the machine. But PCs are pretty quick now - hence people
don't see the great need to keep updating them. 
 
> I think part of the problem there is that in UNIX-land,
> there's a clear distinction between "end user" and "system
> administrator".  But in Windoze-land, the end user is
> quite often the sysadmin - even of they don't know it.

That distinction of the UNIX sysadmin and the UNIX user are I feel
being blurred a lot now - assuming you consider Linux and UNIX
together.

Where I work, all the Suns are administered centrally, although if I
wanted root access on a Sun on a grant of mine, I think I could make a
good case and get it (despite the fact it would upset the system
admin!)

But few of our Suns are used now. They are just too old for
number-crunching and too expensive to replace. The only cost-effective
solution for number-crunching is to use Linux. Those who run Linux,
almost always administer the machines themselves. Sometimes a postdoc
will administer a machine and a few PhD students use it, but there is
no central support for Linux currently. 

I don't bother running a Linux PC at work (preferring to work from
home on my Sun box), but if I did, I'd have root access. (But then I
have administrator access on my Windoze 2000 box).

I'm not saying my experiences are typical, but I suspect a higher
fraction of users of Linux machines at work administer them, compared
to those that administer their own Sun/IBM/HP box at work. So I think
that distinction is being blurred.

You are a professional system admin - I am not, so I'm sure you have
more experience than me on these matters. 
-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

In article <3F296003.A91C5337@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:

> SPARC 4 and older - All too old to bother considering even for the
> poorest student, so I'll forget all them. 
>  
> SPARC 5  

The SS4 is a cheap version of the SS5 (cheaper power supply, only one
disk, no SCA disk cage, etc). The SS4 and SS5 was released after the
SS20 and use the same box as the SS20, so it is not an older system.

-- 
G�ran Larsson     http://www.mitt-eget.com/

Goran Larsson wrote:
> 
> In article <3F296003.A91C5337@ntlworld.com>,
> Dr. David Kirkby <drkirkby@ntlworld.com> wrote:
> 
> > SPARC 4 and older - All too old to bother considering even for the
> > poorest student, so I'll forget all them.
> >
> > SPARC 5
> 
> The SS4 is a cheap version of the SS5 (cheaper power supply, only one
> disk, no SCA disk cage, etc). The SS4 and SS5 was released after the
> SS20 and use the same box as the SS20, so it is not an older system.
> 
> --
> G�ran Larsson     http://www.mitt-eget.com/

Okay, I stand corrected. But the SPARC 4 is of too poor performance to
make them sensible choices to buy. They might be newer than the SS20,
but their performance is nowhere near as good as a half-decent SS20. 


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Mikko Nahkola wrote:

> Well ... how do you _make_ the cost- and risk-calculations without the
> chance that someone's personal preference becomes a factor? Who do you
> trust that much?

Facts and Numbers. A calculation of the inital costs of hardware and
software, of installation costs, maintenance costs, reliability, possible
upgrade paths, reaction time, training costs for employees, and a lot more
factors, separately for every platform that fits the requirements of a
certain computer problem. Since we spend a lot of $$$ on IT equipment we
regularly get evaulation systems before making a final decisions so we can
evaulate the real-life performance and reliability. There is no room for
personal preferences within such calculations. Maybe that there are
companies that do that different, but the fact that even during the current
economical recession our company is growing shows that we do something
right...

> And a second point, I've heard people argue that having the people
> work
> on what they like, or at least are already familiar with, might be
> worth
> some. At least the familiarity part if you can't just retrain/replace
> everyone. This is case-specific again though and should be included in
> the calculations ...

How much is it worth if Your employee has an SGI at home where he is
learning Unix if the companies IT structure relies on AIX? Or better, how
much more is this employee worth over another one that uses Linux at home
and is familiar with that? Even when he's a good employee, should we really
ask them what to buy? While the first one certainly would recommend to go
with SGI the second one for sure would vote for Linux and the Open Source.
Both of them might know their beloved platform quite good but nevertheless
have just a very very limited view over the IT market and the available
platforms. Should we really follow recommendations from people that
recommend something simply because they know nothing else (except Windows
maybe)? We don't, since playing around with a Unix workstation at home might
bring You some unix skills but in no way makes You competent for platform
analysis or IT investent decisions...

>> platform at home, You certainly will also settle on cost/risk
>> analysis to evaluate the best solution for Your emplyee's computing
>> needs, otherwiese You risk to make the wrong decision and maybe make
>> room for someone else that cares more for facts than for feelings
>> when spending other people's money...
>
> What if the feelings aren't yours - say, your boss just loves some
> vendor's stuff for some irrational reason?

Of course my Boss could decide to buy platform X even when the analysis says
the costs are much higher than for platform Y. But at that moment I'm no
longer responsible for this decision since my Boss would totally ignore my
analysis, and sooner or later (mostly sooner) the Boss of my Boss will
notice that my Boss hasn't bought the best Solution but just because of
personal preference. Depending on the size of the financial damage my Boss
either will be deployed to a lower grade job, or get fired and eventually
sued for the financial damage...

But this certainly won't happen as my Boss has not much computing knowledge
but still is aware that in times of worldwide economical depression manager
positions might be a bit rare, especially in a growing company. He certainly
won't jeopardize his job just because of personal preference for a computing
platform...

> Not that this ever happens anywhere, mind you. Surely every company
> has
> made a real cost/risk analysis on Windows as the general-purpose OS
>  too - hey, that's a real computing need nowadays ...

Of course we also use Windows, and yes, we made a cost/risk analysis before
deciding to use Windows for certain tasks. And until now we are very
satisfied with Windows, not because it's generally better than Unix, but
simply because it was the best choice for certain tasks...

Benjamin

In article <3F29AD07.9B491F50@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:

> Okay, I stand corrected. But the SPARC 4 is of too poor performance to
> make them sensible choices to buy. They might be newer than the SS20,
> but their performance is nowhere near as good as a half-decent SS20. 

That is true. The numbers in the SPARCStation series does not indicate
time of release, they more or less indicates performance.

The SS20 is still a very usefull machine. I have two dual processors,
one is used as DNS, web, and mail server, one is a spare.

-- 
G�ran Larsson     http://www.mitt-eget.com/

On Thu, 31 Jul 2003 23:19:52 +0100, drkirkby@ntlworld.com wrote:
>...
>But few of our Suns are used now. They are just too old for
>number-crunching and too expensive to replace. The only cost-effective
>solution for number-crunching is to use Linux.

oh stop your trolling. you've done this before.
What makes cost-effective "number-crunching" for you in these cases
is the HARDWARE, not the software.
Particularly for a university, which gets virtually free solaris licenses.
Solaris on x86 would be just as effective for your stated goals as
Linux on x86



>Dr. David Kirkby,
>Senior Research Fellow,
>Department of Medical Physics,
>University College London,


-- 
  http://www.blastwave.org/ for solaris pre-packaged binaries with pkg-get
    Organized by the author of pkg-get
[Trim the no-bots from my address to reply to me by email!]
         S.1618 http://thomas.loc.gov/cgi-bin/bdquery/z?d105:SN01618:@@@D
                            http://www.spamlaws.com/state/ca1.html

Ken Green wrote:

> I would be very surprised if none of your companies purchasing decisions were
> influenced by by peoples personal preferences.
> 
> Initial purchase price is easy for accountants to look at on their own, but it
> wouldn't be the first time a major system purchase wasn't influenced more
> by what the FD (CFO)'s nephew just started at college's personal preference
> is than anything the technical IT staff in the company felt.

	Except that anyone in accounts seldom has any influence in that
	area. Usually, the decisions are made by department heads based
	on the recommendations of their own staff (possibly influenced
	by the sales people they've dealt with) and the advice they would
	have recieved from their IT people. Costs come out of that
	department's budget so it doesn't really matter to the accountants.
	Smaller companies would operate differently (no departmental
	budgets) but its unlikely that the accountant's nephew (or even
	the janitor's niece) would have much if any influence on purchasing
	decisions.

> Running cost of system are much more difficult to gauge. If the guy costing
> how much time admin'ing the system will be, prefers brand X to brand Y they
> will probably feel brand X costs less to look after, coz for them it does...
> they know it better... every thing is quicker.

	But their wages/salary doesn't change. They still get $N per annum
	regardless of what system. Its only when other, additional staff
	are required that added costs are considered. You would also look
	at annual maintenance contracts and obviously power consumption
	for the system(s) and ancillary costs (air-con in a computer room).

> > For smaller installations maybe. But what renevue do You think such "smaller
> > installations" bring to Sun, HP and Co:? Nothing...
> 
> hat depends on how good the dealer chain is. Lots of small systems are worth
> more than a few big ones. Hence Compaq could buy DEC.

	So what happened to Compaq?

> Linux originally found it's way into lots of companies through the back door.
> In many/most cases the management didn't even know it was there.

	Is that right? Amazing how some people run their companies.

> I've come across cases where an administrator has got fed up with re-installing
> Windows for the n'th time over a weekend trying to solve a problem. Then
> at 6 PM on Sunday evening made a decisions to move the box to Linux so
> that they can just make it work and go home to have a life. The first time
> management knew that the system wasn't running Windows was when it's
> lack of down time stuck out like a saw thumb. And no this wasn't a small site
> this was a major international company.

	Not that I don't believe you but can you cite some actual cases
	with details? How come this was allowed with out proper procedures
	in place? What if Linux didn't solve the problem? Was the software
	compatible?

> If that experience teaches them that some things are easier on one box than
> another, then they would be foolish in the extreme not to use that experience
> at work as well.

	Except that at work its not like it is at home. Its the employers
	who make the rules not the employees. If you don't follow the
	rules you don't keep your job.

-am	� 2003

Philip Brown wrote:
> 
> On Thu, 31 Jul 2003 23:19:52 +0100, drkirkby@ntlworld.com wrote:
> >...
> >But few of our Suns are used now. They are just too old for
> >number-crunching and too expensive to replace. The only cost-effective
> >solution for number-crunching is to use Linux.
> 
> oh stop your trolling. you've done this before.

What is your problem? I am not trying to troll at all. 

> What makes cost-effective "number-crunching" for you in these cases
> is the HARDWARE, not the software.
> Particularly for a university, which gets virtually free solaris licenses.

I'm well aware it's the hardware - I'm not stupid.

If you look at my original post, I mentioned home-users far more than
students/univesity staff. Univesities can get good deals on Sun
equipment, BUT:

a) These deals are not exactly well advertised on www.sun.co.uk The
ScholarPac scheme is not mentioned at all, despite the fact it is
available in the UK.

b) The hardware is still much more expensive than PCs of similar
capablility - not helped by the fact Sun charge much more in the UK. I
bought a couple of Sun compilers under the ScholarPac scheme, which
would be $100 in the USA. But they were 100 pounds in the UK, which is
about 60% more. The same difference appeared on Blade 100's at one
point ($1000=1000 UK pounds). Sun seem to use an exchange rate of $1 =
1 UK pound, which is not too favourable.

If I wanted to buy a Blade 2000, I suspect it would be far cheaper for
me to take a trip on a plane to the USA and bring one back. 

> Solaris on x86 would be just as effective for your stated goals as
> Linux on x86

Agreed. But Solaris x86 would probably be more hassle to install, due
to its much more limited hardware support.

Note the follwing:

1) Home-users, can legally download Solaris x86 at nominal cost and
use it on any PC they like - a PC that Sun never made a single dollar
on. 

2) Home-users can NOT legally download Solaris SPARC at nominal cost
and use that on an old workstation that Sun have already made quite a
bit of money on. 

Sun's so-called 'free binary license' for Solaris 9 is bascially
useless. Nobody cound have moaned about the Solaris 8 license. 


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

"Dr. David Kirkby" wrote:
 
> Of the dealers I know of, most seem to have more SS20's than anything
> else, suggesting that it was the most popular machine. There's one on
> eBay at his minute with a starting price of $9.95 and no bids with
> under and hour to go. They are dirt-cheap.
> 

It appears the U5 was Sun's most popular workstation. 
http://www.sun.com/smi/Press/sunflash/2001-12/sunflash.20011206.2.html

although there are a lot more SS20's on eBay than U5's. Perhaps
because the U5 is not so old, that companies that bought them hold on
to them - we have several at work. In contrast, all the SS20's have
been disposed of.

Whatever the exact numbers, there are a *lot* of Sun SMP capable
machines on the used market at affordable prices. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring wrote:

> >       Which tells us what? The standard of driving between the countries
> >       varies or the quality of vehicles varies or the road conditions are
> >       different? Anything else?
> 
> Probably a combination of factors.

	Yes, I'd think so too.

> Most American drivers fail to pay heed to the traffic conditions around them.

	Some of the multicar pile-ups I see on the news can't be helped
	(foggy conditions, although I would wonder why everyone was driving
	fast). However, most of the reports I see are from the US (but this
	may just be media bias).

> The Autobahn is probably better maintained than most US highways and
> the German automobiles are designed to be run at speed.  Certainly the
> Ferrari was designed to be driven at high speed.

	There's also less of the Autobahn than the American interstates.
	Of course, I could also wonder why there is so much of it (and
	the amount of traffic I've seen on it surprises me too).

> What happened to the Ferrari?  Did it hit a 'roo or wallaby, or perhaps a
> pothole in the road?  We have a situation here with deer.  I've hit 2 of
> them while driving in the wee hours of the morning.

	Not sure. In fact I can't recall if it really was a Ferrari. I'm
	sure it was an Italian car. From what I can recall (it was a few
	years ago now) it was after dark on a straight stretch of road.
	So it could have been driver fatigue. If he veered off the road,
	that would have been enough to make the car roll at high speed.
	I seem to recall that the car was a write-off.

	Hmmm ... I should really do a Google search ... here's some
	references -

	http://www.abc.net.au/am/s663617.htm
	http://www.anarki.net/angry/ap14/ntstate.html
	http://www.ntlib.nt.gov.au/nac/ntchron/19941.html
	http://aussietourism.info/self_drive.htm

	So, from that we can piece together that it was a Ferrari,
	the date it happened, where, how many died and the fact that
	there's a memorial there now so you could visit the site and
	see for yourself.

-am	� 2003

Dave Uhring wrote:

> Perhaps you taped the show and can provide the exact numbers.

	Meanwhile ... I'll listen to the tune by Kraftwerk.

-am	� 2003

Ken Green wrote:

> All I can say is that your very lucky not to live a the same world as I seem to.

	Well, you appear to be in the UK. I'm on the other side of the
	world. So, yes, in a sense I am lucky.

> With out getting into a religious war over Windows against real computers :-)

	Not quite a correct comparison. Windows is an operating system
	(sort of, anyway). Computers are hardware.

> How do you think we have ended up with many companies having made a
> strategic decision to move to a Windows only environment? These decisions
> have almost always been made in direct opposition to all the technical staff in
> a company. The decisions have been made by upper management.

	I've been in clients sites where the staff push Windows as the
	only or best solution (its all they know). Managers would choose
	otherwise (I know one that preferred VMS and Alphas but he relented
	to Suns and Solaris and left Windows to do the crap stuff - although
	he moved their Lotus notes email from an NT box to a Sun). So it
	can vary.

> In every large company I've ever dealt with department heads don't get much
> say about buying $50M worth of kit. They make recommendations yes. but they
> probably don't even get invited to the meeting where the decision is made.

	Its my experience that its the department heads who "own" the
	hardware (they usually buy the solution first - some software
	that does what they want - and then consider recommendations
	for what to run it on). But, yes, if IT only know one OS then
	that's the corporate standard and what gets used. They wouldn't
	take to being excluded from any decision making process that
	affects them. More than likely they may make decisions while
	excluding the staff who do the actual work and best know the
	lay of the land.

> Wages/salaries are different for different kinds of systems. look at job adverts.
> Wages/salaries are paid on the basic of hours spent working, not on amount
> of work done. If you find that it takes twice as long to do some thing on box
> Y as opposed to box X, then it probably costs twice as much as own box Y.
> How long it takes quite often depends on the familiarity of systems.

	Funny, when I was salaried, I never got overtime. But if you're
	already doing the work, you income won't change if you change
	to a different system (unless management wants to screw you -
	which has been known to happen. But then you can always vote
	with your feet).

> > > hat depends on how good the dealer chain is. Lots of small systems are worth
> > > more than a few big ones. Hence Compaq could buy DEC.
> >
> >         So what happened to Compaq?
> 
> Dell

	Wasn't it HP? But no matter, the point was obvious.

> > > Linux originally found it's way into lots of companies through the back door.
> > > In many/most cases the management didn't even know it was there.
> >
> >         Is that right? Amazing how some people run their companies.
> 
> Systems that work are usually invisible to management. They only tend
> to worry about things that don't work. If it ain't broken don't fix it.

	Its in their interest to ensure staff follow procedure. But there
	are managers and there are managers.

> >         Not that I don't believe you but can you cite some actual cases
> >         with details? How come this was allowed with out proper procedures
> >         in place? What if Linux didn't solve the problem? Was the software
> >         compatible?
> 
> No I won't quote companies names, I'd be breaching confidence.

	That's OK but for a multinational company to do this sounds a
	little reckless. They usually follow strict guidelines for QA
	testing etc.

> In this instance the job the windows server was doing was something that
> Linux could do just as well, without any user ever knowing the difference.
> How could they do it? Their job is to make something work. If it works
> then no one notices. If they'd got it wrong, they'd have been screwed for
> do doing it.

	Except that there's a ripple effect. The Windows box was probably
	being backed up (I say probably because we know it probably wasn't).
	But what happens to the Linux box? There may have been a centralised
	backup server that handled the Windows boxes but not Linux (unless
	a backup client was installed). So, errors start to mount up. Sooner
	or later someone would have noticed. And what if the replacement
	failed? Procedures and guidelines are there to cover everyone's
	arse, not just the company's.

> >         Except that at work its not like it is at home.
> 
> No there are many differences between home and work, but there are also many
> similarities.

	Such as?

> One of the major things I learnt from have a rp7400 at home is that it's bloody
> noisy
> and throws out too much heat. This is not relevant to work. At work I keep boxes
> like this is a computer room.
> 
> On the other hand I can learn an awful lot about, partitioning tools, scheduling in
> MP systems, managing memory in systems with multi gigabytes of RAM, IO balancing
> multiple IO connections and raid controllers... etc.

	What? At home?

> In all of the places I've worked you get measured on results. When
> results aren't good, then following rules is important.

	Hmmm ... what the public service like over there? Over here you can't
	get sacked even if you screw up. But since its full of incompetents,
	I can make a nice earning when they call on me to do the work.

> Isn't copywriting your usenet posting just a little bit silly?

	Not since Google claimed that they had Usenet copyrighted.

-am	� 2003

On Fri, 01 Aug 2003, Benjamin Gawert <bgawert@gmx.de> wrote:
> Mikko Nahkola wrote:
> 
>> Well ... how do you _make_ the cost- and risk-calculations without the
>> chance that someone's personal preference becomes a factor? Who do you
>> trust that much?
> 
> Facts and Numbers. A calculation of the inital costs of hardware and
> software, of installation costs, maintenance costs, reliability, possible
> upgrade paths, reaction time, training costs for employees, and a lot more
> factors, separately for every platform that fits the requirements of a
> certain computer problem. Since we spend a lot of $$$ on IT equipment we
> regularly get evaulation systems before making a final decisions so we can
> evaulate the real-life performance and reliability. There is no room for
> personal preferences within such calculations. Maybe that there are
> companies that do that different, but the fact that even during the current
> economical recession our company is growing shows that we do something
> right...

In my field, while you are busy doing calculations and studies
for every platform that fits the requirements, someone else
would be finishing up the job; they would have looked at
whether it ran on a UNIX, Linux, or Windows, and picked
one with about 5 minutes thought, based on their experiences.

The people that are doing these decisions don't have experience
with just one platform, but many. Personal preference does play a
part.

The other problem with extensive paper analysis is that all the
vendors know that this is what people do, and make sure that
they hit all the checkmarks on a list. They aim to look good
on comparisons like that, whereas problems don't show up until
you're running the system for a while -- maybe you discover
that it needs regular, extensive security updates, for example,
that you need to apply within a few days or your audit department
will go through the roof, and you've got to have a few extra
FTE on staff just to test and apply all these patches.
(a personal rant...)


> How much is it worth if Your employee has an SGI at home where he is
> learning Unix if the companies IT structure relies on AIX? Or better, how
> much more is this employee worth over another one that uses Linux at home
> and is familiar with that? Even when he's a good employee, should we really
> ask them what to buy? While the first one certainly would recommend to go
> with SGI the second one for sure would vote for Linux and the Open Source.
> Both of them might know their beloved platform quite good but nevertheless
> have just a very very limited view over the IT market and the available
> platforms. Should we really follow recommendations from people that
> recommend something simply because they know nothing else (except Windows
> maybe)? We don't, since playing around with a Unix workstation at home might
> bring You some unix skills but in no way makes You competent for platform
> analysis or IT investent decisions...

Since such a person likely invariably also has some Windows experience,
it makes them a touch more informed than a lot of others.

> 
>>> platform at home, You certainly will also settle on cost/risk
>>> analysis to evaluate the best solution for Your emplyee's computing
>>> needs, otherwiese You risk to make the wrong decision and maybe make
>>> room for someone else that cares more for facts than for feelings
>>> when spending other people's money...
>>
>> What if the feelings aren't yours - say, your boss just loves some
>> vendor's stuff for some irrational reason?
> 
> Of course my Boss could decide to buy platform X even when the analysis says
> the costs are much higher than for platform Y. But at that moment I'm no
> longer responsible for this decision since my Boss would totally ignore my
> analysis, and sooner or later (mostly sooner) the Boss of my Boss will
> notice that my Boss hasn't bought the best Solution but just because of
> personal preference. Depending on the size of the financial damage my Boss
> either will be deployed to a lower grade job, or get fired and eventually
> sued for the financial damage...

Realistically, this happens all the time. All the time. Everywhere.
Why? Your boss may be a dickhead, or maybe just maybe has contacts
that give them better insight into the future of the vendors, one
of those intangible things that are hard to express as a line item.

Generally, list your supportable options, and recommend the one you
like best. If they pick another supportable option, it doesn't
mean you're no longer responsible. The trouble lies if they
want a solution you consider unsupportable -- then the fur flies.

> But this certainly won't happen as my Boss has not much computing knowledge
> but still is aware that in times of worldwide economical depression manager
> positions might be a bit rare, especially in a growing company. He certainly
> won't jeopardize his job just because of personal preference for a computing
> platform...
> 
>> Not that this ever happens anywhere, mind you. Surely every company
>> has
>> made a real cost/risk analysis on Windows as the general-purpose OS
>>  too - hey, that's a real computing need nowadays ...
> 
> Of course we also use Windows, and yes, we made a cost/risk analysis before
> deciding to use Windows for certain tasks. And until now we are very
> satisfied with Windows, not because it's generally better than Unix, but
> simply because it was the best choice for certain tasks...

Good for you. We need UNIX & Linux though, for certain tasks.

-- 

Leach

On Fri, 01 Aug 2003 22:36:50 +1000, Anthony Mandic wrote:

> 	Some of the multicar pile-ups I see on the news can't be helped
> 	(foggy conditions, although I would wonder why everyone was driving
> 	fast).

American drivers have a dangerous habit of tailgating, i.e., following too
closely, even though such a practice is illegal.  The vast majority of
traffic accidents which I have seen on the interstates can be attributed
directly to this behavior.  With the speed limit on the interstates around
St. Louis, MO, at 60 mph, even the cops travel at 75-80 in the hammer lane
with a train of other cars following them and spaced less than 100 ft
apart.

> However, most of the reports I see are from the US (but this
> 	may just be media bias).

We just have more automobile traffic than other countries so I doubt any
media bias plays here.

> 	There's also less of the Autobahn than the American interstates.
> 	Of course, I could also wonder why there is so much of it (and
> 	the amount of traffic I've seen on it surprises me too).

The heaviest traffic here is during the hours when people are driving to
and from work and school.  It's likely the same occurs in Germany; I know
for sure that happens in Mexico City.

> 	So, from that we can piece together that it was a Ferrari,
> 	the date it happened, where, how many died and the fact that
> 	there's a memorial there now so you could visit the site and
> 	see for yourself.

Yep, a Ferrari driven by a Japanese millionaire.  The race was between
Darwin and Ayres Rock and the Ferrari seems to have gone about 760
straight-line miles (somewhat more distance along the roads) before
hitting the race officials.  That's a fair distance and the driver may
have been, probably was, suffering from exhaustion and fatigue.  The lack
of a speed limit does not appear to have been the primary contribution to
the deaths.

Those people would have been just as dead if the accident had happened at
70 mph, approx. the proposed speed limit of 110 kph.

Dave Uhring wrote:

> American drivers have a dangerous habit of tailgating, i.e., following too
> closely, even though such a practice is illegal.

	Yes, I've noticed that.

> The vast majority of
> traffic accidents which I have seen on the interstates can be attributed
> directly to this behavior.  With the speed limit on the interstates around
> St. Louis, MO, at 60 mph, even the cops travel at 75-80 in the hammer lane
> with a train of other cars following them and spaced less than 100 ft
> apart.

	I don't know what a hammer lane is but don't they book the
	drivers speeding behind them?

> > However, most of the reports I see are from the US (but this
> >       may just be media bias).
> 
> We just have more automobile traffic than other countries so I doubt any
> media bias plays here.

	Also more media and hence more media reports. Plus the fact that
	its cheaper/easier to show foreign English language news reports
	locally. We get a heavy slant towards U.S. and U.K. newsfeeds
	here (except on SBS).

> The heaviest traffic here is during the hours when people are driving to
> and from work and school.  It's likely the same occurs in Germany; I know
> for sure that happens in Mexico City.

	The thing I mentioned before that I found surprising was the
	amount of traffic on an interstate or freeway between Sacramento
	and the San Francisco Bay area. It was the middle of nowhere (no
	development apart from the road - just empty hills and fields),
	night time and a ribbon of headlights and taillights. Nearly choked
	with cars, I wondered where everyone was going and why there was
	no public transit. Anything like that in Sydney would be in town
	and around rush hour. You wouldn't see it out in the country.

> Yep, a Ferrari driven by a Japanese millionaire.  The race was between
> Darwin and Ayres Rock and the Ferrari seems to have gone about 760
> straight-line miles (somewhat more distance along the roads) before
> hitting the race officials.  That's a fair distance and the driver may
> have been, probably was, suffering from exhaustion and fatigue.  The lack
> of a speed limit does not appear to have been the primary contribution to
> the deaths.

	But speed was a factor. If he was going slower, he would have
	probably survived - or, at least, not killed anyone else. I
	seem to recall that there was an inquest but I don't recall the
	details now. His experience may also have been a factor but, yes,
	fatigue was probably the biggest factor. The same conditions
	would have been applied to all the race contestants though. I
	don't recall anyone else having a problem.

> Those people would have been just as dead if the accident had happened at
> 70 mph, approx. the proposed speed limit of 110 kph.

	Yes, probably but the circumstances would all have changed. If he
	was going slower overall, he wouldn't have gotten to that point
	until later, etc.

-am	� 2003

Lord Isildur wrote:
> 
> you have this ongoing obsession with certification. What's the deal with
> that?

	If you consider asking twice or thrice about certification to be
	obsessive, what would you call it when I really sink my teeth in?

	Doug Dingus' original statement was "Your best free UNIX is Linux.".
	So, how can it be really called Unix if its isn't ratified as such?
	Since I questioned that, the terminology changed to Unix-like. Why?

> >       No my question is what advantage does Open Source code offers students.
> 
> umm.. let's see.
> freedom (not just monetary, but license-wise, legal, IP, etc type freedom also),
> the means both technical and authoritative to modify the code to your heart's
> content, the rich, active development community, need I continue?

	Yes, modifying code isn't entirely prudent. What happens when
	you change, say, networking code that does TCP/IP and you then
	are no longer able to communicate with other systems on a network?
	What about modifying a driver so that access to hard disks breaks
	or monitor display screws up. Doesn't sound like modifying code
	is entirely desirable. I would posit that only a handful of
	people can modify code successfully the rest just compile it.
	Those in the latter category could just as easily accept precompiled
	binary packages because the source code doesn't really carry much
	benefit for them.

> I'm not sure what dungeon you live in, surely one that is certified for
> something, but it would behoove you to occasionally let the sun and fresh
> air in.

	I'm reading your post in my loft - where there is plenty of
	fresh air and sunshine. I'm more interested in POSIX compliance.
	Can you, or anyone else, tell me if Linux is POSIX compliant and
	why Microsoft went to the lengths they did to be compliant?

-am	� 2003

Dave Uhring wrote:

> >       No my question is what advantage does Open Source code offers students.
> 
> You might be somewhat hard pressed to run a network without Open Source
> code.  The BSD TCP/IP stack comes to mind as well as BIND, ntpd and
> sendmail.  Your graphics displays would perhaps not work without X11.

	Whilst this is true, you don't really get the source with Solaris.
	And, even if those programs didn't exist, there'd be other software
	the could be used to replace them (I keep thinking of SunView in
	comparison to X11, in particular, but there are plenty of sendmail
	replacements).

> Since those examples of Open Source code seem to be reasonably capable and
> stable, perhaps the study of them would benefit students.

	Yes, but the issue boils down to which and how many. Rather than
	being all or most students its really only a very small proportion
	of students - those studying computer science or information
	technology at an advanced level. To anyone else, its not really
	of much concern if you have the source code or not. As far as end
	users go, having the binaries and being able to configure them would
	suffice. The source code would probably be meaningless and hence
	useless to the majority of users. Modifying said code would be even
	worse (as I've pointed out elsewhere, you can't really change the
	fundamental protocols) and, based on the number of posts I see on
	a regular basis, compiling said code appears to be a complex task
	for a lot of people. Solutions like SunFreeware and Blastwave must
	exist for a reason, no?

-am	� 2003

Anthony Mandic <ql@hotmail.com> writes:

> 	Doug Dingus' original statement was "Your best free UNIX is Linux.".
> 	So, how can it be really called Unix if its isn't ratified as such?
> 	Since I questioned that, the terminology changed to Unix-like. Why?

Even the dimmest of bulbs realise that different people need different
things out of their OS, and put different levels of importance to the
strictness of the Unix label. I don't see why you have to get all
Socratic about it.

> 	I'm reading your post in my loft - where there is plenty of
> 	fresh air and sunshine. I'm more interested in POSIX compliance.
> 	Can you, or anyone else, tell me if Linux is POSIX compliant and
> 	why Microsoft went to the lengths they did to be compliant?

I can tell you that Linux is certainly POSIX compliant. Wether Linux
is POSIX conformant is a more interesting question.

As a curiosity, at least one distribution has claimed to be certified
as conformant to POSIX 1003 in the past

http://www.ukuug.org/newsletter/linux-newsletter/linux@uk21/posix.shtml

*p

In article <3F2B60AD.DA07B392@hotmail.com>, Anthony Mandic <ql@hotmail.com> 
wrote:

[...]
: 	Doug Dingus' original statement was "Your best free UNIX is Linux.".
: 	So, how can it be really called Unix if its isn't ratified as such?
: 	Since I questioned that, the terminology changed to Unix-like. Why?

People don't care that The Open Group owns the UNIX[tm] trademark, and charges 
companies up the wazoo for certification and logo use licences. "Unix-like" and 
"Unix" are interchangeable in regular conversation, but if you get picky most 
people aren't terribly pained to use the 'more correct' terms.

My reading of this page <http://www.opengroup.org/openbrand/register/xu013.htm> 
could mean that IRIX, when running on R12k or later CPUs, isn't UNIX. Does that 
mean IRIX isn't unix in the common sense, and that SGI is no longer sells unix 
systems? Of course not, that would be silly and pedantic, and nobody likes a 
pedant.


[...]
: 	Yes, modifying code isn't entirely prudent. What happens when
: 	you change, say, networking code that does TCP/IP and you then
: 	are no longer able to communicate with other systems on a network?
: 	What about modifying a driver so that access to hard disks breaks
: 	or monitor display screws up. Doesn't sound like modifying code
: 	is entirely desirable. I would posit that only a handful of
: 	people can modify code successfully the rest just compile it.
: 	Those in the latter category could just as easily accept precompiled
: 	binary packages because the source code doesn't really carry much
: 	benefit for them.

"You can modify the code yourself" is shorthand for "You can pay someone 
competant to modify the code for you" in most cases.

It can be the difference between "The application broke, we can't fix it" and 
"We're kinda fucked, but we'll live", or "The vendor says it's secure" and "We 
can personally/independantly audit every single line".

Most of the time (luckily) there's no reason to change the code, but 'most of 
the time' isn't 'all of the time', so it's nice to have the option available.


[...]
: 	Can you, or anyone else, tell me if Linux is POSIX compliant and
: 	why Microsoft went to the lengths they did to be compliant?

People think linux is an idealistic system. POSIX compliance is certainly an 
idealistic goal, perhaps even noble in many aspects.

But no, linux is founded on pragmatism. Pragmatism says that nobody should be 
allowed to call gets(), ever, no matter what the POSIX standards say. Pragmatism 
also says linux pthreads works well enough to not fret about it 'till someone 
else more concerned ponies up to write a better implementation.

If you know ways of changing linux to be more POSIX compliant *in ways that make 
sense*, there's a big line of developers who would love to see your patches.

In other words: No. Linux will never be fully 100% compliant with any currently 
published POSIX standards. Future POSIX standards may be developed such that it 
may eventually make sense for linux to be 100% compliant, but that day is not 
nigh.

The POSIX standard is far from perfect, and linux is far from the only system 
that is willfully non-compliant.


Cheers - Tony 'Nicoya' Mantler :)

-- 
Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@apia.dhs.org
--  http://www.apia.dhs.org/  --  http://nicoya.feline.pp.se/  --

Doug Dingus wrote:
> 
> You are quite possibly the worlds most perfect example of a troll.

	That's probably debatable. I'd pick Kirbu. I'm more interested in
	sorting the wheat from the chaff.

> I took a moment to google your posting history.  Do you always
> communicate like that?

	Like what? Top post, as per your example? Its not something I'm
	in the habit of doing. Its also not the wisest move to telegraph
	your punches in a debate but, while we're sharing, I took the
	trouble to do the same. Not really enough out there to form an
	opinion but I did see one or two posts that I found amusing.

> Anthony Mandic <o6@hotmail.com> wrote:
> > Doug Dingus wrote:
> >
> > > > > Your best free UNIX is Linux.
> > > >
> > > >       Oh? And I suppose you can point us all to Linux's UNIX certification?
> > >
> > > Who cares really.
> >
> >       It matters to those who are concerned with these things.
> 
> Agreed.  I am not worried at all. (About true UNIX or not issues.) The
> Open Group can say what they want.  Linux looks like a UNIX, acts like
> a UNIX.  That is UNIX enough for me without actually being a UNIX!  I
> get all the nice UNIX stuff without any of the costly license issues.
> --No Brainer really.

	That may explain why you seem confused about the state of affairs.
	In one post you state "Your best free UNIX is Linux" while, in
	another, you state "Linux is not a UNIX". So what is it? It either
	is or it isn't? It can't be both and not both can it? The real
	crux should be what is its POSIX compliance?

> > > From a students point of view, Linux is a very
> > > good OS to learn which is why I wrote what I did.
> >
> >       Why?

	OK, now that you've supplied some comments to support your views,
	let's examine them.

> 1.  Because one can learn a lot about *real* computing for very little
> monetary investment.

	When I was a student at university studying computer science, I
	didn't have my own computer. So I would argue that it requires
	no monetary investment. You could do everything on the university
	systems.

> 2.  Linux development is open and very active.  A student can become
> part of the process, learn something, add something to their resume,
> and make good use of others work at the same time.

	I've dissected elsewhere which students fit this category (only
	a small proportion) and how beneficial it really is. I'm not
	sure if I would be impressed by a graduate that listed something
	like this on their resume (I've yet to see one that did). Unless
	they are brilliant, they wouldn't have near enough experience to
	be a competent developer.

> 3.  Linux is a fine example of UNIX even though it isn't one.  For
> those trying to progress in computer science, this is a very nice
> point.

	Its also a tautology. See above.

> 4.  The body of GNU code can be used for almost any sort of project
> without royalty, license agreements (save the GPL), or restrictions.
> This code includes many applications that will do a lot of things for
> only a little time spent learning how.

	That's fine but how does having the code help? All you need
	is the binary applications.

> 5.  Runs on anything that matters.

	That's fine but how many run it on something other than Intel?

> 6.  Is very inexpensive and very flexible.

	You seem to be repeating points 1 and 5.

> 7.  UNIX is like a language.  It is both the art of computing and
> computing itself in the same expression.  This cannot be understood
> running win32 for example.  Linux is more than UNIX enough for this.
> A lawyer would understand this point because the law works much the
> same way.

	I would think a lawyer would more likely focus on "Linux is
	more than UNIX enough for this". Seems you've introduced a
	third definition. Unix is just an operating system. There
	are plenty of those around. Anyone can choose to study or
	use any of them. The concepts are more or less the same and,
	more to the point, the end results of computations should be
	identical.

> If you have to ask why #7 is important...  save some face and just
> don't.

	I'm more bemused by your claim that "UNIX is like a language".

> > > I would likely agree on a number of points there, but Linux has the
> > > mindshare right now along with momentum.
> >
> >       And exactly how much mindshare and momentum does it have?
> 
> Almost all of it.  (Seriously.)  Take a quick look around the web,
> it's not hard to miss.

	What I do from time to time is check what its mindshare in the
	commercial sector is. I don't see many jobs that use it and, of
	those that do, its not the primary focus. I can't see anyone
	studying it likely to get a job with it. They would be better off
	following the market. Here's an opinion from a vendor in the
	commercial sector -

http://resellernews.co.nz/reseller/reseller.nsf/0/86256ABD006E93FCCC256D2B007D35B4?OpenDocument&More=Hot+News

> > > Your point is valid, if hasty.  Maybe I should have written:  "Your
> > > best all-around free UNIX workalike is Linux."  Either way, the intent
> > > is the same.
> >
> >       That would yet need to be determined. What other free UNIX or
> >       UNIX-like choices are available and how do they compare?
> 
> So, maybe we disagree.  Care to explain why your opinion differs?
> Maybe you prefer a BSD?  I know many purists who have good reasons for
> doing so.  What are yours exactly?   Oh, don't forget follow your own
> rules.  (Supporting Facts, clear arguments and all that...)

	I don't have a preference for any of the free Unix wannabe OSes.
	Yes, there are a few BSDs available. There are also other choices
	(I'm closely following Plan9 - the future of Unix). However, for
	the purposes of maintaining my livelihood, I stick to what I know
	best and what keeps me in business.

> > > My personal commercial UNIX preference is IRIX :)  I wonder how things
> > > will continue to go though.
> >
> >       Based on the some posts I see from time to time, I think you
> >       are not the only one who is concerned about the current state
> >       of things in the industry.
> 
> For one preaching clarity of thought and communication, I find this
> statement highly amusing!

	Yes, I meant "Based on some of the posts ...".

> To clarify a bit:
> 
> I am worried about the state of the industry, but not for commercial
> UNIXes.  I worry about patents, Microsoft, and other legal issues that
> threaten our ability to actually own and make use of general purpose
> computers that we have full control of.

	I don't see Microsoft preventing anyone from using any competing
	OS on the computer of their choice.

> BTW, Open Source is a very nice check against that.  (There is no
> better.)

	How is that?

> (wasn't good before...) -- yes.  (How long...) about 10 years I think.
>  (When good...)  About 2 years ago with the arrival of the 2.4 kernel
> series.

	OK. That explains why in 2001/2002 when I was using it, it wasn't good.

> (How determine good...)

	I would either quote Shakespeare's Hamlet "Nothing is either good
	or bad but thinking make it so!" or Plato's Phaedrus "What is good
	and what is not good? Do we need others to teach us these things".

> For me, the "goodness" of Linux is measured by what I can do with it.

	So you're making a subjective value judgement?

> Today, Linux is at the point where I can finally start asking why I
> should buy software.  Some software is worth a purchase because it
> serves a niche need.  (CAD / CAM, Analysis, High End Video and
> publishing are all examples of software that I might purchase.

	So why isn't it free?

> I would
> be reluctant to purchase software that does not run under an OSS
> however.)  Another measure is the need for win32.  I can say that
> win2k is likely to be my last Microsoft OS.  (That's a damn good
> thing.)  Still another measure is its ability to work with other OSes.
>  (Lets just say it does this very well to date.)

	You mean like with nfs?

> So, a big part of the answer to this question is one of personal
> worth.  Each person must answer it on their own terms.  You know mine,
> how about yours?

	You mean like "What's it worth to you?". If everything was
	free, where would the money come from to help support those
	doing development. I think Raj Nathan is right. Its not really
	sustainable.

> One final measure of the "goodness" of Linux would be the number of
> people using it.  Said number is growing nicely at the moment
> indicating a nicely improving level of "goodness" in general.

	Or misguidedness? Measuring how many people use MS OSes, does that
	make them better than Linux?

> >       Being a little loose with words and their meaning doesn't help.
> >       I'd like to know what real advantage Open Source code imparts.
> >       As far as I can tell, its possession is not useful for the vast
> >       majority of students.
> 
> Wow, I am not sure you can really "tell" anything really.  I guess you
> have a problem getting past the literal expression of things.

	When working with computers, how can you be anything other than
	precise?

> Literally, you are right in that possession of Open Source code won't
> do many people any good.  The binaries are useful however.  You really
> need to take a moment to consider the nature of what people say to you
> and not the literal words used.  Until you begin to do this, you will
> find you put off most people while at the same time sharply limiting
> your ability to actually learn anything.

	No, its just deductive reasoning. And now that I've manoeuvred
	you into admitting that its the binaries and not the source that
	is important to the vast majority of people (although we were
	talking students initially and you hadn't really defined which
	students), do it matter how this source is managed?

> > > All you have to do is learn to compile code and understand how the OS
> > > and hardware interact.
> >
> >       And how do rate this learning? Easy or hard?
> 
> It is variable over time and the nature of the student in question.
> It also is variable depending on the learning path chosen.
> 
> For me, it was hard early on, but much easier as time passes.  (I
> started from a win32 perspective which really hosed a lot of basic
> things up. --I'm much better now!)
> 
> For a student wanting to be able to make good use of computers in
> general, the learning is doable and well worth it.  How many things in
> life are truly easy yet deliver high returns?  Of these things, how
> many really are there for the taking?  OSS is both.

	From my observations of the number of posts I see where people
	state that they are having compilation problems, I would have to
	question whether its an issue of the code itself, the compilation
	process or the skill of the person attempting the compile. From
	what I am able to deduce, if they are able to find their way to
	a newsgroup (but not always the most appropriate one), post their
	problem and include an example, they can't be entirely stupid.
	That only leaves the code and the compilation process.

> > > --You can trust the machine to a very high degree.
> >
> >       The machine or the code? And how do you develop this understanding?
> 
> (machine or code) Both.

	Are you familiar with Ken Thompson's compiler trojan horse?

> (understanding?)  One question at a time...  It really is as simple as
> that.

	Couldn't get any simpler than that?

> >       And what happens when everyone changes their own version of the
> >       Open Source code to suit their own particular needs? What
> >       happens to the standards?
> 
> Well, nothing really.  Open Standards are open so people will use
> them.  If someone changes things that a lot of people are using, they
> have little incentive to follow along.  That is what OSS is about
> --choice.

	Except for the issue of protocols. I've mentioned this elsewhere
	so I won't repeat it.

> You see, when a producer of closed binary only software makes a change
> and ends distribution of the older version, everyone involved
> eventually must change.  They call this progress, I call it extortion.
>  Either way, you cannot call it choice by any measure.

	You are free to stay with the old version. I don't view enhancements
	as extortion. I also don't see this as being any different to what
	happens with Open Source. How many people still run very old versions
	of Linux compared to newer ones?

> > > Any student who takes the time to realize this, will be in control of
> > > their computing experience.  They will understand they have choices
> > > beyond those few they are presented and will be in a position to take
> > > advantage of them if they choose.
> >
> >       And do they choose to do so? Why or why not?
> 
> God, you ask such silly things!

	That's only because you haven't yet realised what I'm up to.

> They will choose the path of best
> advantage for them while being able to continue to make use of the
> body of solutions (not just OSS) they believe they need.

	And what is of best advantage to them? Getting a well paid
	job and starting a successful career perhaps? Which enables
	them to better follow their genetic imperative and reproduce?
	Are they more likely to do that with Linux or one of the more
	established commercial OSes? What if they don't value this? Perhaps
	become a junkie and turn to crime. Will Linux help them there?

> >       Secondary or tertiary education? And in exactly what fields of
> >       education? How does it help someone learning dramatic art for
> >       example?
> 
> (Secondary or tertiary?)  --Any educator contributing to ones
> knowledge of computing use or theory.

	What I was getting at was the original remark about students
	benefiting from this. I posit that its only a small fraction
	in the overall number of students. Most wouldn't have a need
	for it.

> The educators who are involved with computing need to make their
> students aware of the differences polarizing the industry today and
> how those decisions will affect them in the future.  We are reaching a
> point where people need to perform basic compuing tasks in order to
> participate in society today.

	You don't really need to know much to use an ATM. It isn't
	that much different to a telephone. What examples can you
	provide when knowledge of computing is needed in society?

> So, how does this affect our dramatic art student?  Well, they might
> get their coursework in electronic format.  If that format is closed,
> they must pay to read their educators instructions as well as pay for
> the educator.  If that format is open, they can receive those
> instructions without payment, or the lowest amount of payment
> depending on their experience.

	That's more than likely stretching it a bit. One would need
	to assume that the educators are also computer literate. I
	don't think its really penetrated that much outside of the
	science and engineering faculties.

> Lets say this student wants to produce a work of their own.  How will
> they do this?  Will they simply perform it and rely on others to
> transribe it.  (Perhaps, but not likely.)  If this work requires a
> group effort, like so many things in theatre do, they must
> communicate.  How?  On what platform and at what cost.

	I don't see it being done with a computer.

> If said student understood the very basics of computing, they could
> use OSS on very common hardware to communicate with others at the
> lowest cost via pre assembled OSS distributions.  They would not need
> to compile anything, just run those good applications that matter on
> their hardware and not worry about anything else.  (Mandrake and
> such.)

	So, is there software available for the performing arts?

> Said student could also trade some work for others to setup an
> environment for them.  Maybe they have needs not easily met with a
> standard distribution or want to be creative in some way with regard
> to computing.  OSS lets them do this to a degree other closed
> solutions, including closed UNIXes, don't.
> 
> In short, people need to know why OSS exists because they might want
> to make use of the choice it provides.  This matters now and is going
> to matter more in the future.

	I think I would have to disagree. I can't see any examples of
	it really mattering now. I can't call the future.

> I am not sure you really know how to argue at all.

	I would have though that my debating skills were evident.

> There is a clear
> difference between the lottery and learning about computing.  You can
> buy all the tickets you want and still never win.

	Yes, the odds are well know. None the less, someone does win.

> Learning how to make good use of Open Source tools, at any skill
> level, will provide a clear return in the form of lower computing
> costs at a minimum.  You will get more out of it than you put in every
> time.  The sharp contrast to the lottery could not be more clear.
> 
> Am I set for life?  As far as computing goes I am.  I can use almost
> any hardware to do most anything I choose to do.  For those things I
> must still pay for, I still have many options not open to those using
> closed solutions only.  Computing is never going to cost me what it
> does others who just pay and trust without asking questions.

	Given the current recession in the IT industry, which appears
	to be a fairly wide phenomenon, I would suggest that there are
	some people out there who don't see themselves now as being
	set for life in this industry. I wonder how many are buying
	lottery tickets and hoping?

> Want to know the funny thing about that?  I did not even go to school
> for any of it.  I know plenty of well educated people paying through
> the nose for their computing needs because they just spent their money
> and trusted without really considering the nature of the deal.

	Are you saying you have no formal qualifications?

> People need educators to tell them things.  Most people won't actually
> pick up a book and read.  (Not sure why, I could never figure it out,
> but it's true.)

	Should I deduce that there aren't many libraries and bookstores
	where you are located?

> It's dead easy if one knows about it (OSS I mean), all you have to do
> is start with a distribution and begin asking questions one at a time
> until you have what you need and understand how to make use of it.
> 
> This is it for me.  Given your posting history and current nature on
> this thread, most of this will be lost on you I am sure.

	You are? I'm not.

> I am basically writing for the benefit of others better equipped to
> make use of my words here.
> 
> You have a nice day.

	I usually do. But thanks anyway.

-am	� 2003

On Sat, 02 Aug 2003 17:23:59 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:

> 	Whilst this is true, you don't really get the source with Solaris.

Sun does supply sources for the GPL code included with Solaris.  In fact
the sources for some non-GPL code is supplied; mozilla and apache come
immediately to mind.

> 	And, even if those programs didn't exist, there'd be other software
> 	the could be used to replace them (I keep thinking of SunView in
> 	comparison to X11, in particular, but there are plenty of sendmail
> 	replacements).

Quite true; indeed, I have replaced sendmail with postfix and bind-8 with
bind-9.
 
>> Since those examples of Open Source code seem to be reasonably capable and
>> stable, perhaps the study of them would benefit students.
> 
> 	Yes, but the issue boils down to which and how many. Rather than
> 	being all or most students its really only a very small proportion
> 	of students - those studying computer science or information
> 	technology at an advanced level.

I thought that was the group of students we were discussing.  Agreed that
students of economics or English lit would have little use for sources.

>       Modifying said code would be even
> 	worse (as I've pointed out elsewhere, you can't really change the
> 	fundamental protocols) and, based on the number of posts I see on
> 	a regular basis, compiling said code appears to be a complex task
> 	for a lot of people. Solutions like SunFreeware and Blastwave must
> 	exist for a reason, no?

Yeah, typing './configure --help' seems to be beyond the capabilities of
quite a few people.  I wonder how many of them are CS or CE graduates. 
One of the problems with such supposed professionals is that the
universities they attended teach nothing but Microsoft crap.

OTOH, those binaries are also useful for people who don't wish to take the
time and effort to compile the sources.

On Sat, 02 Aug 2003 16:29:32 +1000, Anthony Mandic wrote:

> 	I don't know what a hammer lane is but don't they book the
> 	drivers speeding behind them?

The hammer lane is the leftmost lane where drivers can figuratively "drop
the hammer" on the accelerator pedal.  And no, those people driving 15-20
mph above the lime are not ticketed.  There are too many of them doing
that.
 

> 	The thing I mentioned before that I found surprising was the
> 	amount of traffic on an interstate or freeway between Sacramento
> 	and the San Francisco Bay area. It was the middle of nowhere (no
> 	development apart from the road - just empty hills and fields),
> 	night time and a ribbon of headlights and taillights. Nearly choked
> 	with cars, I wondered where everyone was going and why there was
> 	no public transit. Anything like that in Sydney would be in town
> 	and around rush hour. You wouldn't see it out in the country.

With just a few exceptions "public transit" in California seems to be a
curse phrase.  I had a friend who lived in Redlands who thought nothing
odd about driving 65 mi to Los Angleles for dinner.

That particular stretch of highway is one of the most heavily used in the
country and a lot of that traffic was probably trucks.  Long haul truck
drivers have a 20 hour day; 0400 is just as convenient time to drive as
1600.

> 
> 	But speed was a factor. If he was going slower, he would have
> 	probably survived - or, at least, not killed anyone else. I
> 	seem to recall that there was an inquest but I don't recall the
> 	details now. His experience may also have been a factor but, yes,
> 	fatigue was probably the biggest factor. The same conditions
> 	would have been applied to all the race contestants though. I
> 	don't recall anyone else having a problem.

It was a -race-.  Even on closed circuit race tracks people get killed. 
The only victims were involved in the race and public outrage at the lack
of a speed limit because of those deaths is IMHO hardly justified.

In fact the only outrage I saw in the articles you referenced was WRT to
statehood for the NT and that writer used a number of other apparently
specious arguments against statehood.
 
>> Those people would have been just as dead if the accident had happened at
>> 70 mph, approx. the proposed speed limit of 110 kph.
> 
> 	Yes, probably but the circumstances would all have changed. If he
> 	was going slower overall, he wouldn't have gotten to that point
> 	until later, etc.

The same argument could be made for driving faster.  Had that driver got
to Alice Springs earlier he would probably have been less fatigued.  As
you pointed out yourself, the other contestants failed to kill themselves
or others and some of them were probably driving even faster than that
Ferrari driver.  The lack of a speed limit did not seem to have caused
those others to become homicidal maniacs.

On Fri, 01 Aug 2003 01:25:22 +1000
Anthony Mandic <pm@hotmail.com> wrote:

> Lord Isildur wrote:
> > 
> > Lincoln, actually, was the one who put the final nail in the
> > coffin of the Republic, and reaffirmed the nation's true character
> > to be Empire. 'For the people, etc' was rhetoric. Jefferson was the
> > true champion of the idealistic vision of america which for a while
> > was so firmly believed be so many people that it almost came into
> > existence. However, once matters got down to it.. all that was
> > pushed to the side.
> 
> 	Hmmm ... didn't Jefferson own slaves and Lincoln free them?
yes Jefferson was a slave owner, but Lincoln did not free them. He
issued the Emancipation Proclamation that freed slaves in the
Confederate States. He did not free any slaves in the states loyal to
the Union. 

-- 
Jerry Feldman <gaf-nospam-at-blu.org>
Boston Linux and Unix user group
http://www.blu.org PGP key id:C5061EA9
PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9

Jerry Feldman wrote:

> On Fri, 01 Aug 2003 01:25:22 +1000
> Anthony Mandic <pm@hotmail.com> wrote:
> 
> 
>>Lord Isildur wrote:
>>
>>>Lincoln, actually, was the one who put the final nail in the
>>>coffin of the Republic, and reaffirmed the nation's true character
>>>to be Empire. 'For the people, etc' was rhetoric. Jefferson was the
>>>true champion of the idealistic vision of america which for a while
>>>was so firmly believed be so many people that it almost came into
>>>existence. However, once matters got down to it.. all that was
>>>pushed to the side.
>>
>>	Hmmm ... didn't Jefferson own slaves and Lincoln free them?
> 
> yes Jefferson was a slave owner, but Lincoln did not free them. He
> issued the Emancipation Proclamation that freed slaves in the
> Confederate States. He did not free any slaves in the states loyal to
> the Union. 
> 
TAKE IT SOMEHWERE ELSE!

Uli Link wrote:

> How can one be sure that such peak optimized code works correct in all
> circumstances, since it cannot be debugged or tested in full. At highest
> level nearly all modules give a warning of the optimizer.
> No problem if mpg123 crashes one day with a special mp3, but most apps are a
> little more critical.

Automated testing can do a lot to help this. In an application I
wrote, 
http://atlc.sourceforge.net/
a number of programs are used to automatically generate bitmaps. The
md5 checksums of those bitmaps are then compared to what's expected. 

Once the bitmaps are generated, another floating point intensive
program performs a numerically intensive finite difference computation
on these bitmaps, generating some answers as text and some more
bitmaps. The numerical answers are checked against those expected. The
md5 checksums of the bitmaps are checked against those expected. 

When I say against 'what is expected' I'm comparing the results on a
number of platforms, running a number of different operating systems
using a number of different compilers at a number of different
optimisation levels. 

What I have found is that the bitmaps generated by the floating point
code do vary a little between platforms, but only by a single pixel
and then only by a single gray level. In other words, this appears to
be different rounding errors on the different platforms and is so
insignificant.

Of course, the above statergy is not foolproof, but is has helped me
find problems when building multi-threaded code on SPARCs running
Redhat Linux. Without some automated test method, that becomes next to
impossible. 

To make it easy to test on a number of platforms, I wrote a script 

http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/atlc/sources/teston

that copies the source to a remote computer, then runs configure;
make; make check on them. 

However, it is easy to not think to write a specific test - only after
you found a bug and realised my tests don't check for that problem. So
I add a test to the testsuite. 

There's not a lot more one can do I guess, unless of course you have
the time/budget to apply more formal methods, like you would if
writing code for a missile guidance system. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Jerry Feldman wrote:
> 
> On Fri, 01 Aug 2003 01:25:22 +1000
> Anthony Mandic <pm@hotmail.com> wrote:
> 
> > Lord Isildur wrote:
> > >
> > > Lincoln, actually, was the one who put the final nail in the
> > > coffin of the Republic, and reaffirmed the nation's true character
> > > to be Empire. 'For the people, etc' was rhetoric. Jefferson was the
> > > true champion of the idealistic vision of america which for a while
> > > was so firmly believed be so many people that it almost came into
> > > existence. However, once matters got down to it.. all that was
> > > pushed to the side.
> >
> >       Hmmm ... didn't Jefferson own slaves and Lincoln free them?
> yes Jefferson was a slave owner, but Lincoln did not free them. He
> issued the Emancipation Proclamation that freed slaves in the
> Confederate States. He did not free any slaves in the states loyal to
> the Union.
> 
You're going to need the 13th Amendment to the Constitution to
free all the slaves...

--
+----------------------------------------------------------------+
|   Charles and Francis Richmond     richmond at plano dot net   |
+----------------------------------------------------------------+

Dave Uhring wrote:

> Quite true; indeed, I have replaced sendmail with postfix and bind-8 with
> bind-9.

	How do you find postfix in comparison to sendmail, by the way?

> I thought that was the group of students we were discussing.  Agreed that
> students of economics or English lit would have little use for sources.

	Yes, it wasn't clear when "students" was mentioned. So I was angling
	it as all students. This was more to point out that the statements
	made should be more precise.

> Yeah, typing './configure --help' seems to be beyond the capabilities of
> quite a few people.  I wonder how many of them are CS or CE graduates.

	Yes, first you'd have to know the command. Obviously reading
	the README file is a bit of a struggle. You and I see many of
	the same posts, so we've both seen the problems people post
	about. I often wonder about the skill and experience level
	of the poster.

> One of the problems with such supposed professionals is that the
> universities they attended teach nothing but Microsoft crap.

	I know someone who dropped out of university in his first year
	(he claimed he didn't like studying). He didn't have an MCSE
	either but still got a job working with computers as a PC admin
	(Netware and MS). I, and others, didn't think too highly of him.
	I'm not sure how he got the job. Last I heard he was sacked and
	if running a restaurant now.

> OTOH, those binaries are also useful for people who don't wish to take the
> time and effort to compile the sources.

	Yes, that's very true. Someone else has gone to the trouble of
	getting it done right. It then becomes simply a matter of trust.
	This usually works well, but one would wonder what the person or
	people setting it up get out of it.

-am	� 2003

On Sun, 03 Aug 2003 13:49:06 +1000, Anthony Mandic wrote:

> 	OK, the right lane down here. I see lots of signs on the freeways
> 	and motorways that say "Slower vehicles use left lane".

Yes, you people do drive on the wrong side of the road.  Are those rules
about using the left lane enforced?  We have the same regs here in the US
but nobody ever gets a ticket for driving at less than the speed limit in
the fast lane.

>> With just a few exceptions "public transit" in California seems to be a
>> curse phrase.  I had a friend who lived in Redlands who thought nothing
>> odd about driving 65 mi to Los Angleles for dinner.
> 
> 	It does seem that way, doesn't it? I found the BART in SF and
> 	the metro trolley system in San Diego nice but rather limited.

No public transportation system can be designed to be effective and
financially viable in metro areas spread over hundreds or thousands of
square miles.  It's just not possible to provide enough rails or buses.
 
> 	It seemed to be mostly cars. I think it was the I-80 around
> 	Emeryville (where I could view it from) was certainly truck
> 	laden. What was odd was that there was a rail line just a bit
> 	further inland from the bay. That had some freight traffic,
> 	but I would wonder why it wasn't further utilised. Obviously
> 	price competitiveness isn't a major factor.

It's a matter of service.  Rail service simply sucks for most potential
freight customers.  Their sidings are not at many manufacturing and
distribution facilities; trucks can go anywhere a road exists.  The
railroads lose their cars; we had a hopper carload of cement shipped less
than 10 miles line of sight and the car was finally found about 800 miles
away.  A few months later they sent one of our cars to Iowa, about 400
miles away from our siding.  Once the hopper cars did finally arrive we
had to transport the cement by truck from the siding to our plant.

The railroad companies in the US abandoned their passenger service because
they could not meet their timetables.  Trucking companies guarantee
delivery within a limited time tolerance.

> 	Yes, I also found a site that had the govenment debates. But
> 	that was full of waffle so I didn't bother posting it. That
> 	site you mentioned was amusing though. We have plenty of rednecks
> 	here too (the majority of people preferred statehood because
> 	it meant they could manage their own affairs rather than let
> 	a contralised and distant government do it).

One always finds opposition to major changes, irrespective of goals or
benefit/cost ratios.  Consider the ongoing and moronic insistence of
corporate use of Windows.

> 	Yes, very true. Obviously driving experience is a key factor.

It's more than experience.  Dale Earnhart died despite his skill and that
of the other drivers on that track.  Accidents do happen.

Uli Link wrote:

> > Automated testing can do a lot to help this.
> 
> Agree. Let's take a closer look.
> 
> > In an application I wrote,
> > http://atlc.sourceforge.net/
> > a number of programs are used to automatically generate bitmaps. The
> > md5 checksums of those bitmaps are then compared to what's expected.
> >
> 
> That's even closer.

I think that is pretty close, given that 

a) I can visually check any bitmap at this point - they are simple,
consisting of at most 3 colours, with straight lines or circles.

b) Using Gimp to measure distances/diameters, makes it possible to
ensure the bitmaps are what's expected. Sure, Gimp has bugs, but the
chances of measuring a diameter in Gimp and it agreeing with what I
expect, means its highly likely there no errors. 

The 1st, 3rd and 4th examples here 
http://atlc.sourceforge.net/
show the sort of thing I generate. 

So then I generate an md5 checksum of those. That's the easy bit.

> > Once the bitmaps are generated, another floating point intensive
> > program performs a numerically intensive finite difference computation
> > on these bitmaps, generating some answers as text and some more
> > bitmaps. The numerical answers are checked against those expected. The
> > md5 checksums of the bitmaps are checked against those expected.
> >
> > When I say against 'what is expected' I'm comparing the results on a
> > number of platforms, running a number of different operating systems
> > using a number of different compilers at a number of different
> > optimisation levels.
> 
> Sounds like many, many hours of hard work needed before you know what is
> expected.

Yes, it is. I have probably spent 10x longer verifying the code than
writing it! Of course while verifying it, bugs became appernt, so the
two work in parallel to a certain extent. 

But one needs check carefully numerical/scientific programs as it's
easy to generate data that is wrong and you not know it. I know of a
student who was asked to re-submit his PhD after testing his code - he
never re-submitted.

If you build a word processor and it has a bug, it usually becomes
very apparent. Do the same for a numerical simulation and it's often
not the case. So careful testing is really important. 

I should have said I don't make a conserted effort to check the effect
of optimistation on other than SPARC. If someone runs it on another
platform and the results differ, then its up to them to play with the
optimisation. But I verify the code passes all tests on all platforms
I can. That has spotted a bug that was only apparent on 64-bit
platforms and another only apperent when the default char was
unsigned. 

First I set up some test cases, where exact analytical results can be
computed - see 
http://atlc.sourceforge.net/accuracy.html
if interested.

If the numerical results of the program are close to the exact
analytical results (1% errors are more than acceptable here. Better
accuracy takes more CPU time, and is of no practical use.) 

So that allows me to check the numerical results. So I generate
checksums for the important numerical results. 

The bitmaps generated here are less important, as they are only for
visual inspection. They don't contain any scientifically accurate
information.  Binary data is written to contain accurate data. Testing
that is more of a problem, and one I've not fully addressed. 

Once the bitmaps are generated, I generate checksums on a SPARC. Then
test on other systems to see if the checksums agree. When they don't,
I verify the the data is not too far off. In each case, only one pixel
has been affected, so its easy to accept 2 different checksums.
Obviously if there were lots, this would be impossible. 

> The other way round: expect _and_ accept insignificant differences. This
> means defining borders within a result is accepted.

Sort of. But by printing data to 4 significant figures with printf,
that is quite easily. For the bitmaps, I need a bit more work. 

> Real problem are often more complex as at a first thought. Needs a lot of
> investigation finding bitmaps similar but you cannot  easy compare by
> checksum. I know a lot of programmers avoiding floating point arithmetics
> for much easier getting exact measurable results.
> 

Fixed point is fine, but in my case not practical. 

> > However, it is easy to not think to write a specific test - only after
> > you found a bug and realised my tests don't check for that problem. So
> > I add a test to the testsuite.
> 
> Every engineer must lern to check the fault bandwith together with the
> measured results.

True. But sometimes you don't easily realise the fault bandwidth, or
know it is too large. If a program has 10 command line options, its
easy to check those 10 in isolation. More difficult to test all 2^10
possible combinations.


> It is impossible the prove a non-trivial programm as fault free, if built by
> faulty compilers, complex programming languages, and libs by human.

True, but by careful testing on one platform and checking the results
on others don't differ significantly, one has a high degree of
confidence in the results. 

> To my expierience the greatest problem in software projects is, that there
> are no expected results specified, before delivery to test. 

I avoid that. I can compute analytical results in some case, as
described at 
http://atlc.sourceforge.net/accuracy.html
So many tests are against these analytical results. 

> All this is complex enough with a single thread, and then you will need to
> test with mass data for deadlocks between threads.

One of the issues I came across was when using different algorithms
for single and multi-threaded cases. The multi-threaded case would
always converge to a very slightly different result. Not much, perhaps
0.001%, but when that happens on a lots of tests, it makes testing
impossible. The other issue was the results on multi-threaded case
depended on the number of threads run. Again, very small, but it would
have made sensible testing impossible, when numerous pixels in
numerous bitmaps would depend on how many CPUs the system had. 

Hence I changed the algorithm to one where the results are entirely
identical on single-threaded and multi-threaded and the results are
independent of the number of threads run. I loose a bit of speed
because of this, but its the only way to make testing practical. 
 
> Don't believe in missile guidance systems or a recent
> upperclass  automobile, not even those from Germany.

> Too much secrets and too much political-strategic-struggles concerning
> individuals career to give quality a chance. And (too much) 
> formalism  _can_
> kill the individuals responsibility for the quality of its work.
> The absence of those factors is perhaps the greatest influence on the
> quality of open source development.

I can believe that, although I've never worked in such an environment,
I have no experience of it.

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring wrote:

> Yes, you people do drive on the wrong side of the road.

	I think you mean the left side of the road.

> Are those rules about using the left lane enforced?

	Not that I'm aware of. Sometimes a leadfoot will honk you if you
	don't get out of their road.

> We have the same regs here in the US but nobody ever gets a ticket for
> driving at less than the speed limit in the fast lane.

	From what I've seen, every lane is the fast lane.

> >       It does seem that way, doesn't it? I found the BART in SF and
> >       the metro trolley system in San Diego nice but rather limited.
> 
> No public transportation system can be designed to be effective and
> financially viable in metro areas spread over hundreds or thousands of
> square miles.  It's just not possible to provide enough rails or buses.

	True but its generally run as a service rather than at a profit
	(I'm not sure about privately run rail networks. We have a
	combination of public and private buses here. The private ones
	aren't very prevalent but they seem to survive). There used to
	be a poster for an ad campaign a while back that claimed one
	train kept 2,000 cars off the road.

> It's a matter of service.  Rail service simply sucks for most potential
> freight customers.  Their sidings are not at many manufacturing and
> distribution facilities; trucks can go anywhere a road exists.

	It seems that its died out here. Rail freight used to be common
	but now you can't ship from the suburban stations and the freight
	depots are mostly gone. More trucks choke the roads as a result.

> The railroads lose their cars;

	I could argue that trucks can be hijacked but accidents are more
	common. It seems a lot of drivers are under pressure and take
	drugs to stay awake. They've been a few enquiries involving
	suspicious accidents involving trucks.

> we had a hopper carload of cement shipped less than 10 miles line of
> sight and the car was finally found about 800 miles away.

	Tell that to my luggage when I'm next at the airport!

> The railroad companies in the US abandoned their passenger service because
> they could not meet their timetables.  Trucking companies guarantee
> delivery within a limited time tolerance.

	Amtrak still seems to run. The track thru Emeryville and San Diego
	looked to be in good condition when I looked at it. So maintenance
	doesn't seem to be an issue. Perhaps the timetables just aren't
	realistic. But hasn't Amtrak's fortunes picked up since 9/11?

> One always finds opposition to major changes, irrespective of goals or
> benefit/cost ratios.  Consider the ongoing and moronic insistence of
> corporate use of Windows.

	Its interesting to think how it first got into business when
	there were already better solutions in existence.

> >       Yes, very true. Obviously driving experience is a key factor.
> 
> It's more than experience.  Dale Earnhart died despite his skill and that
> of the other drivers on that track.  Accidents do happen.

	Yes, I tried to convince Josh McKee of that some time ago. He
	just wouldn't accept it.

-am	� 2003

Dismayed by the lack of non-commercial licensing programs from
commercial UNIX vendors, I started this thread. Sun's so-called 'free
binary licence' I find a particularly annoying. 

There are now discussions of politics, driving safety, suitability of
Windoze, software testing, name calling etc etc. I do feel that these
topics might be entirely acceptable elsewhere, but not to the
newsgroups I posted this to. 

Sure, if someone else has something *contructive* to say about the
rights and wrongs of non-commercial licensing of UNIX software, than
that is fine, but I hope it's not asking too much to stop the rest,
and perhaps continue those in private or on a more suitable newsgroup.


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

On Sun, 03 Aug 2003 22:02:16 +1000, Anthony Mandic wrote:


> 	I'm doing an install at a client site next week. Setting up some
> 	Suns and installing Trend. Trend uses postfix by default and even
> 	has a web interface for it. The client wants to use sendmail, which
> 	I know better, with some add on for it that I don't know about. Trend
> 	doesn't have an interface for it. I was thinking it might be better
> 	for the client to go with Postfix since the admin side of things
> 	would be easier for them. I don't have an opportunity to test it
> 	though.

Postfix configuration is relatively straight forward, all plain text.  The
config files and samples are well commented.  It's probably worth your
while to install it on a test system.
 
It's primary advantage IMHO is that no root holes have yet been found in
it.  The design of postfix is such that the problems associated with
sendmail just cannot exist.  I see no negatives about postfix at
www.cert.org, only positive comments.

 
> 	Yes, but it does also depend on the student. When I first started
> 	doing CS, I didn't know one bit of code from another. Even when I
> 	did, something very sophisticated was beyond me. It took some time
> 	to become proficient.

The university I attended did not have a CS program at the time.  Hell,
that school did not even have a BS degree and I wound up with a BA instead. 
The only computer related course was Fortran II programming course offered
by the Math department.

>> While the Linux and BSD kernels may not be as well developed as the
>> commercial Unices, they still are excellent examples of reasonably well
>> written code.  Without NDAs those students can hardly study commercial OSs
>> even if the source is available.
> 
> 	I had the luxury of the Lyons' book.

Well, you questioned the usefulness of open source in a CS curriculum and
I pointed out several open source projects which might be studied
usefully.  By its very nature proprietary code such as Solaris and Windows
is generally inaccessible to the students.
 
> 	Agreed. So is the make command by the looks of it too ("I've got
> 	this source, now what do I do?").

Hell, I've given some of them the entire set of configure arguments and
they still couldn't get a set of binaries built.

>> Personal satisfaction and professional recognition, an amount of altruism.
> 
> 	Doesn't pay the bills though.

There are some exceptions.  Most Linux distributors employ coders whose
work becomes free.  Wasabi Systems employs a few NetBSD coders.  But
generally you are right about the bills getting paid.

On Sun, 03 Aug 2003 21:51:25 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:
> 
>> Yes, you people do drive on the wrong side of the road.
> 
> 	I think you mean the left side of the road.

Well, there are a right side and a wrong side.
 
> 	From what I've seen, every lane is the fast lane.

Not at "rush hour".

> 	True but its generally run as a service rather than at a profit
> 	(I'm not sure about privately run rail networks. We have a
> 	combination of public and private buses here. The private ones
> 	aren't very prevalent but they seem to survive). There used to
> 	be a poster for an ad campaign a while back that claimed one
> 	train kept 2,000 cars off the road.

The service still must be funded in some manner.  If the seats get filled
the passenger fares will accomplish that, but there are few routes with
sufficient ridership to be self maintaining.  The taxpayers must account
for the shortfall.
 
> 	It seems that its died out here. Rail freight used to be common
> 	but now you can't ship from the suburban stations and the freight
> 	depots are mostly gone. More trucks choke the roads as a result.

We still have a viable rail freight business serving the major cities. 
Intermodal freight is often used to serve smaller areas.  But unless the
railroad companies have changed their ways, the service is still
unreliable.

> 	I could argue that trucks can be hijacked but accidents are more
> 	common. It seems a lot of drivers are under pressure and take
> 	drugs to stay awake. They've been a few enquiries involving
> 	suspicious accidents involving trucks.

There are railroad accidents, too.  But you are right about the truck
drivers being under pressure.  Very few of them have unions to enforce
such things as service hours.
 
>> we had a hopper carload of cement shipped less than 10 miles line of
>> sight and the car was finally found about 800 miles away.
> 
> 	Tell that to my luggage when I'm next at the airport!

TWA never did find the last of my bags which they lost.  At least the
railroad found our cement hoppers.
 
> 	Amtrak still seems to run. The track thru Emeryville and San Diego
> 	looked to be in good condition when I looked at it. So maintenance
> 	doesn't seem to be an issue. Perhaps the timetables just aren't
> 	realistic. But hasn't Amtrak's fortunes picked up since 9/11?

They claim a higher ridership but a week or two ago Amtrak asked the Feds
for another billion in subsidy.

> 	Its interesting to think how it first got into business when
> 	there were already better solutions in existence.

IBM's gift of a monopoly in 1982.  At the time there was no other DOS for
the 8086/8088 processor, not even CP/M.  And, initially, the only
suppliers of PeeCees shipped MS-DOS with their machines.
 
> 	Yes, I tried to convince Josh McKee of that some time ago. He
> 	just wouldn't accept it.

He shows a remarkable ability to ignore good advice.  I don't see any of
his nonsense now; killfiles are useful.

Benjamin Gawert wrote:
> Rich Teer wrote:
 
>>>> Agreed - but it was you who made the first general statement.  :-)
>>>
>>> Which general statement did I make?
>>
>> This: "Surely every company has made a real cost/risk analysis
>> on Windows as the general-purpose OS too".
 
> If You'd try to read again You will recognise that this statement came from
> Mikko, not from me...

Besides,  I intended it as sarcasm. If you take it as a supposed
generalized factual statement, it is just plain false.

After all, it should be _obivous_ that, whatever the bigger players do,
there are _lots_ of small and medium _non-IT_ businesses where they just
buy whatever is cheap at the local shop. This then can cost them a
bundle in unexpected downtime and service costs ... or, in the case of
the man who did most of the plumbing work for me when I was fixing my
house, at least in having to take the system in for repairs and pick it
up again, rather often.

I usually advise them to at least buy something that has an on-site
3-year warranty if they ask me. Occasionally, some do. 

Things are indeed quite different there than here ... here, we worry
about things like the availability of identical spare parts 10 years 
from now.

>>> at the same time offers the highest flexibility because of the
>>> widest range of available OSes and software. Windows has it's
>>> glitches (like most other OSes, too btw) but for certain tasks it's
>>> still a good choice...

>> IMHO, the only time Windoze is a good choice is when there
>> is no other option.  Admittedly, I'm a biased techie...
 
> That's the problem ,-) In my job I can't allow me being personal biased
> toward or against a platform. I can do that when choosing what I use at
> home, but not when spending the companies money....

There's the thing about measuring field experience then too. You can't
really get that without personal bias, but regardless it _can_ be an
useful data source.

>>> Or they found out that Windows is simply the most cost-effective
>>> solution for their computing needs...

>> Perhaps, but unlikely.  More likely they didn't even bother
>> to find out about, or evaluate, the options.
 
> Or maybe the other options just catapulted themseves out of the race after
> they saw the initial and running costs? Especially small businesses can't
> really pay for Sun/HP/SGI/IBM Unix hardware, software and maintenance,
> leaving Linux/*BSD and Windows as the only alternatives. Now what if Your
> app isn't available for Linux and/or *BSD?

Um. Buy a Mac? 
Hey, don't laugh, the (admittedly few) Mac users that I've talked to
usually would appear to get by without spending much time or money on
that stuff ... even in business environments. (avg. headcount < 5 ...)

And some say that SCO could be an useful alternative. Well, said at
least an year ago. I don't have any personal experience on that myself.

Naturally, if you've gotten _this_ far, you can and naturally will
consider the system as a whole too - and if your main application only
runs on some particular OS, you usually use that.


-- 
Mikko Nahkola <mnahkola@trein.ntc.nokia.com>
#include <disclaimer.h>
#Not speaking for my employer. No warranty. YMMV. 

"Dr. David Kirkby" <drkirkby@ntlworld.com> wrote in message news:<3F2CE0CD.F7F6E3BF@ntlworld.com>...
> Uli Link wrote:

> > > However, it is easy to not think to write a specific test - only after
> > > you found a bug and realised my tests don't check for that problem. So
> > > I add a test to the testsuite.
> > 
> > Every engineer must lern to check the fault bandwith together with the
> > measured results.
> 
> True. But sometimes you don't easily realise the fault bandwidth, or
> know it is too large. If a program has 10 command line options, its
> easy to check those 10 in isolation. More difficult to test all 2^10
> possible combinations.

Or 2^100 combinations of 100 options ...

It is, however, often feasible to test all pairs of options, or even
all triples.  That'll catch the most offensive bugs.  The number of
testcases required for that grows with the log of the number of
options instead of exponentially.  I checked with one such tool. Given
10 options, it took 8 testcases to cover all settings of all pairs of
options and 18 to cover all triples.  100 options required 16
testcases for all pairs and 52 for all triples.

Dave Uhring wrote:

> Postfix configuration is relatively straight forward, all plain text.  The
> config files and samples are well commented.  It's probably worth your
> while to install it on a test system.

	Yes, it looked relatively straight forward. The client wanted to
	use Sendmail Switch though. When I left them, they were still trying
	to figure out how to change the port.

> The university I attended did not have a CS program at the time.  Hell,
> that school did not even have a BS degree and I wound up with a BA instead.

	Yes, I have one of those too.

> The only computer related course was Fortran II programming course offered
> by the Math department.

	I did some Fortran when I was studying architecture. Put me off
	for a while. When I did CS is was separate from maths but they
	had maths prerequisites - which annoyed me.

> Well, you questioned the usefulness of open source in a CS curriculum and
> I pointed out several open source projects which might be studied
> usefully.  By its very nature proprietary code such as Solaris and Windows
> is generally inaccessible to the students.

	Windows definitely is but you can get the Solaris source (or
	could) at an educational institution.

> Hell, I've given some of them the entire set of configure arguments and
> they still couldn't get a set of binaries built.

	I know. It gets beyond a joke sometimes.

> There are some exceptions.  Most Linux distributors employ coders whose
> work becomes free.  Wasabi Systems employs a few NetBSD coders.  But
> generally you are right about the bills getting paid.

	I would imagine the money from CD sales would go towards paying
	them. Which is fair enough.

-am	� 2003

Dave Uhring wrote:

> >       I think you mean the left side of the road.
> 
> Well, there are a right side and a wrong side.

	Or a left side and a wrong side.

> >       From what I've seen, every lane is the fast lane.
> 
> Not at "rush hour".

	Unless you're on a bike!

> The service still must be funded in some manner.  If the seats get filled
> the passenger fares will accomplish that, but there are few routes with
> sufficient ridership to be self maintaining.  The taxpayers must account
> for the shortfall.

	Yes, that's the way it works out here. Except that a few country
	routes were closed over the years and replaced with buses. City
	routes seem to be doing OK.

> There are railroad accidents, too.  But you are right about the truck
> drivers being under pressure.  Very few of them have unions to enforce
> such things as service hours.

	So its left to government regulation?

> >       Tell that to my luggage when I'm next at the airport!
> 
> TWA never did find the last of my bags which they lost.  At least the
> railroad found our cement hoppers.

	I had two incidents. United lost mine on the way to Boulder once.
	They eventually found them and made me pay to get them delivered
	(I didn't really have much choice). Another time, on the way back
	from Orlando, my bags got separated. Qantas delivered them for
	free the next day (and I didn't even fly with them).

> They claim a higher ridership but a week or two ago Amtrak asked the Feds
> for another billion in subsidy.

	So much for increased patronage.

> >       Yes, I tried to convince Josh McKee of that some time ago. He
> >       just wouldn't accept it.
> 
> He shows a remarkable ability to ignore good advice.  I don't see any of
> his nonsense now; killfiles are useful.

	He strikes me as just being a thinkhead. At least his big brother
	can deal with him if he needs to.

-am	� 2003

On Tue, 05 Aug 2003 23:33:26 +1000, Anthony Mandic wrote:

> 	Yes, it looked relatively straight forward. The client wanted to
> 	use Sendmail Switch though. When I left them, they were still trying
> 	to figure out how to change the port.

O'Reilly's "Sendmail", 30.8.38 and 30.4.1.2. But that smells of security
by obscurity.

As long as they're willing and able to chase the patches sendmail is a
good server for even a large company.  BTW postfix-1 just had a CERT
advisory but postfix-2 still is free of such problems.

> 	I did some Fortran when I was studying architecture. Put me off
> 	for a while. When I did CS is was separate from maths but they
> 	had maths prerequisites - which annoyed me.

Heh!  My major was physics; I would have not done well there without some
study of math.  I graduated 6 semester hours short of a second major in
math, then made that up when I returned to grad school.
 
> 	Windows definitely is but you can get the Solaris source (or
> 	could) at an educational institution.

I had forgotten about the edu availablity of Solaris, but still the open
source code trees are freely available to everyone and most of it is well
written.
 
> 	I would imagine the money from CD sales would go towards paying
> 	them. Which is fair enough.

The revenue from the sale of maintenance contracts helps there, too.

On Tue, 05 Aug 2003 23:42:55 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:

>> Well, there are a right side and a wrong side.
> 
> 	Or a left side and a wrong side.

Whatever you get used to.  Only 3 countries on the planet consider the
left side to be the right side.
 
>> Not at "rush hour".
> 
> 	Unless you're on a bike!

Bikes are of little utility when you live 30-35 miles from your workplace
and you have to deal with midwestern US weather.

>> There are railroad accidents, too.  But you are right about the truck
>> drivers being under pressure.  Very few of them have unions to enforce
>> such things as service hours.
> 
> 	So its left to government regulation?

Yep.  Truck drivers are permitted to drive for 10 hours; they have to take
2 one hour breaks, then rest for 8 hours.  This makes for a 20 hour
diurnal cycle which itself may lead to problems.

Many of those drivers maintain two logbooks to deceive the cops and
federal inspectors, especially the drivers pulling loads of fruit and
vegetables.
 
> 	I had two incidents. United lost mine on the way to Boulder once.
> 	They eventually found them and made me pay to get them delivered
> 	(I didn't really have much choice).

Service with a smile?

>       Another time, on the way back
> 	from Orlando, my bags got separated. Qantas delivered them for
> 	free the next day (and I didn't even fly with them).

Different culture.  The US carriers -used to- do that.  The ones in Oz
will get there, too.  I used to work for a company owned by Brambles;
their corporate attitude is just as mercenary as any in the US.
 
>> They claim a higher ridership but a week or two ago Amtrak asked the Feds
>> for another billion in subsidy.
> 
> 	So much for increased patronage.

Incompetent management too accustomed to sucking at the public's teat.  A
public corporation such as Amtrak has no need or incentive to make a
profit.

Uwe Wolfram wrote:
> Dave Uhring wrote:
> 
>>Whatever you get used to.  Only 3 countries on the planet consider the
>>left side to be the right side.
> 
> Although this thread runs on the wrong side of the road for quite a while
> I couldn't resist to count:
> 
> UK, New Zealand, Australia, Japan, Singapore, Malaysia, South Africa,
> Namibia, Zimbabwe, Zambia, India, Ireland, Cyprus and probably more.

If I can recall my history lessons, so did all of pre-Napoleon Europe. On horseback, the 
left is the "natural" side to be on for a right-handed person, as you face a potential 
attacker with your weapon-using arm. Apparently a left-handed miniature megalomaniac 
changed all that to his advantage. Hence the change to driving on the wrong (right) side 
of the road for most of Europe excluding the UK and Cyprus (not really European are we?).


-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GO/! d- s++:+ a+ C++(++++) US++++$ UB++ U*++ P+++ L+++ E--- W+++ N++ w--- O-
M+ V- PS+ PE+ Y+ PGP t+ 5++ X R* tv+ b+ DI++ D G e(*) h++/-- r+++ y?
------END GEEK CODE BLOCK------

-----------------------------------------------------
Bob Hoekstra: APL & Unix Consultant
Telephone:    +44 1483 771028
Mobile:       +44 7710 562345
Email:        Bob.Hoekstra@HoekstraSystems.ltd.uk
-----------------------------------------------------

Dave Uhring wrote:

> O'Reilly's "Sendmail", 30.8.38 and 30.4.1.2. But that smells of security
> by obscurity.

	Thanks. It the sandwich setup Trend's IMSS was designed with.
	Its meant to be done across 3 machines. Client wanted it done
	on one. Yeah, a real fun time.

> As long as they're willing and able to chase the patches sendmail is a
> good server for even a large company.  BTW postfix-1 just had a CERT
> advisory but postfix-2 still is free of such problems.

	Sendmail Switch turns out to be the commercial version of
	sendmail. They can chase support thru the vendor - which is
	good in a way.

> Heh!  My major was physics; I would have not done well there without some
> study of math.  I graduated 6 semester hours short of a second major in
> math, then made that up when I returned to grad school.

	I didn't really see the need for it when doing CS. It just seemed
	to be a lurk to keep the Maths dept. relevant.

> The revenue from the sale of maintenance contracts helps there, too.

	Yes, of course.

-am	� 2003

Dave Uhring wrote:

> >> Not at "rush hour".
> >
> >       Unless you're on a bike!
> 
> Bikes are of little utility when you live 30-35 miles from your workplace
> and you have to deal with midwestern US weather.

	Robert Persig didn't seem to mind. Although he probably stuck to
	back roads and summer.

> >       So its left to government regulation?
> 
> Yep.  Truck drivers are permitted to drive for 10 hours; they have to take
> 2 one hour breaks, then rest for 8 hours.  This makes for a 20 hour
> diurnal cycle which itself may lead to problems.

	They were doing all sorts of strange things here before it
	was regulated. Taking drugs mainly.

> Many of those drivers maintain two logbooks to deceive the cops and
> federal inspectors, especially the drivers pulling loads of fruit and
> vegetables.

	Because they're perishable?

> >       I had two incidents. United lost mine on the way to Boulder once.
> 
> Service with a smile?

	Why else fly United? (Apart from the fact that it make it easier
	for me to use a US carrier if I go on to fly to an internal destination
	after arriving there. I get all the board pass nonsense and bag
	checking done here. Much easier.)

> Different culture.  The US carriers -used to- do that.  The ones in Oz
> will get there, too.  I used to work for a company owned by Brambles;
> their corporate attitude is just as mercenary as any in the US.

	Yes, the only really big Aust. companies got that way because of
	their managers. A few fold really quickly once the top changes.

> Incompetent management too accustomed to sucking at the public's teat.  A
> public corporation such as Amtrak has no need or incentive to make a
> profit.

	Hmmm ... I always thought it was a listed company.

-am	� 2003

On Wed, 06 Aug 2003 13:22:18 +1000, Anthony Mandic wrote:

> 	Sendmail Switch turns out to be the commercial version of
> 	sendmail. They can chase support thru the vendor - which is
> 	good in a way.

It certainly is.  The people who supply an insecure app -should- be made
to be responsible for it.
 
> 	I didn't really see the need for it when doing CS. It just seemed
> 	to be a lurk to keep the Maths dept. relevant.

Presumably the study of linear algebra and numerical analysis is a useful
prerequisite.

On Wed, 06 Aug 2003 13:35:53 +1000, Anthony Mandic wrote:

> 	They were doing all sorts of strange things here before it
> 	was regulated. Taking drugs mainly.

There used to be a big problem with drivers taking drugs here, but that
seems to have abated.  Strict regulation of driving licenses was a factor
there.
 
>> Many of those drivers maintain two logbooks to deceive the cops and
>> federal inspectors, especially the drivers pulling loads of fruit and
>> vegetables.
> 
> 	Because they're perishable?

Primarily.  That and the competition for hauling such freight.  Big rigs
earn little while their drivers are sleeping in them.

> 	Why else fly United? (Apart from the fact that it make it easier
> 	for me to use a US carrier if I go on to fly to an internal destination
> 	after arriving there. I get all the board pass nonsense and bag
> 	checking done here. Much easier.)

No other US carriers serving Oz?  A carrier that made me pay for the
return of my lost baggage would be unlikely to sell me another ticket.

> 	Yes, the only really big Aust. companies got that way because of
> 	their managers. A few fold really quickly once the top changes.

Perhaps corporate nastiness is required for growth.  It does not guarantee
corporate survivability, however.
 
>> Incompetent management too accustomed to sucking at the public's teat.  A
>> public corporation such as Amtrak has no need or incentive to make a
>> profit.
> 
> 	Hmmm ... I always thought it was a listed company.

Amtrak was formed as a government corporation from the surviving passenger
divisions of all the major railroads.

Il giorno Tue, 5 Aug 2003, Uwe Wolfram cos=EC ha scritto:

|Dave Uhring wrote:
|>> Well, there are a right side and a wrong side.
|> >
|> > =09Or a left side and a wrong side.
|>
|> Whatever you get used to.  Only 3 countries on the planet consider the
|> left side to be the right side.
|
|Although this thread runs on the wrong side of the road for quite a while
|I couldn't resist to count:
|
|UK, New Zealand, Australia, Japan, Singapore, Malaysia, South Africa,
|Namibia, Zimbabwe, Zambia, India, Ireland, Cyprus and probably more.

  I think Thailand is also part of the pack.
  This OT amuses me!  :-)


  Sandro


--=20
Bellum se ipsum alet
       La guerra nutre se stessa

Livio, Ab urbe condita, XXXIV,9

Alessandro Selli wrote:

> |UK, New Zealand, Australia, Japan, Singapore, Malaysia, South Africa,
> |Namibia, Zimbabwe, Zambia, India, Ireland, Cyprus and probably more.
> 
>   I think Thailand is also part of the pack.

	Hmmm ... in that case Burma might be too. What about Pakistan?

>   This OT amuses me!  :-)

	Yep.

-am	� 2003

Dave Uhring wrote:

> >       I didn't really see the need for it when doing CS. It just seemed
> >       to be a lurk to keep the Maths dept. relevant.
> 
> Presumably the study of linear algebra and numerical analysis is a useful
> prerequisite.

	If you're interested in those areas. I seem to recall a reader
	from the maths dept. because a database lecturer in CS. Didn't
	seem to know much.

-am	� 2003

On Wed, 06 Aug 2003 23:06:22 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:

>> Presumably the study of linear algebra and numerical analysis is a useful
>> prerequisite.
> 
> 	If you're interested in those areas. I seem to recall a reader
> 	from the maths dept. because a database lecturer in CS. Didn't
> 	seem to know much.

At one time those areas were the primary use for computers.  Such
prerequisites are probably an anachronism these days.  Not many businesses
have a need for computing ballistic trajectories.

On Wed, 06 Aug 2003 23:11:59 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:

>> Primarily.  That and the competition for hauling such freight.  Big rigs
>> earn little while their drivers are sleeping in them.
> 
> 	They don't share the rig with other drivers? As far as I know, they
> 	don't do it here but they do do it with taxis.

I don't know the exact numbers but it seems like most rigs are driven by
solo drivers rather than teams.  Many trucks are operated by their owners
and working under contract to carrier companies.  It's a problem finding a
co-driver one can live with 24/7.
 
> 	Yes. United was one of the better ones. I joined their mileage
> 	program fairly early on. Got too many points in them now.

Those bonus miles really do lock in customer loyalty.  ISTR that some of
the carriers have merged their mileage programs.  Might be worth checking.
 
> 	I was watching a documentary on the WorldCom collapse last
> 	night. I was surprised the sharks got away relatively unscathed.

CNN's Lou Dobbs airs the scoreboard daily.  So far only one of those
scumbags has gone to the slammer.  But "justice" does grind inexorably if
slowly; some of the others may yet see hard time.
 

>> Amtrak was formed as a government corporation from the surviving passenger
>> divisions of all the major railroads.
> 
> 	Maybe its time to sell it off.

No viable buyers exist even if the price were zero.  Amtrak was formed in
the first place because maintaining railway passenger service is in the
national interest.  The rail carriers were about to discontinue that
service.

Dave Uhring wrote:

> I don't know the exact numbers but it seems like most rigs are driven by
> solo drivers rather than teams.  Many trucks are operated by their owners
> and working under contract to carrier companies.  It's a problem finding a
> co-driver one can live with 24/7.

	It might be better to have the rigs owned by companies and do the
	same thing bus companies do. Have depots where the drivers can do
	a shift change with someone else.

> Those bonus miles really do lock in customer loyalty.  ISTR that some of
> the carriers have merged their mileage programs.  Might be worth checking.

	Well, I could fly with anyone in Star Alliance. But to the US and
	back, its only a few of them and United is the only one where I
	get the benefit of bag check thru and boarding passes issued up
	front.

> CNN's Lou Dobbs airs the scoreboard daily.  So far only one of those
> scumbags has gone to the slammer.  But "justice" does grind inexorably if
> slowly; some of the others may yet see hard time.

	It would be interesting to see what ends up happening with Enron.

> >       Maybe its time to sell it off.
> 
> No viable buyers exist even if the price were zero.  Amtrak was formed in
> the first place because maintaining railway passenger service is in the
> national interest.  The rail carriers were about to discontinue that
> service.

	Well, there's only one real way to find out. Privatise it and
	do a float. Maybe it would sell now.

-am	� 2003

Poor Jefferson, he was a deeply troubled man.

It might have been Lincoln who laid the final nail, I won't argue that, but
it was Jefferson who first set us on that road with the Louisiana Purchase.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

Dave Uhring wrote:

> Well, that would require having terminals located less than 10 hours apart
> and covering the entire country.  Only the very largest carriers have that
> capability, which means the smaller ones would die.  Then we would have
> poor service and high rates.  Competition is vital to the industry.

	Given the amount of haulage, I can't seeing it being that bad.
	Not sure what percantage is long distance but it could be looked
	at with some of the more popular long routes. Drivers could
	form cooperatives too.

> One carrier, Werner, has a better solution.  Equip every tractor with
> Qualcomm GPS units which transmit the rigs' locations and velocities to
> the home office at regular intervals.  The driver's logs are maintained
> within the carrier's computer systems.

	That only really handles the logbook side of things.

> But it is that baggage service which caused your problem :-)

	Yes, that fact wasn't lost on me. Once in a blue moon isn't
	a bad record (I was going thru Denver too).

> >       It would be interesting to see what ends up happening with Enron.
> 
> Several of their execs have been indicted.  Justice is generally not too
> swift here and sometimes is non-existent.  See the outcome of the
> anti-trust case against MSFT.

	Yep, much the same as here. Crime may not really pay but
	white collar crime has better rates when it does.

> >       Well, there's only one real way to find out. Privatise it and
> >       do a float. Maybe it would sell now.
> 
> Would not work.  Maintaining a viable passenger rail service remains in
> the national interest.  The company could not be allowed to fail, thus no
> incentive for its management to earn profit.

	What if its employee-owned? Or just the popular routes privatised.
	I gather there are some (Orient Express are doing something out
	here but its too early to tell if it'll succeed).

-am	� 2003

On Thu, 07 Aug 2003 23:30:08 +1000, Anthony Mandic wrote:

> 	Given the amount of haulage, I can't seeing it being that bad.

There remains the need for the ability of small carriers to grow large. 
Your proposal would doom the small carriers and establish monopoly in the
industry.

> 	Not sure what percantage is long distance but it could be looked
> 	at with some of the more popular long routes. Drivers could
> 	form cooperatives too.

A large percentage of OTR freight is long haul.  California's farms
provide fresh fruit and vegetables for most of the country, and as I
mentioned earlier, transportation by rail is unreliable.

The US also has a high volume of trade with western Pacific nations and
much of the resulting cargo is transported across the US by truck.  There
is a high volume of traffic from eastern seaports to the remainder of the
US.  There also is a high volume of traffic between Canada and Mexico.

> 	That only really handles the logbook side of things.

Plus company oversight.  But the logbook is the critical document since it
makes it impossible for the driver to fake his hours of service to law
enforcement.

> 	Yep, much the same as here. Crime may not really pay but
> 	white collar crime has better rates when it does.

The best paying crime and least likely to be paid for is hacking banks and
other financial institutions.  Almost no bank is willing to publicly admit
that its security is nonexistent.

> 	What if its employee-owned? Or just the popular routes privatised.
> 	I gather there are some (Orient Express are doing something out
> 	here but its too early to tell if it'll succeed).

That just would not work with the -only- passenger carrier.  The incentive
to earn profit still would not exist.

Additionally the Amtrak routes which are popular and presumably profitable
provide the subsidy to maintain the less popular and less profitable
routes.  Spinning off the Miami-Boston routes to create a profitable
company would mean even higher Federal subsidies to the remainder of the
network.  Moreover there is no guarantee that such a spinoff would produce
a self-sustaining carrier.  The Long Island Railroad was unprofitable
before Amtrak was created, and LIRR probably has the highest ridership of
any passenger line in the country.

In comp.unix.solaris Peter da Silva <peter@abbnm.com> wrote:
>   Poor Jefferson...

Oh, so sorry, but the Tsar has redacted your post as off-topic.

Go back to your Useless2 sandbox.

On Thu, 07 Aug 2003 19:02:45 +0000, Cypherpun wrote:

> In comp.unix.solaris Peter da Silva <peter@abbnm.com> wrote:
>>   Poor Jefferson...
> 
> Oh, so sorry, but the Tsar has redacted your post as off-topic.
> 
> Go back to your Useless2 sandbox.

The entire thread is OT; it started that way.  But as someone who makes a
lifestyle of personally attacking other subscribers to the NG, you are the
pot calling the kettle black.

Anthony Mandic <qo@hotmail.com> wrote in message news:<3F2B7E75.AE80EACE@hotmail.com>...

(fucked up discussion snipped)

Look, I hold the opinion that Linux is a fine operating system to
learn computing on.  You can nitpick the benefits of Open code vs
closed code all you want, but it boils down to this:

Open Code is code that people can use how they see fit to use it. 
There are no use restrictions on Open Code.  Closed binary code and
systems all have various use restrictions attached that limit what you
can do, plus you get to pay for the experience.

Everyone has the right to learn.  Open code assures that right will
continue.  Having a well established body of Open code also does a lot
to insure the right to learn can happen at the lowest cost possible
for those who need it.

When you learn how to make use of binary code, your knowledge becomes
dependant on that particular piece of code.  If you want to exercise
what you have learned, you or sombody else must pay the creator of the
code in order to proceed.  If you cannot use what you know without
payment, who owns the knowledge?

When one learns about Open code, though binaries or source or a
combination of both, their knowledge is not dependant on any one
entity.  So, exercising said knowledge becomes a simple matter of will
--not payment.  Consider again who owns what in this case.

What happens when said student wants to try something new?  Maybe this
something will end up being disruptive.  (To the market)

The binary only people will not have an incentive to promote the
action because it could contribute to lower sales; thus, they will
resist it and act against it.  These actions could include preventing
you from using the very tools that you need to compete!?!

The Open people could care less.  If its really better, others will
use it.  If not, others can learn from it.  Either way, everyone will
eventually gain benefit from it, but nobody is prevented from doing
what they want with it.

Maybe I can say it one more way that is a bit simpler.  If someone
learned math using only calculators, they would have to own a
calculator in order to perform calculations.  To them, doing more
advanced calculations would involve the purchase of a better
calculator.

If someone else learned both the calculator way, and the pencil and
paper way, they have the option of using calculators, but are free to
perform any calculation they want without as well.

Considering the licensing issue that started this thread, my statement
about Linux and its utility for education has merit enough without all
the little annoying side arguements.

If the Dr. really wants to compute for lowest cost, he would be wise
to start investing into Open Source now.  He should further favor
applications and hardware that make best use of said Open systems for
the best return and flexibility in the future.

Given the state of things today, Linux appears to be a very good
choice in that regard with the BSD operating systems being another.

The rest is minor bullshit that I do not have time for.

In comp.unix.solaris Dave Uhring <daveuhring@yahoo.com> wrote:
>  Cypherpunk wrote:
>  > Peter da Silva <peter@abbnm.com> wrote:
>  >>   Poor Jefferson...
>  > 
>  > Oh, so sorry, but the Tsar has redacted your post as off-topic.
>  > 
>  > Go back to your Useless2 sandbox.
>   
>   The entire thread is OT; it started that way.

I was not complaining that he posted off-topic, despite it
looking that way superficially.

You'll have to dig deeper with Groups Google.

Try: tsar "Peter da Silva"

The first result has enough of the info, including
the link to the failed attempt to take over Usenet.

Of course, that's a whole other flame war. ;-)

Doug Dingus wrote:
> 
> Anthony Mandic <qo@hotmail.com> wrote in message news:<3F2B7E75.AE80EACE@hotmail.com>...
> 
> (fucked up discussion snipped)
>
Good. 
 
> Look, I hold the opinion that Linux is a fine operating system to
> learn computing on.  You can nitpick the benefits of Open code vs
> closed code all you want, but it boils down to this:
> 
> Open Code is code that people can use how they see fit to use it.
> There are no use restrictions on Open Code.  Closed binary code and
> systems all have various use restrictions attached that limit what you
> can do, plus you get to pay for the experience.

> If the Dr. really wants to compute for lowest cost

Cost is not the over-riding issue for me. If it was, I'd steer well
away from UNIX boxes from HP, Sun, IBM, SGI etc. Even used, they are
expensive compared to PCs. So cost is not my *only* concern.

> he would be wise
> to start investing into Open Source now.  


> He should further favor
> applications and hardware that make best use of said Open systems for
> the best return and flexibility in the future.

I do use open-source. I have here systems running Linux (PC and
SPARC), NetBSD and OpenBSD. However, to be honest, I've never looked
at the source code for the operating system, which is I guess the same
for 99% of UNIX users. I see the operating system as a tool to do
something constructive with, like I consider a saw and chisel tools to
do constructive work with. I've never had much interest in making a
saw or a chisel.

I have also developed open-source code
http://atlc.sourceforge.net/
which I've released under the GPL. I also often build open-source code
on both commercial and free operating systems. 
 
> Given the state of things today, Linux appears to be a very good
> choice in that regard with the BSD operating systems being another.

I accept Linux and *BSD are useful. However, there is a serious
problem if these are used in isolation, which is one reason I'd like
to see free or cheap versions of the major UNIX operating systems and
development tools for non-commerical use. The problem is that a high
percentage of developers develop code on a machine which is 
a) 32-bit 
b) big-endian
c) treats char as 'signed' by default. 
d) runs only one compiler (gcc), assembler (gas) and linker (GNU
linker)
e) has only one version of make installed (GNU make).
f) Uses the same compiler when compiling single-threaded and
multi-threaded code.
g) Has local files in /usr/local/*
h) Has a certain way of maintaining it
i) etc etc etc.

95% of those machines are PC, and a very large proportion run Linux,
with most being Redhat. Hence people tend to produce unportable code.
Now whether you approve or not, a large number of commercial operating
systems are run, so those making unportable code are a nuisance. Once
a program gets big, or starts to use lots of Linux specific kernel
details, it becomes a nightmare to port. 

In contrast, I take a very different attitude. I accept there are
several UNIX and UNIX-like operating systems. I try to make my code
work on them all, using both gcc and the manufacturer's compilers. So
'atlc'
http://atlc.sourceforge.net/
gets tested on Linux (SPARC + PC), NetBSD, OpenBSD, Solaris 9, Tru64
5.1B, HP-UX 11, AIX 5.2 and IRIX 6.5.16. Where portability issues
exist I try to correct them, or at least put details in the README's. 

I'd like to read the temperature of my SCSI disks. There are several
Linux utilities to do this, but not one will work on SPARC. Yet the
author of 'cdrecord' has clearly demonstrated it's possible to write
low-level SCSI code that is portable.

I also learn to use different systems.  For an administrator, the
commercial operating systems are very different. Despite the fact a
large number of people may use those operating systems, fewer will get
the chance to be root, unless they get a machine at home. Only then
can they learn the OS more fully. 

So I think open-source is good. I usually prefer to use open-source
tools. Despite the fact I don't work in the computer group, my UNIX
knowledge of systems like Solaris has been invaluable at work. I've
recently been arguing we should use amanda rather than the commercial
Veritas for backups. The system admin is not too keen, but the head of
department, who holds the purse strings, is likely to force the
sysadmin to use amanda. (Let's not start a discussion about the
benifits/pitfalls of amanda vs Veritas. That one has been discussed
elsewhere, and is nothing to do with what I started this thread with.)

So my feelings are that open-source is good, but commercial UNIX
systems exist and will continue to do so for several years at least.
Interested parties should have the option of using those for
non-commercial use. That is good for the individual and is good I feel
of benefit for the vendor too. 

Home users are never going to keep Sun, IBM, HP, SGI etc in business,
but very often they do spend money. I've bought personally several
*new* Sun items, from Sun resellers, totalling perhaps 400 UK pounds
($650). Several home-users have bought new Sun workstations for home
use. So again, by making Solaris available to home users, Sun will
often get some extra $$. It won't make them a fortune, but will
perhaps keep their balance sheet slightly more healthy, so reducing
the number of employees they have made redundant in the past, due to
poor sales. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring wrote:

> There remains the need for the ability of small carriers to grow large.
> Your proposal would doom the small carriers and establish monopoly in the
> industry.

	I can't see it becoming a monopoly. Cartels could possibly form
	but there's always government to regulate things.

> A large percentage of OTR freight is long haul.  California's farms
> provide fresh fruit and vegetables for most of the country, and as I
> mentioned earlier, transportation by rail is unreliable.
> 
> The US also has a high volume of trade with western Pacific nations and
> much of the resulting cargo is transported across the US by truck.  There
> is a high volume of traffic from eastern seaports to the remainder of the
> US.  There also is a high volume of traffic between Canada and Mexico.

	It would be interesting to see what happens when petroleum based
	fuels eventually run out. I'm sure hydrogen fuel would make it by
	then but who's to say.

> The best paying crime and least likely to be paid for is hacking banks and
> other financial institutions.  Almost no bank is willing to publicly admit
> that its security is nonexistent.

	Do the exact details really need to come out in court? Or can it
	be kept in camera?

> That just would not work with the -only- passenger carrier.  The incentive
> to earn profit still would not exist.

	Unless there were no subsidies?

> Additionally the Amtrak routes which are popular and presumably profitable
> provide the subsidy to maintain the less popular and less profitable
> routes.  Spinning off the Miami-Boston routes to create a profitable
> company would mean even higher Federal subsidies to the remainder of the
> network.  Moreover there is no guarantee that such a spinoff would produce
> a self-sustaining carrier.  The Long Island Railroad was unprofitable
> before Amtrak was created, and LIRR probably has the highest ridership of
> any passenger line in the country.

	That's interesting. And with Mugger trying to rob one of the subway
	vending machines (according to a post of his some months back), it
	doesn't help its profitability. I wonder what patronage falloff
	would be like as fares went up? I seem to recall that there's no
	service between JFK airport and NYC (or from there to Newark for
	that matter but they probably don't extend their lines to New Jersey
	anyway). I would have tried it when I was there if it was available
	(I do it here and its far cheaper than a taxi or shuttle - but not
	entirely as convenient).

-am	� 2003

Dave Uhring wrote:
> 
> On Thu, 07 Aug 2003 19:02:45 +0000, Cypherpun wrote:
> 
> > In comp.unix.solaris Peter da Silva <peter@abbnm.com> wrote:
> >>   Poor Jefferson...
> >
> > Oh, so sorry, but the Tsar has redacted your post as off-topic.
> >
> > Go back to your Useless2 sandbox.
> 
> The entire thread is OT; it started that way.  But as someone who makes a
> lifestyle of personally attacking other subscribers to the NG, you are the
> pot calling the kettle black.

	He's just jealous because they wouldn't let him on Usenet 2.
	His rabbit, however, had gained some modicum of popularity.

-am	� 2003

Dave Uhring wrote:
> 
> On Fri, 08 Aug 2003 00:17:09 +0000, Cypherpun wrote:
> 
> > I was not complaining that he posted off-topic, despite it
> > looking that way superficially.
>
> You are ignoring my statement about personal attacks.  It's time to stop
> that.  Even if this particular thread is OT, the NG is a technical NG and
> such crap has no place here.

	Well said. But as we've seen with many of my replies to him
	and his followups, he's too thick to get the hint.

> If you don't like the opinions of your enemies, then killfile them.

	Then he'd have to put everyone in his killfile.

-am	� 2003

On Fri, 08 Aug 2003 13:18:45 +1000, Anthony Mandic wrote:

> 	I can't see it becoming a monopoly. Cartels could possibly form
> 	but there's always government to regulate things.

Government regulation of commerce has many problems associated with it. 
The system works well as it is, although some fine tuning of the system
could be done.  Requiring Qualcomm equipment would be a good start.
 

> 	It would be interesting to see what happens when petroleum based
> 	fuels eventually run out. I'm sure hydrogen fuel would make it by
> 	then but who's to say.

Good point.  We may have to re-establish the draft to rebuild our army, at
least to postpone that eventuality.


> 	Do the exact details really need to come out in court? Or can it
> 	be kept in camera?

If the bank can claim that the hack was a terrorist act then USA PATRIOT
will ensure that the details are kept in camera.  Star chambers seem to be
in our future.

>> That just would not work with the -only- passenger carrier.  The incentive
>> to earn profit still would not exist.
> 
> 	Unless there were no subsidies?

No subsidies implies the eventual cessation of the service given the
present management.  This would be unacceptable given the "national
interest" in maintaining a rail passenger service.
 

> 	That's interesting. And with Mugger trying to rob one of the subway
> 	vending machines (according to a post of his some months back), it
> 	doesn't help its profitability.

A trivial concern considering the mismanagement of the railroad companies
in the past century and the losses suffered from that.


>       I wonder what patronage falloff would be like as fares went up?

Probably minimal on the routes between Washington and Boston.  That area
has a high population density and poor facilities for automobile traffic.

>       I seem to recall that there's no
> 	service between JFK airport and NYC (or from there to Newark for
> 	that matter but they probably don't extend their lines to New Jersey
> 	anyway). I would have tried it when I was there if it was available
> 	(I do it here and its far cheaper than a taxi or shuttle - but not
> 	entirely as convenient).

The only time that NYC was my airline destination I rented a car to drive
to Connecticut for my work.  Since I needed local transportation there my
employer paid for the car, and did so quite happily.  The US Air Force was
paying cost plus on that contract.

In article <3F330255.B90F491B@ntlworld.com>,
 "Dr. David Kirkby" <drkirkby@ntlworld.com> wrote:

: The problem is that a high
: percentage of developers develop code on a machine which is 
: a) 32-bit 
: b) big-endian

x86 is little-endian, actually as is the Itanic and x86-64. SPARC is big-endian, 
68k is big-endian, MIPS is bi-endian, PPC is bi-endian (though if you ever use 
it in LE mode, be prepared for pain), ARM is bi-endian, Alpha is little-endian. 
I don't know what endian Clipper is, but Clipper is dead, so there's not much 
point worrying.

Thanks primarily to Debian, endian errors are getting very rare, and 
pointer-size errors are slowly on their way out too.

[...]
: I'd like to read the temperature of my SCSI disks. There are several
: Linux utilities to do this, but not one will work on SPARC. Yet the
: author of 'cdrecord' has clearly demonstrated it's possible to write
: low-level SCSI code that is portable.
[...]

Possible but difficult. Every single platform uses a different API to send 
commands to and from drives. cdrecord gets around this by (no surprise) writing 
different drive interface code for every single platform.

One of these days I might port smartmontools or some other package to IRIX, but 
I don't have a Sun so you're out of luck there.


Cheers - Tony 'Nicoya' Mantler :)

-- 
Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@apia.dhs.org
--  http://www.apia.dhs.org/  --  http://nicoya.feline.pp.se/  --

"Dr. David Kirkby" <drkirkby@ntlworld.com> wrote in message news:<3F330255.B90F491B@ntlworld.com>...
> Doug Dingus wrote:
> > 
> > Anthony Mandic <qo@hotmail.com> wrote in message news:<3F2B7E75.AE80EACE@hotmail.com>...
> > 
> > (fucked up discussion snipped)
> >
> Good. 

:)

(snip)
> I do use open-source. I have here systems running Linux (PC and
> SPARC), NetBSD and OpenBSD. However, to be honest, I've never looked
> at the source code for the operating system, which is I guess the same
> for 99% of UNIX users. I see the operating system as a tool to do
> something constructive with, like I consider a saw and chisel tools to
> do constructive work with. I've never had much interest in making a
> saw or a chisel.

Figured you did.  I agree with your tool analogy.  It does break down
though when you compare the obvious and open nature of the saw to the
closed and murky nature of binary only software.  However, my point
was not to exclude commercial UNIXes, only to make sure Open code was
part of the package.  In this, I believe we agree and the analogy is
sound given the Open package will indirectly educate whoever is
interested how exactly that 'saw' does its thing.

(snip)
> > Given the state of things today, Linux appears to be a very good
> > choice in that regard with the BSD operating systems being another.
> 
> I accept Linux and *BSD are useful. However, there is a serious
> problem if these are used in isolation, which is one reason I'd like
> to see free or cheap versions of the major UNIX operating systems and
> development tools for non-commerical use. The problem is that a high
> percentage of developers develop code on a machine which is 
> a) 32-bit 
> b) big-endian
> c) treats char as 'signed' by default. 
> d) runs only one compiler (gcc), assembler (gas) and linker (GNU
> linker)
> e) has only one version of make installed (GNU make).
> f) Uses the same compiler when compiling single-threaded and
> multi-threaded code.
> g) Has local files in /usr/local/*
> h) Has a certain way of maintaining it
> i) etc etc etc.

Agree with all of these in principle.  Multi-os exposure is important.
 I use IRIX often with Linux.  The differences between the two are
annoying, but educational at the same time.  There are plenty of
things Linux does not do *well* yet that commercial UNIX does do. 
When people learn one platform, open or closed, they learn that
platform.  Once diversity enters the mix, they learn the art of
computing itself.

> 
> 95% of those machines are PC, and a very large proportion run Linux,
> with most being Redhat. Hence people tend to produce unportable code.
> Now whether you approve or not, a large number of commercial operating
> systems are run, so those making unportable code are a nuisance. Once
> a program gets big, or starts to use lots of Linux specific kernel
> details, it becomes a nightmare to port.

To me, the hardware problem is a big one because I generally dislike
PC hardware.  However, history shows us, simple yet adaptable systems
tend to be the winners.  The PC is very adaptable and cheap. Hard to
beat that combination however much of a kludge it might be currently.

Unportable code happens when developers focus on a platform.  In this,
my favor is split.  On one hand, the attention Open Source, Linux in
particular, is getting is a damn good thing.  On the other, porting
becomes more of a problem as it gains mindshare.

I am not sure a good solution exists for this; thus, your position on
many platforms is a sound one.  (Not sure it will win the war though
it will generate people who will continue to understand the issue.)
 
(snip)

> 
> I also learn to use different systems.  For an administrator, the
> commercial operating systems are very different. Despite the fact a
> large number of people may use those operating systems, fewer will get
> the chance to be root, unless they get a machine at home. Only then
> can they learn the OS more fully. 

There is always ebay...  Seriously, this is true.  As is your point
about free licenses.  I am just not sure things will change however
much you and I would like them to.  Creative students can network
around and get what they want behind the scenes, creative educators
have more limitations I am afraid...

(snip)

> Home users are never going to keep Sun, IBM, HP, SGI etc in business,
> but very often they do spend money. I've bought personally several
> *new* Sun items, from Sun resellers, totalling perhaps 400 UK pounds
> ($650). Several home-users have bought new Sun workstations for home
> use. So again, by making Solaris available to home users, Sun will
> often get some extra $$. It won't make them a fortune, but will
> perhaps keep their balance sheet slightly more healthy, so reducing
> the number of employees they have made redundant in the past, due to
> poor sales. 

Your point about hobby level advocacy is a good one that I share.  If
they are careful about how they run their program, they can get more
than dollars out of the deal.  They can get the engineering and ideas
for improvement from people outside their enterprise without undue
cost.

There is one positive here though.  Perhaps the attention Open Source
is getting might cause a few things to happen:

- UNIX becomes a bit more unified in how it does things.  Personally,
I do not consider the diversity a problem, but many do.

- Commercial vendors just might get a little more free with their
licenses once they realize the effect this growing OSS mindshare will
have on their market position.

- Though porting is a problem, people *will* do it.  Commercial UNIX
already benefits nicely from the existing body of OSS.  As this
continues, UNIX workstations will hold their value to potential users.
 This combined with the mindshare issue above just might be a good
thing overall.

Tony 'Nicoya' Mantler wrote:

> In article <3F330255.B90F491B@ntlworld.com>,
>  "Dr. David Kirkby" <drkirkby@ntlworld.com> wrote:
>
> : The problem is that a high
> : percentage of developers develop code on a machine which is
> : a) 32-bit
> : b) big-endian
>
> x86 is little-endian, actually as is the Itanic and x86-64. SPARC is big-endian,

IA64 is bi-endian

Tony 'Nicoya' Mantler wrote:
> 
> In article <3F330255.B90F491B@ntlworld.com>,
>  "Dr. David Kirkby" <drkirkby@ntlworld.com> wrote:
> 
> : The problem is that a high
> : percentage of developers develop code on a machine which is
> : a) 32-bit
> : b) big-endian
> 
> x86 is little-endian

Sorry it was late and I did not even think about what way around it
was. Yes, I accept that the bog-standard PC, which is what I was
meaning, is little-endian. 

In my life I've even come across a machine which I guess you would
call mid-endian or something like that. It was a 24-bit machine, with
either most (or might have been least) significant byte in the
middle!! That was on a Bruker NMR (or MRI as it's now called) clinical
system. 

> Thanks primarily to Debian, endian errors are getting very rare, and
> pointer-size errors are slowly on their way out too.

I don't know what Debian's contribution is, but it can't get around
the fact that if people write binary data, it will be unreadable on
another system, unless they make provision by having a switch to read
data in the other format, or they did what I once did, which was to
write a program such that the binary data was written in one format,
irrespective of the platform. 

> [...]
> : I'd like to read the temperature of my SCSI disks. There are several
> : Linux utilities to do this, but not one will work on SPARC. Yet the
> : author of 'cdrecord' has clearly demonstrated it's possible to write
> : low-level SCSI code that is portable.
> [...]
> 
> Possible but difficult. Every single platform uses a different API to send
> commands to and from drives. cdrecord gets around this by (no surprise) writing
> different drive interface code for every single platform.

In which case, if you develop a program that needs low-level SCSI, why
not start from J�rg Schilling's code, where the API is I assume
consistent for cd-record, and the interface code is platform specific?
J�rg Schilling might like to comment on that. 

I accept low-level code is more difficult, but if people had access to
different systems cheaply, they might think about these issues more. 

But there are many bits of code around where the too much is assumed
about the platform, that does not need to be. It might assume 'memset'
exists, which is not always the case. It might assume 'long long'
exists, which is not always the case. I can't help feeling that making
things like the OS and development tools readily available for
non-commercial use, will encourage more people to look to use other
hardware, and perhaps realise gcc is not the only compiler in the
world, and some of gcc's options are not accepted by someone else's
compiler. 

> One of these days I might port smartmontools or some other package to IRIX, but
> I don't have a Sun so you're out of luck there.

Well I have an SGI too, but are less concerned about that, as it only
runs a few hours/week and don't bother me too much. Yesterday and
Wednesday I had a lot of temperature problems on my older SPARC 20's
which have fast disks and fast CPUs in them. Hence my 'brute force'
approach to over-temperature protection
http://www.medphys.ucl.ac.uk/~davek/sun/cool.html
came into effect. 
-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Doug Dingus wrote:
> 
> "Dr. David Kirkby" <drkirkby@ntlworld.com> wrote in message 
> > I do use open-source. I have here systems running Linux (PC and
> > SPARC), NetBSD and OpenBSD. However, to be honest, I've never looked
> > at the source code for the operating system, which is I guess the same
> > for 99% of UNIX users. I see the operating system as a tool to do
> > something constructive with, like I consider a saw and chisel tools to
> > do constructive work with. I've never had much interest in making a
> > saw or a chisel.
> 
> Figured you did.  I agree with your tool analogy.  It does break down
> though when you compare the obvious and open nature of the saw to the
> closed and murky nature of binary only software.  However, my point
> was not to exclude commercial UNIXes, only to make sure Open code was
> part of the package.  In this, I believe we agree and the analogy is
> sound given the Open package will indirectly educate whoever is
> interested how exactly that 'saw' does its thing.

We agree there. 

> (snip)
> > > Given the state of things today, Linux appears to be a very good
> > > choice in that regard with the BSD operating systems being another.
> >
> > I accept Linux and *BSD are useful. However, there is a serious
> > problem if these are used in isolation, which is one reason I'd like
> > to see free or cheap versions of the major UNIX operating systems and
> > development tools for non-commerical use. The problem is that a high
> > percentage of developers develop code on a machine which is
> > a) 32-bit
> > b) big-endian

My mistake, I should have said little-endian.

<snip>
> > i) etc etc etc.
> 
> Agree with all of these in principle.  Multi-os exposure is important.
>  I use IRIX often with Linux.  The differences between the two are
> annoying, but educational at the same time.  

Exactly. 

> > 95% of those machines are PC, and a very large proportion run Linux,
> > with most being Redhat. Hence people tend to produce unportable code.

> Unportable code happens when developers focus on a platform.  In this,
> my favor is split.  On one hand, the attention Open Source, Linux in
> particular, is getting is a damn good thing.  On the other, porting
> becomes more of a problem as it gains mindshare.
> I am not sure a good solution exists for this; thus, your position on
> many platforms is a sound one.  (Not sure it will win the war though
> it will generate people who will continue to understand the issue.)
> 

Well if it makes a few people inside Sun, IBM, HP, SGI etc think,
there might be some benefit to it. 

If nothing else, it might make people inside Sun realise that there is
some ambiguity of how their free binary licence is being interpreted.
It might say *exactly* what Sun want it to say, it might be legally
sound, but the wording on different web pages clearly creates
confusion. 

I personally thought the conditions were clear, but there are many
others (including someone at Sun) who have different views on what it
means on SPARC, and others who have different views about what it
means on x86. The number of mis-interpretations seems to be huge. So
perhaps Sun could at least re-word it, so it's less ambiguous. 

> > I also learn to use different systems.  For an administrator, the
> > commercial operating systems are very different. Despite the fact a
> > large number of people may use those operating systems, fewer will get
> > the chance to be root, unless they get a machine at home. Only then
> > can they learn the OS more fully.
> 
> There is always ebay...  Seriously, this is true.  As is your point
> about free licenses.  I am just not sure things will change however
> much you and I would like them to.  Creative students can network
> around and get what they want behind the scenes, creative educators
> have more limitations I am afraid...

eBay is a good source of hardware and OS software, but try getting
compiler tools. Many are tied to specific host ID's. In any case, not
everyone wants to do something illegal. 

> Your point about hobby level advocacy is a good one that I share.  If
> they are careful about how they run their program, they can get more
> than dollars out of the deal.  They can get the engineering and ideas
> for improvement from people outside their enterprise without undue
> cost.

Yes they can. I've tried to highlight to Sun (via one of their web
pages), the errors in the definition of 'load average' as defined in
the 'uptime' man page. An error which seems to be in every 'uptime'
man page I've come across. I only came across that by using a Sun as a
hobby and realising the error. 

I don't suppose it will get published, but only this month I wrote an
email to the letters page of 'Personal Computer World' pointing out
the error in their statement about Apple being the first to bring a
64-bit machine to the desktop, pointing out Sun had done it nearly a
decade ago. 


> - Commercial vendors just might get a little more free with their
> licenses once they realize the effect this growing OSS mindshare will
> have on their market position.

Yes, that might be true. 

> - Though porting is a problem, people *will* do it. 

Sometimes. If it's a 'killer application' that everyone wants, that
probably it will be ported. Make it a more specialised tool, with a
smaller potential user base, and that application probably won't get
ported. If it's a very large application, where the code gets hard to
follow, then again it might not get ported. 

> Commercial UNIX
> already benefits nicely from the existing body of OSS. 

Yes, they have gained a lot. It would not hurt them to put a bit more
back. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

In article <3F336F45.400A411A@ntlworld.com>,
Dr. David Kirkby <drkirkby@ntlworld.com> wrote:

>> [...]
>> : I'd like to read the temperature of my SCSI disks. There are several
>> : Linux utilities to do this, but not one will work on SPARC. Yet the
>> : author of 'cdrecord' has clearly demonstrated it's possible to write
>> : low-level SCSI code that is portable.
>> [...]
>> 
>> Possible but difficult. Every single platform uses a different API to send
>> commands to and from drives. cdrecord gets around this by (no surprise) writing
>> different drive interface code for every single platform.
>
>In which case, if you develop a program that needs low-level SCSI, why
>not start from J�rg Schilling's code, where the API is I assume
>consistent for cd-record, and the interface code is platform specific?
>J�rg Schilling might like to comment on that. 

libscg is giving you a platform independant SCSI transport interface.
It is easy to use and there are remplate programs that you can easily
use to put your code into.

The simples way to start is to fetch cdrtools and modify a small SCSI program
like "readcd" or "scgcheck".

-- 
EMail:joerg@schily.isdn.cs.tu-berlin.de (home) J�rg Schilling D-13353 Berlin
      js@cs.tu-berlin.de		(uni)  If you don't have iso-8859-1
      schilling@fokus.fraunhofer.de	(work) chars I am J"org Schilling
URL:  http://www.fokus.fraunhofer.de/usr/schilling ftp://ftp.berlios.de/pub/schily

Dave Uhring wrote:

> Government regulation of commerce has many problems associated with it.
> The system works well as it is, although some fine tuning of the system
> could be done.  Requiring Qualcomm equipment would be a good start.

	Breaking MS's "monopoly" would be another. Its been done before
	with Bell. How well do Americans consider that worked anyway?

> >       It would be interesting to see what happens when petroleum based
> >       fuels eventually run out. I'm sure hydrogen fuel would make it by
> >       then but who's to say.
> 
> Good point.  We may have to re-establish the draft to rebuild our army, at
> least to postpone that eventuality.

	The rail system could back to steam, at least.

> >       Do the exact details really need to come out in court? Or can it
> >       be kept in camera?
> 
> If the bank can claim that the hack was a terrorist act then USA PATRIOT
> will ensure that the details are kept in camera.  Star chambers seem to be
> in our future.

	Hmmm ... it might be a stretch but it could work. Imagine hackers
	ending up in Camp X-Ray.

> >       Unless there were no subsidies?
> 
> No subsidies implies the eventual cessation of the service given the
> present management.  This would be unacceptable given the "national
> interest" in maintaining a rail passenger service.

	Yes but I was thinking it might give them some incentive to become
	profitable and keep their jobs.

> >       That's interesting. And with Mugger trying to rob one of the subway
> >       vending machines (according to a post of his some months back), it
> >       doesn't help its profitability.
> 
> A trivial concern considering the mismanagement of the railroad companies
> in the past century and the losses suffered from that.

	Yes. He didn't get away with it anyway. The guard chased him off.

> >       I wonder what patronage falloff would be like as fares went up?
> 
> Probably minimal on the routes between Washington and Boston.  That area
> has a high population density and poor facilities for automobile traffic.

	I was under the impression that the area was choked with interstates.

> The only time that NYC was my airline destination I rented a car to drive
> to Connecticut for my work.  Since I needed local transportation there my
> employer paid for the car, and did so quite happily.  The US Air Force was
> paying cost plus on that contract.

	Consider yourself lucky.

-am	� 2003

On Fri, 08 Aug 2003 23:26:31 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:

> 	Breaking MS's "monopoly" would be another. Its been done before
> 	with Bell. How well do Americans consider that worked anyway?

It produced lower quality of service and higher costs.  The breakup didn't
last either since the baby Bells are recombining.

> 	The rail system could back to steam, at least.

The tree huggers would raise holy hell about the resulting pollution and
timber consumption.

> 	Hmmm ... it might be a stretch but it could work. Imagine hackers
> 	ending up in Camp X-Ray.

Not the least far-fetched idea.  We have a Himmler wannabe who would be
happy to do that.  He lost the election to the US Senate to a dead man so
Dubya gave him a job where he could get revenge for that affront.
 
> 	Yes but I was thinking it might give them some incentive to become
> 	profitable and keep their jobs.

I guess you have never managed union people.  Besides, their jobs would be
protected anyway.  The feds would just take over the passenger rail
service when the private companies failed again.

>> Probably minimal on the routes between Washington and Boston.  That area
>> has a high population density and poor facilities for automobile traffic.
> 
> 	I was under the impression that the area was choked with interstates.

In 1968 I worked for 6 months at Aberdeen Proving Ground, about 30 miles
from Baltimore.  Many times it took over 2 hours to drive to Baltimore for
supper because of traffic congestion.  I'm sure that the situation is even
worse now.

>> The only time that NYC was my airline destination I rented a car to drive
>> to Connecticut for my work.  Since I needed local transportation there my
>> employer paid for the car, and did so quite happily.  The US Air Force was
>> paying cost plus on that contract.
> 
> 	Consider yourself lucky.

Better yet, I arrived shortly after noon and avoided the rush hour traffic
getting out of the city.

Dave Uhring wrote:

> It produced lower quality of service and higher costs.  The breakup didn't
> last either since the baby Bells are recombining.

	Are they allowed to? I gather they are also up against competition
	from other telcos.

> >       The rail system could back to steam, at least.
> 
> The tree huggers would raise holy hell about the resulting pollution and
> timber consumption.

	I haven't seen them use timber in years. Usually its coal
	(however, in years past they've been known to use Egyptian
	mummies).

> >       Hmmm ... it might be a stretch but it could work. Imagine hackers
> >       ending up in Camp X-Ray.
> 
> Not the least far-fetched idea.  We have a Himmler wannabe who would be
> happy to do that.  He lost the election to the US Senate to a dead man so
> Dubya gave him a job where he could get revenge for that affront.

	Is that Poindexter? If we ask nicely, will he send Tugger there?

> I guess you have never managed union people.

	Actually, I had the misfortune of working with them when I had
	a summer job with the SRA (State Rail Authority; now called
	StateRail). I put myself thru uni that way for a few years.
	Its also where I developed a loathing for unionists.

> Besides, their jobs would be
> protected anyway.  The feds would just take over the passenger rail
> service when the private companies failed again.

	A real no-win situation.

> >       I was under the impression that the area was choked with interstates.
> 
> In 1968 I worked for 6 months at Aberdeen Proving Ground, about 30 miles
> from Baltimore.  Many times it took over 2 hours to drive to Baltimore for
> supper because of traffic congestion.  I'm sure that the situation is even
> worse now.

	After 35 years, I'd imagine it would be.

-am	� 2003

On Sat, 09 Aug 2003 00:30:43 +1000, Anthony Mandic wrote:

> 	Are they allowed to? I gather they are also up against competition
> 	from other telcos.

The original anti-trust judgment applied to AT&T, not the baby Bells. 
Sure they are free to recombine and have done so.  I don't recall the
exact number that remain, only the one which services me.  SBC ate
Ameritech and Pac Bell, possibly another one or three.

The other telcos need to use the Bells' infrastructure and even though the
Bells are required to permit that usage they can and do make it difficult.

> 	I haven't seen them use timber in years. 

Well, in the days of steam coal was not readily available everywhere and
many of those engines used wood as fuel.  

>       Usually its coal
> 	(however, in years past they've been known to use Egyptian
> 	mummies).

Never heard about the mummies.  Diesel replaced coal; higher energy
content per pound and cleaner.  Besides it's pretty hard to burn coal in a
Diesel engine, and those old steamers were dangerous.  They had an
annoying tendency to explode.

> 	Is that Poindexter? If we ask nicely, will he send Tugger there?

Poindexter is on the way out because of the terrorism futures market he
proposed.

http://www.bayarea.com/mld/mercurynews/news/world/6434410.htm

No, the Himmler wannabe is our chief law enforcement officer, the Attorney
General, former governor of Missouri and failed Senatorial candidate, Mr.
John Ashcroft.  I'm not sure if he can be asked nicely unless you're
willing to out-donate Microsoft for his future political efforts.

> 	Actually, I had the misfortune of working with them when I had
> 	a summer job with the SRA (State Rail Authority; now called
> 	StateRail). I put myself thru uni that way for a few years.
> 	Its also where I developed a loathing for unionists.

Then you should understand how likely they are to improve productivity and
profit for their employers just to preserve their jobs.

>> In 1968 I worked for 6 months at Aberdeen Proving Ground, about 30 miles
>> from Baltimore.  Many times it took over 2 hours to drive to Baltimore for
>> supper because of traffic congestion.  I'm sure that the situation is even
>> worse now.
> 
> 	After 35 years, I'd imagine it would be.

Despite ongoing highway construction and maintenance traffic volume seems
always to outpace the improvements.

Anthony Mandic <s8@hotmail.com> wrote in message news:<3F331319.23FEDE7C@hotmail.com>...
> Doug Dingus wrote:
> 
(snip)
> > If someone else learned both the calculator way, and the pencil and
> > paper way, they have the option of using calculators, but are free to
> > perform any calculation they want without as well.
> 
> 	You could also write programs to do calculations. But note
> 	that calculations are not learning maths. Its simply a means
> 	of solving equations.

(was not factoring a computer into the simple analogy)

> 
> > Considering the licensing issue that started this thread, my statement
> > about Linux and its utility for education has merit enough without all
> > the little annoying side arguements.
>  ...
> > Given the state of things today, Linux appears to be a very good
> > choice in that regard with the BSD operating systems being another.
> 
> 	If you were willing to continue this, my arguments would
> 	have been that GNU and Open Source have been around a lot
> 	longer than Linux. They didn't need Linux to become popular
> 	or worthwhile but Linux really needed them to become anything
> 	worthwhile.



That's not a bad discussion; however, it is off topic in this thread
and fairly clear to both of us as well.  You are correct in the
relationship between Linux and GNU, though the path you were taking
was not obvious to me at all.  Perhaps showing a couple of cards would
have gotten us here quicker no?

I would also point out that Linux/GNU have mutually gained from each
others presence and development.  Today they both need each other. 
Maybe GNU will finish their kernel one day so we can see what a GNU OS
would be like.

You sound awfully bitter. Cheer up, things can't possibly be as bad as
they seem.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

In article <bgbk0q$msvr2$1@ID-22005.news.uni-berlin.de>,
Benjamin Gawert <bgawert@gmx.de> wrote:
> Such general statements are nothing but false. Windows isn't always a worse
> choice over some Unix operating system, it simply depends on the task and
> the configuration...

That's not the point. The point is that whether or not it's a better choice,
it's the one that's picked, and it's often picked despite being dramatically
worse for the job. Microsoft's own experience in converting Hotmail from
FreeBSD to Windows is a perfect object lesson.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

In article <3F336F45.400A411A@ntlworld.com>,
 "Dr. David Kirkby" <drkirkby@ntlworld.com> wrote:

: Tony 'Nicoya' Mantler wrote:
[...]
: > x86 is little-endian
: 
: Sorry it was late and I did not even think about what way around it
: was. Yes, I accept that the bog-standard PC, which is what I was
: meaning, is little-endian. 
: 
: In my life I've even come across a machine which I guess you would
: call mid-endian or something like that. It was a 24-bit machine, with
: either most (or might have been least) significant byte in the
: middle!! That was on a Bruker NMR (or MRI as it's now called) clinical
: system. 

I seem to recall there's one machine that's byte swapped on 16-bit boundaries, 
but I can't remember which one it is.


: > Thanks primarily to Debian, endian errors are getting very rare, and
: > pointer-size errors are slowly on their way out too.
: 
: I don't know what Debian's contribution is, but it can't get around
: the fact that if people write binary data, it will be unreadable on
: another system, unless they make provision by having a switch to read
: data in the other format, or they did what I once did, which was to
: write a program such that the binary data was written in one format,
: irrespective of the platform. 

Well, debian's contribution is to fix exactly that. Since they support so many 
different platforms, many packages have to get marked with bugs for incorrect 
endian support, or incorrect 64-bit support. Depending on the severity of the 
error, it *has* to be fixed by the next debian release, or the package faces 
exclusion.

These fixes usually get pushed back upstream to the original programmers who are 
usually more than happy to see them.

MacOS X is also helping things, being it's rather popular and runs on big-endian 
and now 64-bit hardware.


[...]
: > Possible but difficult. Every single platform uses a different API to send
: > commands to and from drives. cdrecord gets around this by (no surprise) writing
: > different drive interface code for every single platform.
: 
: In which case, if you develop a program that needs low-level SCSI, why
: not start from J�rg Schilling's code, where the API is I assume
: consistent for cd-record, and the interface code is platform specific?
: J�rg Schilling might like to comment on that. 

Well, two factors:

1) cdrecord records CDs. It's SCSI library isn't (last I checked) packaged in a 
seperate conviniently reusable form.
2) The programmer who wrote the SMART tools woke up one morning and said "I want 
to get SMART data on my linux box", not "I want to get SMART data on any unix 
box".

From a simplicity standpoint, it's much easier to write a program for one 
system, then worry about portability, than it is to fuss about portability from 
the start, especially when it comes to APIs as varied as SCSI.


: I accept low-level code is more difficult, but if people had access to
: different systems cheaply, they might think about these issues more. 

Well, people need to have access, *and* want to use the machines. Sometimes I 
think that Linux/SPARC is more popular than Solaris/SPARC.


: But there are many bits of code around where the too much is assumed
: about the platform, that does not need to be. It might assume 'memset'
: exists, which is not always the case. It might assume 'long long'
: exists, which is not always the case. I can't help feeling that making
: things like the OS and development tools readily available for
: non-commercial use, will encourage more people to look to use other
: hardware, and perhaps realise gcc is not the only compiler in the
: world, and some of gcc's options are not accepted by someone else's
: compiler. 

Porting an application that hasn't been ported much is always a bit of a 
headache. The biggest trick is learning the zen of autoconf.


: > One of these days I might port smartmontools or some other package to IRIX, 
: > but I don't have a Sun so you're out of luck there.
: 
: Well I have an SGI too, but are less concerned about that, as it only
: runs a few hours/week and don't bother me too much. Yesterday and
: Wednesday I had a lot of temperature problems on my older SPARC 20's
: which have fast disks and fast CPUs in them. Hence my 'brute force'
: approach to over-temperature protection
: http://www.medphys.ucl.ac.uk/~davek/sun/cool.html
: came into effect.

Heh, ouch!

I have to say I've never had any temperature problems with my SGI systems, even 
when the ambient temperature crawls over 30c.


Cheers - Tony 'Nicoya' Mantler :)

-- 
Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@apia.dhs.org
--  http://www.apia.dhs.org/  --  http://nicoya.feline.pp.se/  --

Tony 'Nicoya' Mantler wrote:
> 
> In article <3F336F45.400A411A@ntlworld.com>,
>  "Dr. David Kirkby" <drkirkby@ntlworld.com> wrote:
> 
> : Tony 'Nicoya' Mantler wrote:

> : In my life I've even come across a machine which I guess you would
> : call mid-endian or something like that. It was a 24-bit machine, with
> : either most (or might have been least) significant byte in the
> : middle!! That was on a Bruker NMR (or MRI as it's now called) clinical
> : system.
> 
> I seem to recall there's one machine that's byte swapped on 
> 16-bit boundaries,

nothing surprises me.

> : > Thanks primarily to Debian, endian errors are getting very rare, and
> : > pointer-size errors are slowly on their way out too.
> :
> : I don't know what Debian's contribution is, but it can't get around
> Well, debian's contribution is to fix exactly that. 
> Since they support so many
> different platforms, many packages have to get marked 
> with bugs for incorrect
> endian support, or incorrect 64-bit support. Depending 
> on the severity of the error, it *has* to be fixed 
> by the next debian release, or the package faces
> exclusion.


Em, but a lot of packages for which the source is available, are not
distributed by Debian. 

> MacOS X is also helping things, being it's rather popular and 
> runs on big-endian
> and now 64-bit hardware.

Same as SPARC. 

> : In which case, if you develop a program that needs low-level SCSI, why
> : not start from J�rg Schilling's code, where the API is I assume
> : consistent for cd-record, and the interface code is platform specific?
> : J�rg Schilling might like to comment on that.
> 
> Well, two factors:
> 
> 1) cdrecord records CDs. It's SCSI library isn't (last I checked) packaged in a
> seperate conviniently reusable form.

See J�rg Schilling's comments. 

> 2) The programmer who wrote the SMART tools woke up one 
> morning and said "I want to get SMART data on my linux box", 
> not "I want to get SMART data on any unix box".

Exactly. That is what I see is wrong with a lot of Linux development. 
 
> From a simplicity standpoint, it's much easier to write a program for one
> system, then worry about portability, than it is to fuss 
> about portability from the start, especially when it comes to 
> APIs as varied as SCSI.

Sure it is easier, if you consider making it publicly available, it
would be helpful if you do put a bit of effort in at the beginning
thinking about portability.
 
> : I accept low-level code is more difficult, but if people had access to
> : different systems cheaply, they might think about these issues more.
> 
> Well, people need to have access, *and* want to use the machines.

True. 
 
> Porting an application that hasn't been ported much is always a bit of a
> headache. The biggest trick is learning the zen of autoconf.

autoconf does help a lot. You still need to think about it. The other
day I changed one of my variables from type 'long long' to 'double'
when I realised that I could achieve exactly the same result using a
more portable data type. 

> : Yesterday and
> : Wednesday I had a lot of temperature problems on my older SPARC 20's
> : which have fast disks and fast CPUs in them. Hence my 'brute force'
> : approach to over-temperature protection
> : http://www.medphys.ucl.ac.uk/~davek/sun/cool.html
> : came into effect.
> 
> Heh, ouch!
> 
> I have to say I've never had any temperature problems 
> with my SGI systems, even when the ambient temperature crawls over 30c.

It is not really Sun's fault. At the time the SS20 was made, the CPUs
were slower (used less power) and disks used less power too. Then Ross
came along and made Hyersparcs, which use quite a lot of power. I tend
to put 10 k rpm disks in the SS20's, as they are the only disks
readily available now in the larger sizes. The temperatures here have
been 36 dec C lately, but I suspect my garage is over 40 deg C
internally. The disks have a maximum temperature of around 55 deg C,
so there is not a lot of room. So my 50 deg C cut-outs are tripping. I
think the cooling was a weak point on the SS20, but I've tended to
make the situation worst by the faster disks. I'm willing to accept
that though - I only use the machines for portability testing, so it
does not matter to me if I run them in the evening only. 



-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring <daveuhring@yahoo.com> wrote:
>  Cypherpunk wrote:
>   
>  > I was not complaining that he posted off-topic, despite it
>  > looking that way superficially.
>  > 
>  > You'll have to dig deeper with Groups Google.
>  > 
>  > Try: tsar "Peter da Silva"
>   
>   You are ignoring my statement about personal attacks.  It's time to stop
>   that.  Even if this particular thread is OT, the NG is a technical NG and
>   such crap has no place here.
>   
>   If you don't like the opinions of your enemies, then killfile them.

Oh, so sorry Dave, but you've been chattering away with Anythony,
who has repeately (at least five in the past few days) made homophobic
and/or other "personal attack" comments in response to my posts.

Not once did you say anything about that, and now he's done it in this thread.

Check out his post today under "script question - urgent" in case
you missed that beauty.

Feel free to continue encouraging him merely by talking to him
(treating him as if he were normal) while he posts crap at me
even when I don't mention him, and feel free to killfile me.

Or, you could refuse to talk to him while he posts like a kook.

Whatever you choose, try to be consistent in your dealings with people.

----

In comp.unix.solaris Peter da Silva <peter@abbnm.com> wrote:
>
>   You sound awfully bitter.

Bitter? Heck no, it was an enjoyable dig at one of the goofballs
who tried to take over Usenet.

>   Cheer up, things can't possibly be as bad as they seem.

That's right, your attempt tanked.

Ironically, your crosspost to four groups wouldn't have
even been allowed. ;-)

On Fri, 8 Aug 2003, Dr. David Kirkby wrote:

> > 2) The programmer who wrote the SMART tools woke up one
> > morning and said "I want to get SMART data on my linux box",
> > not "I want to get SMART data on any unix box".
>
> Exactly. That is what I see is wrong with a lot of Linux development.

And I think that THAT problem is engendered by the lack of
experience and professionalism amongst many Linux programmers
(not, I hasten to add, all Linux programmers!).

I'd wager that many Linux programmers have little or no
computer science education or exposure to industry.  Hence
they grow up with Linux on x86 being their only world, and
get the misguided impression that Linux is all that counts.
It's a bit of a vicious circle, really.

I've heard many people saying that it's much easier to port
something TO Linux, rather than FROM it, and I think that
says a lot.

-- 
Rich Teer, SCNA, SCSA

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-online.net

In article <nicoya-825927.13001808082003@shawnews.wp.shawcable.net>,
Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
> I seem to recall there's one machine that's byte swapped on 16-bit
> boundaries, 

Well, um, sorta.

The PDP-11 doesn't have any 32-bit integer operations, but the floating
point operations are big-endian and the integer operations are little-endian,
so if you try and mix the two you end up with middle-endian layout.

Of course anyone trying to mix floating and integer operations that way,
particularly back in the '70s when there were at least twice as many
floating point formats as there were processors that supported floats,
was just screaming for the foulup fairy to pay them a visit.

There has also been some argument that the double-register operations
are middle-endian, but since registers aren't dynamically addressible
it's just a convention and invisible to all but the most clear and pure
academic assembly programmer.

Yes, I know that's a contradition in terms.

That's the point.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

In article <nicoya-D7C64A.15214408082003@news.wp.shawcable.net>,
Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
> No doubt it would be nice, but it's not always practical to write something 
> portably on the first version.

It's rarely practical to do otherwise, unless you really DO build one to
throw away and actually throw it away. Why? Two reasons:

1. Writing portably really works best when it's done from the start.

2. Writing portably doesn't just help other people, it helps you, because
   the system you first port to is often the next major release of the one
   you started with.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

Tony 'Nicoya' Mantler wrote:

> : Em, but a lot of packages for which the source is available, are not
> : distributed by Debian.
> 
> I rarely find myself in want of a package that isn't distributed in Debian.
> They're very good about including everything and the kitchen sink.

I was not aware of that. Have they got my finite difference package
for computing the properties of transmission lines?

http://atlc.sourceforge.net/ 

It is in NetBSD 
http://wuarchive.wustl.edu/mirrors/NetBSD/NetBSD-current/pkgsrc/cad/atlc/README.html
(a very old version) and someone has produced a binary for a Sony
Playstation 2. I've never heard from a user running it on a Sony
Playstation 2, but apparently it has been built. 

I did have a problem on Debian SPARC
http://www.mail-archive.com/autoconf@gnu.org/msg08981.html
but by testing that problem was found and resolved. 

> : > 2) The programmer who wrote the SMART tools woke up one
> : > morning and said "I want to get SMART data on my linux box",
> : > not "I want to get SMART data on any unix box".
> :
> : Exactly. That is what I see is wrong with a lot of Linux development.
> 
> That's why it's 'linux development', not 'unix development'. ;)

Yes, that is a shame. 

> : Sure it is easier, if you consider making it publicly available, it
> : would be helpful if you do put a bit of effort in at the beginning
> : thinking about portability.
> 
> No doubt it would be nice, but it's not always practical 
> to write something portably on the first version.

No, but having thought about portability issues, you can then make it
a lot easier for others, with little work for yourself, by thinking
about portability issues and then doing your best to remove them. 

> : autoconf does help a lot. You still need to think about it. The other
> : day I changed one of my variables from type 'long long' to 'double'
> : when I realised that I could achieve exactly the same result using a
> : more portable data type.
> 
> I hope you're not substituting a floating point type for an integer type.

Yes I am substituting a floating point type for an integer type! In
this instance I think that is a good idea, as it avoids the use of a
'long long' on 32-bit systems, which would otherwise be needed. I
could write my own 64-bit integer multiplication routine (or even link
with one that exists), but I feel the use of a double in this instance
was okay. I might be wrong, but at least I have given it some thought,
which is more than seems to happen with many people who only develop
only on Linux. 

I needed to find the ram in the system in Mb on Solaris (other methods
are used on other systems), so only Solaris is considered. Any Solaris
system will have ram which is an integer multiple of a Mb. I first
determine the number of pages of ram:

sysconf(_SC_PHYS_PAGES); /* returns pages as a long */

and put that in a double. Next I determine the size of the pages in
bytes

sysconf(_SC_PAGESIZE); /* returns page size as a long */

These are multiplied together, which will overflow a long if the
machine has more than 2 Gb of ram. Next it gets divided by the number
of bytes per Mb, and cast to the nearest long. Ignoring the error
checks, the code looks something like this:

#define BYTES_PER_MB  1048576

  double ram;
  long memory_in_mb;
 
  ram=(double) sysconf(_SC_PAGESIZE); /* size of page in bytes */
  ram*= (double) sysconf(_SC_PHYS_PAGES); /* ram now in bytes */
  ram=ram/BYTES_PER_MB; /* convert to Mb, as a double */
  memory_in_mb=(long) (ram + 0.5); /* cast to nearest long. */

before adding 0.5 and casting to a long, the number of Mb is likely to
be non-exact due to rounding errors. But the error should never be as
much as 0.5. Even on a system with 1 Tb of RAM, the precision only
needs to be about 6 significant digits for this to be a valid
assumption. 

I eventually cast to a long again, rounding to the nearest integer
value. I think that is acceptable. It works okay on a Solaris box with
4 Gb of RAM, as it does on several others with smaller amounts of ram.
The same principle has been tested on several other non-Solaris
machines, all of which result in the exact right number. 

You might not like the above code. But it works, and is more portable
than trying to use 'long long'. So I have:

a) Realised using 'long long' is not portable.
b) Thought about ways around the problem.
c) Determined for any practical system it should work.
d) Tested my solution and verified it works on several Solaris
machines, which is all this needs to work on.
e) Tested the same principle (but slightly different code) on several
other non-Solaris machines and verified it is okay.

Whether or not the above code is good or bad is to a certain extent
immaterial. I have at least thought about the issue and tried to keep
it portable. That is something that a lot of people who only develop
on Linux and gcc seem to ignore. 

Hence I feel introducing people to other compilers, and systems is
good for computing as a whole. 


-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

Dave Uhring <daveuhring@yahoo.com> wrote:
>  Cypherpunk wrote:
>
>  > In this thread alone, let alone for the past few months at least,
>  > he has been serially forging (exact address forgery) people's
>  > accounts at hotmail.com through the Deutsche newsserver.
>
>   How are these addresses obtained?

He's probably got a script that is generating two character
combinations for now, automatically each time he posts. He
started off with one character names at hotmail.com.

>  How do you know they are legitimate addresses?

The implicit answer is the same for domainnames: all the  one/two-letter
combinations will be registered.

Here is the range of his forgeries:

http://groups.google.com/groups?num=50&hl=en&lr=&ie=UTF-8&oe=UTF-8&scoring=d&as_drrb=b&q=author%3Aanthony+author%3Amandic&btnG=Google+Search&as_mind=27&as_minm=5&as_miny=2003&as_maxd=8&as_maxm=8&as_maxy=2003

Pretty massive. I don't know anyone who has ever forged more addresses than he.

The explicit answer is I checked Groups Google when I noticed his
address was changing rapidly, when he posted as "ie@hotmail.com."

author ie@hotmail.com:

   france.ile-de-france.creteil - Aug 4, 2003 by babyfoot - View Thread (1 article)
   comp.unix.solaris - Jul 8, 2003 by Anthony Mandic - View Thread (28 articles)

There are over 200 posts by "babyfoot" as ie@hotmail.com,
indicating that is the true owner.

Of course, this will only verify he is forging people
who post to Usenet. You'd have to actually email the
others to check the address is in use.

Peter da Silva wrote:
> 
> In article <nicoya-D7C64A.15214408082003@news.wp.shawcable.net>,
> Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
> > No doubt it would be nice, but it's not always practical to write something
> > portably on the first version.
> 
> It's rarely practical to do otherwise, unless you really DO build one to
> throw away and actually throw it away. Why? Two reasons:
> 
> 1. Writing portably really works best when it's done from the start.

I could not agree more. One really needs to think about the possible
issues right from the start - even if you can't resolve all of them.
It's much harder to correct the issues later. 

> 2. Writing portably doesn't just help other people, it helps you, because
>    the system you first port to is often the next major release of the one
>    you started with.

I could not agree more. 

By testing your current project on a number of systems, it also
teaches you about portability issues that you can use in your next
project. I recently learn that doing 

char foo;

is unportable, so won't do that again, unless I'm sure it is okay in
the context of using it. 

You can often do this for free, using resources such as the
SourceForge compile farm (for Sourceforge hosted projects) 
http://sourceforge.net/docman/display_doc.php?docid=762&group_id=1
and the HP test drive (well worth a look)
http://www.testdrive.compaq.com/
Other will often test code for you. 

-- 
Dr. David Kirkby,
Senior Research Fellow,
Department of Medical Physics,
University College London,
11-20 Capper St, London, WC1E 6JA.
Tel: 020 7679 6408 Fax: 020 7679 6269
Internal telephone: ext 46408
e-mail davek@medphys.ucl.ac.uk

In article <bh160b$1pjj$3@jeeves.eng.abbnm.com>,
 peter@abbnm.com (Peter da Silva) wrote:

: In article <nicoya-D7C64A.15214408082003@news.wp.shawcable.net>,
: Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
: > No doubt it would be nice, but it's not always practical to write something 
: > portably on the first version.
: 
: It's rarely practical to do otherwise, unless you really DO build one to
: throw away and actually throw it away. Why? Two reasons:
: 
: 1. Writing portably really works best when it's done from the start.
: 
: 2. Writing portably doesn't just help other people, it helps you, because
:    the system you first port to is often the next major release of the one
:    you started with.

Well, there's different classifications for portability issues.

Some are very simple, like bcopy() vs memcpy(), or differences in header files. 
Thankfully in linux (at least in debian) there's a 'conforming to' section in 
the manpages that can give you some indication of the portability of a function 
without having to search though the manpages of 10 different systems. These 
kinds of issues are reasonably easy to address at any point in development.

Other issues can be a bit more complex - for example the raw storage command 
interface, where there is no standard nor even the slightest resemblence between 
the APIs on different platforms. This is the case where I would argue that it's 
easier to make the first stab non-portable (while keeping an eye toward making 
sure you can make it portable later) then add portability in the next iteration.

If you try to write both your application *plus* a portability library at the 
same time, you end up effectively writing two projects at once which makes 
things a wee bit more difficult than they need to be.

Of course, you can then argue whether writing portable code means 'writing code 
which works on any platform' or 'writing code which can easily be made to work 
on any platform'. I would argue that it's OK to fall short of the former, but 
you should be very careful not to slip up on the latter. (with the exceptions of 
applications which only logically apply to one platform)


Cheers - Tony 'Nicoya' Mantler :)

-- 
Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@apia.dhs.org
--  http://www.apia.dhs.org/  --  http://nicoya.feline.pp.se/  --

Tony 'Nicoya' Mantler wrote:
> In article <bh160b$1pjj$3@jeeves.eng.abbnm.com>,
>  peter@abbnm.com (Peter da Silva) wrote:
> 
> : In article <nicoya-D7C64A.15214408082003@news.wp.shawcable.net>,
> : Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
> : > No doubt it would be nice, but it's not always practical to write something 
> : > portably on the first version.
> : 
> : It's rarely practical to do otherwise, unless you really DO build one to
> : throw away and actually throw it away. Why? Two reasons:
> : 
> : 1. Writing portably really works best when it's done from the start.
> : 
> : 2. Writing portably doesn't just help other people, it helps you, because
> :    the system you first port to is often the next major release of the one
> :    you started with.
> 
> Well, there's different classifications for portability issues.
> 
> Some are very simple, like bcopy() vs memcpy(), or differences in header files. 
> Thankfully in linux (at least in debian) there's a 'conforming to' section in 
> the manpages that can give you some indication of the portability of a function 
> without having to search though the manpages of 10 different systems. These 
> kinds of issues are reasonably easy to address at any point in development.
> 
> Other issues can be a bit more complex - for example the raw storage command 
> interface, where there is no standard nor even the slightest resemblence between 
> the APIs on different platforms. This is the case where I would argue that it's 
> easier to make the first stab non-portable (while keeping an eye toward making 
> sure you can make it portable later) then add portability in the next iteration.
> 
> If you try to write both your application *plus* a portability library at the 
> same time, you end up effectively writing two projects at once which makes 
> things a wee bit more difficult than they need to be.
> 
> Of course, you can then argue whether writing portable code means 'writing code 
> which works on any platform' or 'writing code which can easily be made to work 
> on any platform'. I would argue that it's OK to fall short of the former, but 
> you should be very careful not to slip up on the latter. (with the exceptions of 
> applications which only logically apply to one platform)
> 
> 
> Cheers - Tony 'Nicoya' Mantler :)
> 
PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT 
YOU ARE TALKING ABOUT!

Dr. David Kirkby wrote:

> Peter da Silva wrote:
> 
>>In article <nicoya-D7C64A.15214408082003@news.wp.shawcable.net>,
>>Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
>>
>>>No doubt it would be nice, but it's not always practical to write something
>>>portably on the first version.
>>
>>It's rarely practical to do otherwise, unless you really DO build one to
>>throw away and actually throw it away. Why? Two reasons:
>>
>>1. Writing portably really works best when it's done from the start.
> 
> 
> I could not agree more. One really needs to think about the possible
> issues right from the start - even if you can't resolve all of them.
> It's much harder to correct the issues later. 
> 
> 
>>2. Writing portably doesn't just help other people, it helps you, because
>>   the system you first port to is often the next major release of the one
>>   you started with.
> 
> 
> I could not agree more. 
> 
> By testing your current project on a number of systems, it also
> teaches you about portability issues that you can use in your next
> project. I recently learn that doing 
> 
> char foo;
> 
> is unportable, so won't do that again, unless I'm sure it is okay in
> the context of using it. 
> 
> You can often do this for free, using resources such as the
> SourceForge compile farm (for Sourceforge hosted projects) 
> http://sourceforge.net/docman/display_doc.php?docid=762&group_id=1
> and the HP test drive (well worth a look)
> http://www.testdrive.compaq.com/
> Other will often test code for you. 
> 
PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT 
YOU ARE TALKING ABOUT!

Rich Teer wrote:

> On Fri, 8 Aug 2003, Dr. David Kirkby wrote:
> 
> 
>>>2) The programmer who wrote the SMART tools woke up one
>>>morning and said "I want to get SMART data on my linux box",
>>>not "I want to get SMART data on any unix box".
>>
>>Exactly. That is what I see is wrong with a lot of Linux development.
> 
> 
> And I think that THAT problem is engendered by the lack of
> experience and professionalism amongst many Linux programmers
> (not, I hasten to add, all Linux programmers!).
> 
> I'd wager that many Linux programmers have little or no
> computer science education or exposure to industry.  Hence
> they grow up with Linux on x86 being their only world, and
> get the misguided impression that Linux is all that counts.
> It's a bit of a vicious circle, really.
> 
> I've heard many people saying that it's much easier to port
> something TO Linux, rather than FROM it, and I think that
> says a lot.
> 
PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT 
YOU ARE TALKING ABOUT!

Cypherpunk@nyc.rr.com wrote:

> Dave Uhring <daveuhring@yahoo.com> wrote:
> 
>> Cypherpunk wrote:
>>  
>> > I was not complaining that he posted off-topic, despite it
>> > looking that way superficially.
>> > 
>> > You'll have to dig deeper with Groups Google.
>> > 
>> > Try: tsar "Peter da Silva"
>>  
>>  You are ignoring my statement about personal attacks.  It's time to stop
>>  that.  Even if this particular thread is OT, the NG is a technical NG and
>>  such crap has no place here.
>>  
>>  If you don't like the opinions of your enemies, then killfile them.
> 
> 
> Oh, so sorry Dave, but you've been chattering away with Anythony,
> who has repeately (at least five in the past few days) made homophobic
> and/or other "personal attack" comments in response to my posts.
> 
> Not once did you say anything about that, and now he's done it in this thread.
> 
> Check out his post today under "script question - urgent" in case
> you missed that beauty.
> 
> Feel free to continue encouraging him merely by talking to him
> (treating him as if he were normal) while he posts crap at me
> even when I don't mention him, and feel free to killfile me.
> 
> Or, you could refuse to talk to him while he posts like a kook.
> 
> Whatever you choose, try to be consistent in your dealings with people.
> 
> ----
> 
> In comp.unix.solaris Peter da Silva <peter@abbnm.com> wrote:
> 
>>  You sound awfully bitter.
> 
> 
> Bitter? Heck no, it was an enjoyable dig at one of the goofballs
> who tried to take over Usenet.
> 
> 
>>  Cheer up, things can't possibly be as bad as they seem.
> 
> 
> That's right, your attempt tanked.
> 
> Ironically, your crosspost to four groups wouldn't have
> even been allowed. ;-)
PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT 
YOU ARE TALKING ABOUT!

On Fri, 8 Aug 2003, Alan Johnson wrote:

> PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT
> YOU ARE TALKING ABOUT!

If you'd take half a second to tell us which newsgroup(s)
you're talking about, I'm sure we'll be happy to oblige.

Now, was it really necessary to post your comment half a
dozen times, without snipping the articals you replied to?

-- 
Rich Teer, SCNA, SCSA

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-online.net

On Fri, 08 Aug 2003 23:33:47 +0000, Cypherpun wrote:

> 
> He's probably got a script that is generating two character
> combinations for now, automatically each time he posts. He
> started off with one character names at hotmail.com.
> 
>>  How do you know they are legitimate addresses?
> 
> The implicit answer is the same for domainnames: all the  one/two-letter
> combinations will be registered.

Then your accusation about people receiving increased spam because of
those fraudulent email addresses is unjustified.  You do not -know- that,
do you?

Are you also unaware that spammers run dictionary generated mailings,
particulary at hotmail.com?

In comp.unix.solaris Alan Johnson <adjtech@ulster.net> wrote:
>   
>   PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT 
>   YOU ARE TALKING ABOUT!

You have net.loon Peter duh Silva to thank for that.

He is the one who added to a comp.unix.solaris-only post the other
three newsgroups in message ID bgti6j$18m8$1@jeeves.eng.abbnm.com

Complain to him directly at: peter@abbnm.com

Dave Uhring <daveuhring@yahoo.com> wrote:
>  Cypherpunk wrote:
>  > 
>  > He's probably got a script that is generating two character
>  > combinations for now, automatically each time he posts. He
>  > started off with one character names at hotmail.com.
>  > 
>  >>  How do you know they are legitimate addresses?
>  > 
>  > The implicit answer is the same for domainnames: all the  one/two-letter
>  > combinations will be registered.
>   
>   Then your accusation about people receiving increased spam because of
>   those fraudulent email addresses is unjustified.  You do not -know- that,
>   do you?

Let me repeat this again, for the hard of hearing ;-)

Cypherpunk wrote:
>
>   Of course, this will only verify he is forging people
>   who post to Usenet. You'd have to actually email the
>   others to check the address is in use.

Remember, most people don't know what Usenet is.

>   Are you also unaware that spammers run dictionary
>   generated mailings, particulary at hotmail.com?

What's your point, that it's okay to *guarantee additional*
spam to an address?

That it's okay to do exact-address forgeries in Usenet?

If you're trying to defend Anythony on any grounds,
prepare to have the expression of Wile E. Coyote.

Dave Uhring wrote:

> The original anti-trust judgment applied to AT&T, not the baby Bells.

	Ah, of course.

> The other telcos need to use the Bells' infrastructure and even though the
> Bells are required to permit that usage they can and do make it difficult.

	It might have been better to split the systems up between
	infrastructure and service. Of course, those owning the
	infrastructure could screw those providing the service.

> Never heard about the mummies.

	Saw it in a documentary. Apparently they didn't burn all that
	well.

> Besides it's pretty hard to burn coal in a Diesel engine,

	Extremely hard.

> and those old steamers were dangerous.  They had an
> annoying tendency to explode.

	Yep.

> Poindexter is on the way out because of the terrorism futures market he
> proposed.

	Ah, that's what I'd heard about.

> No, the Himmler wannabe is our chief law enforcement officer, the Attorney
> General, former governor of Missouri and failed Senatorial candidate, Mr.
> John Ashcroft.  I'm not sure if he can be asked nicely unless you're
> willing to out-donate Microsoft for his future political efforts.

	No chance of that but I'm sure we can get Tugger to call him a
	maroon.

> Then you should understand how likely they are to improve productivity and
> profit for their employers just to preserve their jobs.

	Yes, but its a 50/50 proposition. Some should have lost their
	jobs a long time ago. Others were fairly dedicated.

> Despite ongoing highway construction and maintenance traffic volume seems
> always to outpace the improvements.

	I'm sure that there's an old adage for this phenomenon.

-am	� 2003

Doug Dingus wrote:

> >       You could also write programs to do calculations. But note
> >       that calculations are not learning maths. Its simply a means
> >       of solving equations.
> 
> (was not factoring a computer into the simple analogy)

	Would you believe programmable calculators?

> >       If you were willing to continue this, my arguments would
> >       have been that GNU and Open Source have been around a lot
> >       longer than Linux. They didn't need Linux to become popular
> >       or worthwhile but Linux really needed them to become anything
> >       worthwhile.
> 
> That's not a bad discussion; however, it is off topic in this thread
> and fairly clear to both of us as well.

	Yes. One of the reasons I was continuing this was to teach
	Kirbu a lesson about his troll posts. He still quite hasn't
	gotten the hint.

> You are correct in the
> relationship between Linux and GNU, though the path you were taking
> was not obvious to me at all.  Perhaps showing a couple of cards would
> have gotten us here quicker no?

	Yes, I suppose I could have done that (but its not usually
	wise in a debate).

> I would also point out that Linux/GNU have mutually gained from each
> others presence and development.  Today they both need each other.

	Yes, it does appear that way, doesn't it.

> Maybe GNU will finish their kernel one day so we can see what a GNU OS
> would be like.

	No doubt it will have a kernalised emacs with all the usual bells
	and whistles.

-am	� 2003

Dave Uhring wrote:
> 
> On Fri, 08 Aug 2003 21:14:41 +0000, Cypherpun wrote:
> 
> > Oh, so sorry Dave, but you've been chattering away with Anythony,
> > who has repeately (at least five in the past few days) made homophobic
> > and/or other "personal attack" comments in response to my posts.
> 
> Your entire posting was nothing but a personal attack.  Nothing else, not
> even OT matters.

	As you probably recall, Dave, he first started up here by
	following Gary Burnore and attacking him. It seems to be
	his entire rationale. He's then proceeded to try and
	justify his existence here by posting OT observations
	and providing links and posting their contents as well,
	abusing people for top posting or posting in HTML (like
	he's doing anyone a service) or calling people "maroons".

> > Feel free to continue encouraging him merely by talking to him
> > (treating him as if he were normal) while he posts crap at me
> > even when I don't mention him, and feel free to killfile me.
> 
> You would have to get pretty offensive to get into my killfile and once in
> a while your contribution here is worthwhile.

	As a script kiddie, he doesn't contribute much. By the looks
	of it, most everyone has him killfiled (except unsuspecting
	newbies. But I usually ensure that they get the hint fairly
	quickly).

	His only chance to earn some credibility and respectability is
	to change his attitude completely. Post under his real name
	(the email address doesn't matter due to the spam problem),
	post without insulting or abusing people, make no off-topic
	posts or silly references to inconsequential nonsense etc.
	Then, and only then, will he get treated as an equal.

-am	� 2003

Dave Uhring wrote:
> 
> On Fri, 08 Aug 2003 22:28:17 +0000, Cypherpun wrote:
> 
> > In this thread alone, let alone for the past few months at least,
> > he has been serially forging (exact address forgery) people's
> > accounts at hotmail.com through the Deutsche newsserver.
> 
> How are these addresses obtained?  How do you know they are legitimate
> addresses?

	They're not. Hotmail doesn't allow 2 letter/digit email addresses.

-am	� 2003

Dave Uhring wrote:
> 
> On Fri, 08 Aug 2003 23:33:47 +0000, Cypherpun wrote:
> 
> >
> > He's probably got a script that is generating two character
> > combinations for now, automatically each time he posts. He
> > started off with one character names at hotmail.com.
> >
> >>  How do you know they are legitimate addresses?
> >
> > The implicit answer is the same for domainnames: all the  one/two-letter
> > combinations will be registered.
> 
> Then your accusation about people receiving increased spam because of
> those fraudulent email addresses is unjustified.  You do not -know- that,
> do you?

	As per usual, his accusations are baseless and the typical
	ravings of a paranoid individual. The email addresses I use
	are not possible at Hotmail.

> Are you also unaware that spammers run dictionary generated mailings,
> particulary at hotmail.com?

	One of the reasons I chose Hotmail to do this was because of
	the amount of spam I do get with the legitimate Hotmail accounts
	I do have. However, this is only a minor divergence. The real
	reason was to give Charlie Brown something to do -
	http://www.freep.com/money/tech/mwend22_20021122.htm
	(I'd recommend others do likewise to really keep him busy) and
	the fact the it really riles Guy because his ISP told him to
	post with his real email address.

-am	� 2003

Anthony Mandic <sp@hotmail.com> wrote in message news:<3F345F60.E5337D58@hotmail.com>...
> Doug Dingus wrote:
> 
> > >       You could also write programs to do calculations. But note
> > >       that calculations are not learning maths. Its simply a means
> > >       of solving equations.
> > 
> > (was not factoring a computer into the simple analogy)
> 
> 	Would you believe programmable calculators?
Ok, you have me on this point.  Next time a better analogy...

> > Maybe GNU will finish their kernel one day so we can see what a GNU OS
> > would be like.
> 
> 	No doubt it will have a kernalised emacs with all the usual bells
> 	and whistles.

Dang! Had not thought of that... I'm a vi guy when forced to choose;
otherwise, nedit all the way.

Peace

Jack Hamilton wrote:
> 
> Alan Johnson <adjtech@ulster.net> wrote:
....
> >PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT
> >YOU ARE TALKING ABOUT!
> 
> How are we to know which groups those may be?  If indeed there are any
> such groups.  I'm reading the thread in comp.unix.tru64, and I think
> it's on-topic.

	A check with Google suggests he lurks in a HP newsgroup.
	He could also be a freemason.

> And please stop shouting.

	Maybe its a freemason trait.

-am	� 2003

Doug Dingus wrote:
>
> > > (was not factoring a computer into the simple analogy)
> >
> >       Would you believe programmable calculators?
> Ok, you have me on this point.  Next time a better analogy...

	I was half expecting you to reply "or a boyscout with a pocket knife?"

> > > Maybe GNU will finish their kernel one day so we can see what a GNU OS
> > > would be like.
> >
> >       No doubt it will have a kernalised emacs with all the usual bells
> >       and whistles.
> 
> Dang! Had not thought of that... I'm a vi guy when forced to choose;
> otherwise, nedit all the way.

	Yes, I prefer vi too. However, lately I'm coming across a lot
	of people with a preference for pico.

-am	� 2003

Dave Uhring wrote:

> SBC certainly has been screwing the smaller telcos.  They pay millions per
> year in fines and consider that expense to be less than their projected
> losses for obeying the laws.

	Fines? OK, you've lost me here.

> For years, as the ILEC locally, SBC prevented CLECs from utilizing their
> facilities for DSL service in violation of the telecommunications act of
> 1996.  Monopolistic practices, once learned, are not forgotten easily.

	For some reason I keep thinking of Motorola and their legion
	of satellites and those big satellite phones that didn't work
	too well. Not sure why they decided to bring some of them
	down (apart from poor patronage). Perhaps if the handsets got
	better, things would have worked out better for them.

> It's rare that the dedicated ones have any impact on the viability of the
> venture.  Worse, those workers are generally the junior ones and the first
> to be let go as profit declines.

	Not always. The only juniors when I was working with them were
	the uni students. But I understand what you mean.

-am	� 2003

"Alan Johnson" <adjtech@ulster.net> wrote in message
news:pZucnc53Ar6XpqmiXTWJiw@thebiz.net...
> PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT
> YOU ARE TALKING ABOUT!

I agree with Jack, and I'm from comp.sys.hp.hpux.
And if you *really* think that everyone in your
group feels the same way you do, then please tell
us which group that is so we know where *not* to
send these posts.

However, you could always do as I do, and simply
ignore the threads you don't wish to read.

Vu.

"Anthony Mandic" <sv@hotmail.com> wrote in message
news:3F34CA7F.DAC7A73C@hotmail.com...
> Doug Dingus wrote:
> > > > Maybe GNU will finish their kernel one day so we can see what a GNU
OS
> > > > would be like.
> > >
> > >       No doubt it will have a kernalised emacs with all the usual
bells
> > >       and whistles.
> >
> > Dang! Had not thought of that... I'm a vi guy when forced to choose;
> > otherwise, nedit all the way.
>
> Yes, I prefer vi too. However, lately I'm coming across a lot
> of people with a preference for pico.

So I'm not the only one who prefers pico???
I'm kinda learning vi, but it's so damn confusing :(

Vu.

In article <nicoya-D2804C.19020608082003@news.sc.shawcable.net>,
Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
> Other issues can be a bit more complex - for example the raw storage command 
> interface, where there is no standard nor even the slightest resemblence
> between the APIs on different platforms.

I find that these kinds of problems can often be avoided, by looking at
the APIs you're using and simply avoiding ones that aren't standardised
or are in transition.

The usual response to this suggestion is that these APIs are so much
more efficient than the standard ones than it wouldn't be practical to
use anything else. Which would have some weight if it weren't for the
tendency of the applications in question to simultaneously use expensive
libraries or components, or include expensive and cumbersome features like
"skinnable interfaces" that simultaneously make them harder to use.

> If you try to write both your application *plus* a portability library at the 
> same time, you end up effectively writing two projects at once which makes 
> things a wee bit more difficult than they need to be.

Isn't that a bit of an exaggeration? You don't need to write a portability
library to encapsulate the problematic component such that porting doesn't
turn into a complete rewrite of substantial portions of code (or, worse,
hastily writing an implementation of the parts of API #1 youhappen to use
in API #2).

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

In article <pZucnc53Ar6XpqmiXTWJiw@thebiz.net>,
Alan Johnson  <adjtech@ulster.net> wrote:
> PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT 
> YOU ARE TALKING ABOUT!

Alas, non-portable code is a plague on all the operating systems discussed
in the groups this is crossposted to.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

> > Alan Johnson  <adjtech@ulster.net> wrote:
> > PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO
> >  INTEREST IN  WHAT YOU ARE TALKING ABOUT!
>
> Alas, non-portable code is a plague on all the operating systems discussed
> in the groups this is crossposted to.
>

I'm following the thread in comp.unix.aix, and portability is no problem
with only a single platform.
But the real *NIX world is heterogenous.

And even if this thread is some kind of excessive in many ways, it is
interesting BECAUSE the initial crossposting.

---
Uli

On Sat, 09 Aug 2003 20:25:21 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:
> 
>> SBC certainly has been screwing the smaller telcos.  They pay millions per
>> year in fines and consider that expense to be less than their projected
>> losses for obeying the laws.
> 
> 	Fines? OK, you've lost me here.

http://www.eetimes.com/sys/news/OEG20010524S0064
http://www.bizjournals.com/stlouis/stories/2002/10/07/daily45.html
http://sanjose.bizjournals.com/sanantonio/stories/2002/05/27/daily5.html


> 	For some reason I keep thinking of Motorola and their legion
> 	of satellites and those big satellite phones that didn't work
> 	too well. Not sure why they decided to bring some of them
> 	down (apart from poor patronage). Perhaps if the handsets got
> 	better, things would have worked out better for them.

If useless satellites can be deorbited they should be.  There is already
enough junk in low earth orbit.

But Motorola sold that galaxy of satellites and the Iridium system.  The
satellites are still there and being used.

http://www.iridium.com/

>> It's rare that the dedicated ones have any impact on the viability of the
>> venture.  Worse, those workers are generally the junior ones and the first
>> to be let go as profit declines.
> 
> 	Not always. The only juniors when I was working with them were
> 	the uni students. But I understand what you mean.

Junior as regards job seniority, not a school class.

On Sat, 9 Aug 2003, Vu wrote:

> I'm kinda learning vi, but it's so damn confusing :(

Vi is very confusing at first, until you grok how it
works.  The it goes form very confusing to very powerful.

A long time vi user,

-- 
Rich Teer, SCNA, SCSA

President,
Rite Online Inc.

Voice: +1 (250) 979-1638
URL: http://www.rite-online.net

Rich Teer <rich.teer@rite-group.com> writes:

> On Sat, 9 Aug 2003, Vu wrote:
> 
> > I'm kinda learning vi, but it's so damn confusing :(
> 
> Vi is very confusing at first, until you grok how it
> works.  The it goes form very confusing to very powerful.

I have to disagree. vi remains very confusing no matter how much one
attempts to learn it- sort of like X keyboard bindings.  You must
think vi is powerful because it beeps a lot and messes up your text.

 
> A long time vi user,

They have 12 step programs for this kind of thing, you know.

;)

A long time emacs user,

Gregm
 

Rich Teer wrote:

> On Fri, 8 Aug 2003, Alan Johnson wrote:
> 
> 
>>PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT
>>YOU ARE TALKING ABOUT!
> 
> 
> If you'd take half a second to tell us which newsgroup(s)
> you're talking about, I'm sure we'll be happy to oblige.
> 
> Now, was it really necessary to post your comment half a
> dozen times, without snipping the articals you replied to?
> 
comp.sys.hp.hpux, which I posted a request to over a week ago requesting 
that this thread stop being posted as it was totally unrelated to HP-UX 
issues, just a general rant about pretty much everything. The reason 
that I DID post like that is to let you people see just how annoying it 
is to have this kind of thread running on a newsgroup that has no 
interest in it from my point of view, as I am sure others in the group 
will agree.

In article <bh2ur1$7gs$1@jeeves.eng.abbnm.com>,
 peter@abbnm.com (Peter da Silva) wrote:

: In article <nicoya-D2804C.19020608082003@news.sc.shawcable.net>,
: Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
: > Other issues can be a bit more complex - for example the raw storage command 
: > interface, where there is no standard nor even the slightest resemblence
: > between the APIs on different platforms.
: 
: I find that these kinds of problems can often be avoided, by looking at
: the APIs you're using and simply avoiding ones that aren't standardised
: or are in transition.
: 
: The usual response to this suggestion is that these APIs are so much
: more efficient than the standard ones than it wouldn't be practical to
: use anything else. Which would have some weight if it weren't for the
: tendency of the applications in question to simultaneously use expensive
: libraries or components, or include expensive and cumbersome features like
: "skinnable interfaces" that simultaneously make them harder to use.

Well, this discussion started with the portability concerns of sending raw 
commands to storage devices to retrieve SMART data. The APIs to do so on each OS 
come in two flavours: non-portable and non-existant.

Of course, nowadays there's the libscg compatability library to get around that, 
but if you didn't have that you'd be forced to either a: write non-portable 
code, or b: write your own portability library.


: > If you try to write both your application *plus* a portability library at the 
: > same time, you end up effectively writing two projects at once which makes 
: > things a wee bit more difficult than they need to be.
: 
: Isn't that a bit of an exaggeration? You don't need to write a portability
: library to encapsulate the problematic component such that porting doesn't
: turn into a complete rewrite of substantial portions of code (or, worse,
: hastily writing an implementation of the parts of API #1 youhappen to use
: in API #2).

Well it's certainly more trouble than it's worth when the original author has no 
immediate need to run their application on any other platform.

It's wise to make all your code clean so that you can make it portable more 
easily at a later date. It's idealistic to make all your code portable 
regardless of the needs of your current project.

For example, right now I'm writing an ethernet driver in IRIX. I could, in 
theory, make it portable between IRIX, Linux, MacOS, Solaris, QNX, BeOS, 
AmigaOS, DOS and whatnot, but that would be one excedrin-sized headache I would 
be wise to avoid. In this case, the most sane path (should I ever need the 
driver on a different platform) would be a complete rewrite when porting.


Cheers - Tony 'Nicoya' Mantler :)

-- 
Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@apia.dhs.org
--  http://www.apia.dhs.org/  --  http://nicoya.feline.pp.se/  --

In article <m33cgadb9t.fsf@europa.pienet>,
Greg Menke  <gregm-news@toadmail.com> wrote:
> I have to disagree. vi remains very confusing no matter how much one
> attempts to learn it-

Ah, then you don't know the secret.

There is a secret to understanding vi, one which was hinted at in a
recent popular movie, and once you have absorbed this knowledge and,
with some modest amount of earnest meditation, integrated it into
your worldview... vi will be no mystery at all.

I will, in defiance perhaps of common sense, reveal that secret. Bend
your eyes, concentrate your will, breathe in, pause, and then breathe
out completely. And, now...

	"There is no insert mode"

And there you have it. Understand that, and you will understand vi.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

In article <nicoya-55C0FB.13314209082003@news.wp.shawcable.net>,
Tony 'Nicoya' Mantler  <nicoya@apia.dhs.org> wrote:
> Well, this discussion started with the portability concerns of sending raw 
> commands to storage devices to retrieve SMART data. The APIs to do so on
> each OS come in two flavours: non-portable and non-existant.

Ah, well, in that circumstance there's little point in writing anything
that you intend to actually keep except by creating a wrapper, since for
that kind of tool the wrapper... the portability library, pretty much IS
the application.

> It's wise to make all your code clean so that you can make it portable more 
> easily at a later date. It's idealistic to make all your code portable 
> regardless of the needs of your current project.

If the code is written cleanly and compartmentalised it is portable, even
if it has not yet been ported.

> For example, right now I'm writing an ethernet driver in IRIX.

Ah, but a driver is in some sense an implementation of the portability
library itself, with the driver API being that library's API, as seen
from the OS side. :)

So the question here is not "is the driver portable", but "is the driver
interface portable"? In some cases it is, at least portable enough that
one can use the same driver with some modification on multiple operating
systems.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

In article <97qcnR5OcK4IuqiiXTWJhQ@thebiz.net>,
Alan Johnson  <adjtech@ulster.net> wrote:
> comp.sys.hp.hpux, which I posted a request to over a week ago requesting 
> that this thread stop being posted as it was totally unrelated to HP-UX 
> issues, just a general rant about pretty much everything.

Ah, well, as it turns out... for those of us currently working on Tru64 the
relationship between HPUX and portability issues is uniquely obvious.

I'd add a smiley except I don't much feel like smiling about it.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

peter@abbnm.com (Peter da Silva) writes:

> In article <m33cgadb9t.fsf@europa.pienet>,
> Greg Menke  <gregm-news@toadmail.com> wrote:
> > I have to disagree. vi remains very confusing no matter how much one
> > attempts to learn it-
> 
> Ah, then you don't know the secret.
> 
> There is a secret to understanding vi, one which was hinted at in a
> recent popular movie, and once you have absorbed this knowledge and,
> with some modest amount of earnest meditation, integrated it into
> your worldview... vi will be no mystery at all.

Is this like staying in the bathroom long enough that you no longer
notice the smell?
 
> I will, in defiance perhaps of common sense, reveal that secret. Bend
> your eyes, concentrate your will, breathe in, pause, and then breathe
> out completely. And, now...
> 
> 	"There is no insert mode"
> 
> And there you have it. Understand that, and you will understand vi.

So success with vi is choosing which text to overwrite?

Using vi makes me feel like there is a 3rd arm growing out of the top
of my head that keeps making obscene gestures at the passers-by.

Gregm

On Sat, 9 Aug 2003 21:57:16 +0000 (UTC), Peter da Silva <peter@abbnm.com>
wrote:
>I will, in defiance perhaps of common sense, reveal that secret. Bend
>your eyes, concentrate your will, breathe in, pause, and then breathe
>out completely. And, now...
>	"There is no insert mode"
>And there you have it. Understand that, and you will understand vi.

You know how long I've been wrestling vi. In all that time, this insight has
eluded me.

It still eludes me.

What has Stephanie been putting in your dinner?

In article <m3wudmbh0g.fsf@europa.pienet>,
Greg Menke  <gregm-news@toadmail.com> wrote:
> > 	"There is no insert mode"

> > And there you have it. Understand that, and you will understand vi.

> So success with vi is choosing which text to overwrite?

No, grasshopper, you have failed to understand.

There is no insert mode, there is no replace mode, there is no command
mode. There are no modes at all.

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

I worked for a guy that learned Teco using RSX11-M Plus. He did amazing
things, he actually wrote a small inventory management program with Teco,
and another chap I had met in the same firm wrote a Contact Management
database with Teco.

rtt

"Peter da Silva" <peter@abbnm.com> wrote in message
news:bh4ack$19pc$2@jeeves.eng.abbnm.com...
> In article <slrnbjauu2.nsq.jmaynard@thebrain.conmicro.cx>,
> Jay Maynard <jmaynard@conmicro.cx> wrote:
> > On Sat, 9 Aug 2003 21:57:16 +0000 (UTC), Peter da Silva
<peter@abbnm.com>
> > wrote:
> > >I will, in defiance perhaps of common sense, reveal that secret. Bend
> > >your eyes, concentrate your will, breathe in, pause, and then breathe
> > >out completely. And, now...
> > > "There is no insert mode"
> > >And there you have it. Understand that, and you will understand vi.
>
> > You know how long I've been wrestling vi. In all that time, this insight
has
> > eluded me.
>
> Consider TECO. Does TECO have an insert mode?
>
> -- 
> I've seen things you people can't imagine. Chimneysweeps on fire over the
roofs
> of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.
All
> these things will be lost in time, like chalk-paintings in the rain.
`-_-'
> Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?
'U`

I have seen lot of nashing teeth and grinding of molars over the years. All
kinds of things come up over the horizon thought up by some shmuck, C, C++,
C#, RAD, SQL, JAVA, .NET, etc...

It's all half baked trash. People used to have this great argument about
compiled versus interpreted. Written in a high level language for
portability or assembler for efficiency. Clap trap I say.

Standard API's libraries and what knot's. We can't even agree on a standard
interface in this world. I recently worked on a project run by a bunch of
looney tunes from Nortel, cowboys for sure, Java was chosen for it's code
reuse and portability. Funny things is, three different programmers wrote
thier own search program, rather than one person and two reusing. All kinds
of assinine stuff like that.

I want some smart stuff now, like, when I put a CD-ROM in my CD-ROM player,
I want it to tell my Amplifier to turn on at a low initial volume, switch to
CD-ROM and let fly. Or when I put a DVD in the DVD player, I want the
Amplifier to come on and the Video screen too and the whole thing to start
up. ANd I want all that to work amongst my Sony, Philips, Panasonic, NAD,
Protel, Macintosh, Peerless and Haethkit systems please.


Standard API's, puh lease. ;-)

Vu wrote:

> > Yes, I prefer vi too. However, lately I'm coming across a lot
> > of people with a preference for pico.
> 
> So I'm not the only one who prefers pico???

	The guys using it are all fairly young (compared to me anyway).
	It was probably all the rage at uni when they were studying.
	It is a fairly simple editor.

> I'm kinda learning vi, but it's so damn confusing :(

	Its not very hard. Once you unleash its power, there's no looking
	back.

-am	� 2003

Dave Uhring wrote:

> >       Fines? OK, you've lost me here.
> 
> http://www.eetimes.com/sys/news/OEG20010524S0064
> http://www.bizjournals.com/stlouis/stories/2002/10/07/daily45.html
> http://sanjose.bizjournals.com/sanantonio/stories/2002/05/27/daily5.html

	OK, I get the picture. Much the same thing was happening out here.

> If useless satellites can be deorbited they should be.  There is already
> enough junk in low earth orbit.

	But were they really useless? From my recollection, they weren't.
	Agree about the junk though.

> >       Not always. The only juniors when I was working with them were
> >       the uni students. But I understand what you mean.
> 
> Junior as regards job seniority, not a school class.

	Yes, I was saying much the same thing. The students who had parttime
	summer jobs were the juniors and did the lowest level work. All the
	permanents had been there for years.

-am	� 2003

Cypherdork@nyc.rr.com wrote:

> Who's "Guy Polis"?

	Dunno. Do you?

> >       You keep saying that. I'm not sure why you think I'm homophibic.
> 
> Why, it might have something to do with calling me 'slugger', 'plugger',
> 'rent boy' and posting an URL to a HS for gays and asking if I had applied.

	Yeah, I had all those nicknames mapped out ages ago. Then it was
	just a matter of you making the right keyword for the match (I
	was going to mention the NYGP last year when you mentioned the
	Euro love fest but missed the opportunity because you atarted
	laying low for a while. No matter, this year was just as good.
	Tugger was a good one. You mentioned South Park so I took an
	opportunity to watch it. And what do you think showed up? Some
	Australian taking his tugboat to New York and beating up the
	maroons. Brilliant! Slugger, of course, is in honour of the
	Yankees; or maybe the NY Mets).

> >       Yes, I know. Until Hotmail do something like validating incoming
> >       emails (which should reduce spam there by over 95%), I'm not
> >       concerned by how little extra spam my email addresses cause them.
> >       If anything, Charlie Brown might decide its not worth bothering
> >       with Hotmail at all.
> 
> Charlie Brown? That used to be the name of the chairman of AT&T circa 1981.
> 
> Who are you referring to? The chairman of another company?

	Since when is it my problem if you're too lazy to cut and paste
	the link I gave you and read it for yourself.

> >  > I don't know or care what the URL is, as usual. Last time it was
> >  > homophobic stuff. I'm sure it's even more clever this time.
> >
> >       No, its legitimate.
> 
> Sorry, I'm still not pulling up your URL references.

	Pity, its the sort of thing you'd go ballistic over.

> >       So why did you change it?
> 
> Because 'cypher@punk.net' was a sinkhole anyone could use.
> I've gotten a number of nutjobs spanked who forged me, something
> I couldn't do with a non-replyable address. Since I use this
> address only for Usenet, I expect it to be a spam attractor,
> and so it's easy to delete the spam since it comes in large
> chunks.

	Do you get any of the forged email bounces as discussed in
	RISKS 21.76 and later? -
http://groups.google.com/groups?selm=CMM.0.90.4.1006278573.risko%40chiron.csl.sri.com&oe=UTF-8&output=gplain

	Just another reason not to advertise a real email address
	on Usenet.

> >       That's what I read in someone else's post. You didn't deny it
> >       at the time when they said you got spanked.
> 
> Oh, what a genius! So that's why you think my name is 'Guy Polis' too.

	Do you think I really care?

> Gee, then I must have HIV/AIDS, am gay, have a micropenis (the last one
> is by youz BTW ;-), that I followed Burnore into CUS, that most people
> have me plonked, ugh I don't even remember all the crap nutjobs like
> you have posted. And I've never had an account spnaked.

	OK, if you say so, I believe you.

> >       You just walked into that one too easily.
> 
> You must mean I responded to a nutjob.

	You mean there's more than one of you?

> >       I don't change my real name so they are free to killfile me
> >       on that if they so wish.
> 
> Addresses are unique, names aren't.

	Really?

> Besides, there's not the slightest reason for you to morph
> your address.

	I did use a real email address at Start once (before they stopped
	their free email service). I just make it look like a spam-blocked
	one. It didn't stop the clueless rubes trying to email me. What was
	interesting, though, was the amount of spam. Its better to push the
	problem back to them.

-am	� 2003

In article <bh4aaf$19pc$1@jeeves.eng.abbnm.com>,
 peter@abbnm.com (Peter da Silva) wrote:

[...]
: There is no insert mode, there is no replace mode, there is no command
: mode. There are no modes at all.

Nonsense, there's two modes: the one in which it beeps at you, and the one in 
which it doesn't.

I still remember the first time I ran into vi. Nothing but a screen full of ~s, 
accompanied by the sinking realization that I can't, for the life of me, 
remember how to exit.

Oh, to be young again.


Cheers - Tony 'Nicoya' Mantler :)

-- 
Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@apia.dhs.org
--  http://www.apia.dhs.org/  --  http://nicoya.feline.pp.se/  --

Anythony Manic <anthony@isug.com> wrote:
>  Cypherpunk@nyc.rr.com wrote:
>   
>  You keep saying that. I'm not sure why you think I'm homophibic.
>   
>  > Why, it might have something to do with calling me 'slugger', 'plugger',
>  > 'rent boy' and posting an URL to a HS for gays and asking if I had applied.
>   
>   	Yeah, I had all those nicknames mapped out ages ago. Then it was
>   	just a matter of you making the right keyword for the match (I
>   	was going to mention the NYGP last year when you mentioned the
>   	Euro love fest but missed the opportunity because you started
>   	laying low for a while. No matter, this year was just as good.
>   	Tugger was a good one. You mentioned South Park so I took an
>   	opportunity to watch it. And what do you think showed up? Some
>   	Australian taking his tugboat to New York and beating up the
>   	maroons. Brilliant! Slugger, of course, is in honour of the
>   	Yankees; or maybe the NY Mets).

Oh, yeah, right. Reality check:

Anythony Manic wrote in message ID 3CF4B55A.58632455@start.com.au:
#
#    [Don Kool's] buttplugged boyfield [is cypherpunk].

----

Anythony Manic <anthony@isug.com> wrote:
>  Cypherpunk@nyc.rr.com wrote:
>
>  So why did you change it?
>   
>  > Because 'cypher@punk.net' was a sinkhole anyone could use.
>  > I've gotten a number of nutjobs spanked who forged me, something
>  > I couldn't do with a non-replyable address. Since I use this
>  > address only for Usenet, I expect it to be a spam attractor,
>  > and so it's easy to delete the spam since it comes in large
>  > chunks.
>   
>   	Do you get any of the forged email bounces as discussed in
>   	RISKS 21.76 and later? -
>   http://groups.google.com/groups?selm=CMM.0.90.4.1006278573.risko%40chiron.csl.sri.com&oe=UTF-8&output=gplain
>   
>   	Just another reason not to advertise a real email address
>   	on Usenet.

A real email address is only a problem like that when
there are kooks actively operating with your address.

All I have to do is zap 'cypherpunk' via a screen menu and
instantiate another one, should I feel I need to.

>   
>  I don't change my real name so they are free to killfile me
>  on that if they so wish.
>   
>  > Addresses are unique, names aren't.
>   
>   	Really?

Really. Here Google...

http://www.cotse.net/tos.html#forging
#    
#    Q: Will we terminate an account for forging?
#    
#    A: Yes. If it really is a forge. A forge is not having the same 
#    name, it is having the same e-mail address. If a user is posting 
#    with your e-mail address in their from line we will terminate 
#    the account.

>   
>  > Besides, there's not the slightest reason for you to morph
>  > your address.
>   
>   	I did use a real email address at Start once (before they stopped
>   	their free email service). I just make it look like a spam-blocked
>   	one. It didn't stop the clueless rubes trying to email me. What was
>   	interesting, though, was the amount of spam. Its better to push the
>   	problem back to them.

What do you do for <mailto:anthony@isug.com> ? ;-)

>   
>   -am	� 2003

No need to copyright yourself. No one wants to be you.

In article <_0jZa.6098$ox5.711003@news20.bellglobal.com>,
Richard Tomkins <tomkinsr@@istop.com> wrote:
> I worked for a guy that learned Teco using RSX11-M Plus. He did amazing
> things, he actually wrote a small inventory management program with Teco,
> and another chap I had met in the same firm wrote a Contact Management
> database with Teco.

And does TECO have an insert mode?

-- 
I've seen things you people can't imagine. Chimneysweeps on fire over the roofs
of London. I've watched kite-strings glitter in the sun at Hyde Park Gate.  All
these things will be lost in time, like chalk-paintings in the rain.   `-_-'
Time for your nap.  | Peter da Silva | Har du kramat din varg, idag?    'U`

On Sun, 10 Aug 2003 14:11:41 +1000, Anthony Mandic wrote:

> Dave Uhring wrote:

>> If useless satellites can be deorbited they should be.  There is already
>> enough junk in low earth orbit.
> 
> 	But were they really useless? From my recollection, they weren't.

They were useless to their owner since Motorola had decided to discontinue
the service; that's the only utility which counts.  It's a "good thing"
that the service was rescued by a buyer; there is no alternative for the
people who are able and willing to pay for the service.

IIRC, the US military used Iridium during the Afghanistan and Iraq
actions.  The high orbit military satellites were saturated with all the
RT video imaging traffic from the UAVs.

> 	Yes, I was saying much the same thing. The students who had parttime
> 	summer jobs were the juniors and did the lowest level work. All the
> 	permanents had been there for years.

Yes, a confusing coincidence of word usage.

"Alan Johnson" <adjtech@ulster.net> wrote in message
news:97qcnR5OcK4IuqiiXTWJhQ@thebiz.net...
> Rich Teer wrote:
>
> > On Fri, 8 Aug 2003, Alan Johnson wrote:
> >
> >
> >>PLEASE STOP CROSS POSTING TO NEWSGROUPS THAT HAVE NO INTEREST IN WHAT
> >>YOU ARE TALKING ABOUT!
> >
> >
> > If you'd take half a second to tell us which newsgroup(s)
> > you're talking about, I'm sure we'll be happy to oblige.
> >
> > Now, was it really necessary to post your comment half a
> > dozen times, without snipping the articals you replied to?
> >
> comp.sys.hp.hpux, which I posted a request to over a week ago requesting
> that this thread stop being posted as it was totally unrelated to HP-UX
> issues, just a general rant about pretty much everything. The reason
> that I DID post like that is to let you people see just how annoying it
> is to have this kind of thread running on a newsgroup that has no
> interest in it from my point of view, as I am sure others in the group
> will agree.
>

lol - I'm from the same group, and I've been reading this with a bit
of interest, so if people stopped posting to this group, I'd miss out,
cos not everyone in the HP-UX group thinks it's pointless.

If you don't care for this topic, ignore it. Simple.

Vu.

"Anthony Mandic" <t5@hotmail.com> wrote in message
news:3F35C39E.76C61E84@hotmail.com...
> Vu wrote:
>
> > > Yes, I prefer vi too. However, lately I'm coming across a lot
> > > of people with a preference for pico.
> >
> > So I'm not the only one who prefers pico???
>
> The guys using it are all fairly young (compared to me anyway).
> It was probably all the rage at uni when they were studying.
> It is a fairly simple editor.

I can't remember why I started using pico, must have been when
I first got a unix account at uni. I don't wade through thousands
of lines of code, so pico suits me fine. Except when I forget to
install it. Took me ages to figure out that it's included in the
'Pine' package :)

>
> > I'm kinda learning vi, but it's so damn confusing :(
>
> Its not very hard. Once you unleash its power, there's no looking
> back.

I keep hearing that, but what I like about pico is that the shortcut
commands are there in front of you, while with vi, i have to remember
things like that, then forget them, and end up doing something stupid.
The only shortcut i remember is :q - and I found that doesn't always
work either :P

Vu.

Vu wrote:

> I can't remember why I started using pico, must have been when
> I first got a unix account at uni. I don't wade throu