Hi!

I wonder, what do you think, will Solaris really slowly die, and be
replaced with GNU-izied OpenSolaris? As far as I can see from
opensolaris developers and Sun personel, that is the intended way for
the future...

What do you think about that? I'm sad to see it happening, but with Ian
Murdoch being the chief, it seems inevitable...




-- 
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |

Jakov Sosic wrote:

>  I wonder, what do you think, will Solaris really slowly die, and be
>  replaced with GNU-izied OpenSolaris? As far as I can see from
>  opensolaris developers and Sun personel, that is the intended way for
>  the future...

What is exactly GNU-izied OpenSolaris?

>  What do you think about that? I'm sad to see it happening, but with Ian
>  Murdoch being the chief, it seems inevitable...

Some of the free software utilities are extremely useful. Some are less
so. Just like with any other large group of software. I don't see what
that has to do with Ian Murdock.

Besides, any reasonably competent chief has to look in the future and
not in the past. OS distributions with GNU "personality" have been done.
There is no money, no market share, no vibrant community and no nothing
in doing exactly that again. Free individuals could do that as a hobby,
but commercial company cannot afford that luxury.

Except that some people might think that it would be beneficial if
people familiar with Linux could feel more familiar with Solaris. Just
like AIX got Linux personality almost a decade ago. But implementing
that is hardly a vision for the future. It's in the "must have these
days" department. Just like POSIX compliance once was.

-- 
 .-.   .-.    Yes, I am an agent of Satan, but my duties are largely
(_  \ /  _)   ceremonial.
     |
     |        dave@fly.srk.fer.hr

On 2009-04-18, Drazen Kacar <dave@fly.srk.fer.hr> wrote:

> What is exactly GNU-izied OpenSolaris?

Solaris with bash as default and root shell, GNU coreutils replacing
SysV, and so on.


> Besides, any reasonably competent chief has to look in the future and
> not in the past. OS distributions with GNU "personality" have been done.
> There is no money, no market share, no vibrant community and no nothing
> in doing exactly that again. Free individuals could do that as a hobby,
> but commercial company cannot afford that luxury.

You mean Nexenta? Well it's happening again with OpenSolaris :) What I
am interested in how will Sun succed to replace Solaris 10 with OpenSolaris
in the enterprise... Maybe we are about to witness another demise.


> Except that some people might think that it would be beneficial if
> people familiar with Linux could feel more familiar with Solaris. Just
> like AIX got Linux personality almost a decade ago. But implementing
> that is hardly a vision for the future. It's in the "must have these
> days" department. Just like POSIX compliance once was.

I don't see point in canibalizing SysV just for the sake of few
newcomers to feel more welcome. You could just rebrand Ubunto to Sun
Something, and they would feel really welcome.

I mean, you have different package management, you have different
kernel, different init system, scripts, management... And you think
someone will feel more welcome if you drop out sh and replace it with
bash? I don't get it.


-- 
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |

Jakov Sosic wrote:
>  On 2009-04-18, Drazen Kacar <dave@fly.srk.fer.hr> wrote:
> 
> > What is exactly GNU-izied OpenSolaris?
> 
>  Solaris with bash as default and root shell, GNU coreutils replacing
>  SysV, and so on.

That? Why fussing about that? I couldn't care less about the default
root shell because I've been execing zsh from root's .profile (or
equivalent) when my user becomes root for ages. I never understood how
the default root shell managed to become some kind of symbol. And what
in the world it's supposed to symbolize.

Having GNU coreutils in the default Solaris would be pretty nice. And so
on.

> > in doing exactly that again. Free individuals could do that as a hobby,
> > but commercial company cannot afford that luxury.
> 
>  You mean Nexenta?

I don't mean any concrete example. I meant on principle. Commercial
companies have far less freedom than private individuals. But they have
more power, if you care about that kind of power. You have the choice,
as always. You just can't have both, as usual.

>  Well it's happening again with OpenSolaris :)

Is it? It looks like a confused dot.com company to me.

>  What I am interested in how will Sun succed to replace Solaris 10
>  with OpenSolaris in the enterprise... Maybe we are about to witness
>  another demise.

Why should they? What's wrong with releasing Solaris 11?

>  I don't see point in canibalizing SysV just for the sake of few
>  newcomers to feel more welcome.

They are not few. Unless you mean that only a few would come even if
Linux personality is implemented. Besides, Solaris has a long history of
being compatible with itself. There would be new directories to put in
the PATH, and other confusing mess like that.

>  You could just rebrand Ubunto to Sun Something, and they would feel
>  really welcome.

They, as far as I can tell, don't care. I don't care either. I don't
know who cares. Who's the customer for that effort?

>  I mean, you have different package management, you have different
>  kernel, different init system, scripts, management... And you think
>  someone will feel more welcome if you drop out sh and replace it with
>  bash? I don't get it.

Well, certain scripts would work out of the box. That's important for
people who can't write shell scripts and don't want to learn. It's also
important to those who can, but don't like to do it.

Different kernel shouldn't matter much. Debian just adopted FreeBSD
kernel for one of its ports, IIRC. There are libraries on top of the
kernel and their job is to isolate applications from different kernels.

One init system is a mess just like any other init system. SMF has
potential, but I still don't know whether to like it or not. It's not a
pain in the ass if you have shell completions for svcadm and friends.
But without... ugh.

I never liked shell scripts, so I don't care if they are a different
kind of ugly or the same kind of ugly.

I don't know what your "management" is supposed to be. If you mean
Solaris command line utilities, then supplying zsh (and bash, if
necessary) completions would do wonders.

-- 
 .-.   .-.    Yes, I am an agent of Satan, but my duties are largely
(_  \ /  _)   ceremonial.
     |
     |        dave@fly.srk.fer.hr

Jakov Sosic wrote:
> Hi!
> 
> I wonder, what do you think, will Solaris really slowly die, and be
> replaced with GNU-izied OpenSolaris? As far as I can see from
> opensolaris developers and Sun personel, that is the intended way for
> the future...
> 
> What do you think about that? I'm sad to see it happening, but with Ian
> Murdoch being the chief, it seems inevitable...
> 

Yes... I'm going to miss the broken patch mgmt that Sun provided.
:)

Hate to see everything working... broken software means more help
requests = job security.

I told you what I'm going to miss... what are you going to miss?

Jakov Sosic wrote:
> Hi!
> 
> I wonder, what do you think, will Solaris really slowly die, and be
> replaced with GNU-izied OpenSolaris? As far as I can see from
> opensolaris developers and Sun personel, that is the intended way for
> the future...
> 
> What do you think about that? I'm sad to see it happening, but with Ian
> Murdoch being the chief, it seems inevitable...
> 
> 
> 
> 

No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
Solaris 11 will still have much of the Solaris we have today with a lot 
of the newer gnome packages.  It should still offer binary compatibility 
as in the past.  Opensolaris will not offer binary compatibility.  The 
only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
the Sparc platform is XSun.  Without this package, I cannot upgrade my 
Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
graphics card to replace the xvr1200.  As far as I know, there is no 
such animal.

Paul

On 2009-04-19, Paul Gress <pgress@pb.net> wrote:

> No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
> Solaris 11 will still have much of the Solaris we have today with a lot 
> of the newer gnome packages.  It should still offer binary compatibility 
> as in the past.  Opensolaris will not offer binary compatibility.  The 
> only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
> the Sparc platform is XSun.  Without this package, I cannot upgrade my 
> Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
> graphics card to replace the xvr1200.  As far as I know, there is no 
> such animal.

So you basically think that SX:CE will be Solaris Next, and not Indiana
with all the new stuff?



-- 
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |

Paul Gress wrote:
> Jakov Sosic wrote:
>> Hi!
>>
>> I wonder, what do you think, will Solaris really slowly die, and be
>> replaced with GNU-izied OpenSolaris? As far as I can see from
>> opensolaris developers and Sun personel, that is the intended way for
>> the future...
>>
>> What do you think about that? I'm sad to see it happening, but with Ian
>> Murdoch being the chief, it seems inevitable...
>>
>>
>>
>>
> 
> No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
> Solaris 11 will still have much of the Solaris we have today with a lot 
> of the newer gnome packages.  It should still offer binary compatibility 
> as in the past.  Opensolaris will not offer binary compatibility.  The 
> only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
> the Sparc platform is XSun.  Without this package, I cannot upgrade my 
> Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
> graphics card to replace the xvr1200.  As far as I know, there is no 
> such animal.
> 
> Paul
Having survived the migration from SunOS 3.4 to Solaris 1 (BSD to SysV 
migration) including the GNU utilities is just another migration process 
as the OS evolves.
In respect of the SunBlade 2500, Sun don't make UltraSPARC workstations 
so sadly you will need to think about replacing it with an x64 machine.
UltraSPARC appears to be only a server based now.

In article <6MmdnRGchOP7rHbUnZ2dnUVZ8t6dnZ2d@pipex.net>,
	solx <nospam@example.net> writes:
> In respect of the SunBlade 2500, Sun don't make UltraSPARC workstations 
> so sadly you will need to think about replacing it with an x64 machine.
> UltraSPARC appears to be only a server based now.

There are no more Ultrasparc servers now either (except possibly
48VDC Netras for the Telco industry). Sparc servers are T series
(Niagra and decendents) or M series (sparc64 VI and later), in
addition to the x64 servers and workstations of course.

-- 
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

Jakov Sosic wrote:
> On 2009-04-19, Paul Gress <pgress@pb.net> wrote:
> 
>> No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
>> Solaris 11 will still have much of the Solaris we have today with a lot 
>> of the newer gnome packages.  It should still offer binary compatibility 
>> as in the past.  Opensolaris will not offer binary compatibility.  The 
>> only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
>> the Sparc platform is XSun.  Without this package, I cannot upgrade my 
>> Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
>> graphics card to replace the xvr1200.  As far as I know, there is no 
>> such animal.
> 
> So you basically think that SX:CE will be Solaris Next, and not Indiana
> with all the new stuff?
> 
> 
> 

Yes, I also believe Sun announced this, but I'm not going to search for 
it.  Solaris 11 will be based off of Opensolaris source, but not the 
same.  It will also include closed binary for backwards compatibility. 
They may even include all the gnu utilities, but not in the Path first.

As for my Blade 2500, I'm taking the time to write it off in my mind, 
when it's time to replace it most likely it's going to be an X64 based 
workstation.  Maybe an Ultra 27.  I haven't decided yet.  I'm also 
looking at the XI Computer Workstations 
(http://www.xicomputer.com/products/mtowerxeon.asp), they seem to be 
better equipt then the Ultra 27, capable of much more ram, and dual 5500 
series processors.  Sun should take a look at them.  Here we have a 
small workstation company outperforming what Sun offers.

Paul

quoting Jakov Sosic (Sun, 19 Apr 2009 11:33:11 +0200):
> So you basically think that SX:CE will be Solaris Next, and not
> Indiana with all the new stuff?

I've read articles (from inside SUN) stating that SXCE will be replaced
by OpenSolaris. A bit like happened to the SXDE. It just seized to
exist.

-- 
Dick Hoogendijk -- PGP/GnuPG key: 01D2433D
+http://nagual.nl/ | SunOS 10u6 10/08 ZFS+

Dick Hoogendijk wrote:
> quoting Jakov Sosic (Sun, 19 Apr 2009 11:33:11 +0200):
>> So you basically think that SX:CE will be Solaris Next, and not
>> Indiana with all the new stuff?
> 
> I've read articles (from inside SUN) stating that SXCE will be replaced
> by OpenSolaris. A bit like happened to the SXDE. It just seized to
> exist.
> 

SXCE will be replaced by Opensolaris externally.  SXCE will still be 
developed for Solaris next internally, as it use to be prior to Opensolaris.

On 2009-04-19, Paul Gress <pgress@pb.net> wrote:

> SXCE will be replaced by Opensolaris externally.  SXCE will still be 
> developed for Solaris next internally, as it use to be prior to Opensolaris.

This doesn't make sense to me, although I would like it to happen this
way though...

In the other hand, Sun is well know for those 'doesnt make sense' moves :)



-- 
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |

"Jakov Sosic" <jsosic@jsosic.homeunix.org> wrote in message news:6bmpb6-m93.ln1@pc-jsosic.srce.hr...
> Hi!
>
> I wonder, what do you think, will Solaris really slowly die, and be
> replaced with GNU-izied OpenSolaris? As far as I can see from
> opensolaris developers and Sun personel, that is the intended way for
> the future...
>
> What do you think about that? I'm sad to see it happening, but with Ian
> Murdoch being the chief, it seems inevitable...
>
>

Sun needs urgently to reconsider its enterprise customers, and
slow down the pace of change. Ironically, the linux vendors realised
this years ago so, for instance, Red Hat has Fedora to trial the cutting
edge stuff, and RHEL which changes at a more glacial pace.

You might say OpenSolaris leads Solaris but Solaris 10 still has
too many changes in my view, both from Solaris 9 and between
releases.

Sun could also learn from Microsoft, and ensure that before something
is released, it is checked for usability and consistency. Quick: what
is the name of that SMF service you need to start: is it named after
its functionality, its daemon, or something a bit like it?

Sun's hardware range too has long been an incoherent mess, and
why do they keep mucking around with the names? The successor
to the 880 was the 890, which begot, what, the M5000?
Whereas HP Proliant 580 went from 580 G3 to 580 G4 to 580 G5.

It ought to be easy to buy Sun, and it isn't.

And what is the point of restricting information about bugs, patches
and so on? Who gains from Sun system administration getting harder?
IBM? HP? Certainly not Sun because if problems are harder to
solve, perceived reliability goes down and cost of ownership rises.

And it is very nice that Sun has spent billions of dollars so we can run
MySql on Solaris 10 on laptops in our bedrooms but, really, that is not
where the money is.

-- 
John.

Jakov Sosic wrote:
>  On 2009-04-19, Paul Gress <pgress@pb.net> wrote:
> 
> > SXCE will be replaced by Opensolaris externally.  SXCE will still be 
> > developed for Solaris next internally, as it use to be prior to Opensolaris.
> 
>  This doesn't make sense to me, although I would like it to happen this
>  way though...

Did you ever try to imagine what software development process looks like
when you have millions of lines of code *and* you want to release stable
and reliable product?

-- 
 .-.   .-.    Yes, I am an agent of Satan, but my duties are largely
(_  \ /  _)   ceremonial.
     |
     |        dave@fly.srk.fer.hr

John L wrote:

> Sun could also learn from Microsoft, and ensure that before something
> is released, it is checked for usability and consistency. 

Like Vista?

-- 
Ian Collins

John L wrote:
> "Jakov Sosic" <jsosic@jsosic.homeunix.org> wrote in message news:6bmpb6-m93.ln1@pc-jsosic.srce.hr...
>> Hi!
>>
>> I wonder, what do you think, will Solaris really slowly die, and be
>> replaced with GNU-izied OpenSolaris? As far as I can see from
>> opensolaris developers and Sun personel, that is the intended way for
>> the future...
>>
>> What do you think about that? I'm sad to see it happening, but with Ian
>> Murdoch being the chief, it seems inevitable...
>>
>>
> 
> Sun needs urgently to reconsider its enterprise customers, and
> slow down the pace of change. Ironically, the linux vendors realised
> this years ago so, for instance, Red Hat has Fedora to trial the cutting
> edge stuff, and RHEL which changes at a more glacial pace.
> 
> You might say OpenSolaris leads Solaris but Solaris 10 still has
> too many changes in my view, both from Solaris 9 and between
> releases.
> 
> Sun could also learn from Microsoft, and ensure that before something
> is released, it is checked for usability and consistency. Quick: what
> is the name of that SMF service you need to start: is it named after
> its functionality, its daemon, or something a bit like it?
> 
> Sun's hardware range too has long been an incoherent mess, and
> why do they keep mucking around with the names? The successor
> to the 880 was the 890, which begot, what, the M5000?
> Whereas HP Proliant 580 went from 580 G3 to 580 G4 to 580 G5.

The problem for Sun is that it has been forced into the most competitive 
market having to sell against HP and Dell in the x64 server and 
workstation arena. Not only that it has an operating system Solaris 
which can run on its competitor's computers. Sun hardware too expensive 
then use HP, Dell or local company which uses Gigabyte, ASUS or 
Supermicro motherboards. Sun needs to generate a revenue stream from 
support on the x64 platform.
Sun should have in my opinion been pushing the 64bit capabilities of 
Solaris. I found it very funny that while running a 64bit desktop for 
years, it was only last year that my Windows colleagues migrated in XP 
x64 and even then driver support was very patchy.
The problem for Sun is that they are heavily reliant on the sales of 
Sparc based hardware but they are not producing a T series workstation 
for developers to work on. So developers either have to switch to 
Solaris x86 or Windows or Linux not a particularly good position for Sun 
to put its customers in. Some companies will switch to Windows as they 
see it as an opportunity to get rid of Solaris, some switch to Linux and 
a few might switch to Solaris/OpenSolaris. Where companies are reliant 
on 3rd Party libraries paying for Solaris/OpenSolaris x86 support 
requires effort but they probably already have a Linux port so the 
decision could made on cost.
For Sun and Solaris/OpenSolaris to survive, applications need to be 
easily ported. Having tried to build a few applications, it has been a 
long annoying process. The situation has definitely improved compared to 
three years ago but it would be good to be able to download prebuilt 
libraries that are part of Solaris/OpenSolaris and were selectable.
I have used sunfreeware, blastwave  and opencsw but if I am building an 
application I would have to libraries available as part of the core OS.
It is a pity that the maintainers of these sites could not provide the 
libraries though the Software Management facility as part of OpenSolaris.

> 
> It ought to be easy to buy Sun, and it isn't.

Sun has a lot of resellers which have traditionally provided presales 
and postsales support.
> 
> And what is the point of restricting information about bugs, patches
> and so on? Who gains from Sun system administration getting harder?
> IBM? HP? Certainly not Sun because if problems are harder to
> solve, perceived reliability goes down and cost of ownership rises.
They need to make money from support.
> 
> And it is very nice that Sun has spent billions of dollars so we can run
> MySql on Solaris 10 on laptops in our bedrooms but, really, that is not
> where the money is.
MySQL was a ill thoughtout purchase unless Sun is actually making money 
from it.

> 

On Apr 20, 12:47 am, Ian Collins <ian-n...@hotmail.com> wrote:
> John L wrote:
> > Sun could also learn from Microsoft, and ensure that before something
> > is released, it is checked for usability and consistency.
>
> Like Vista?
>

Probably more like XP.  Vista (and ME) demonstrate that everybody
makes misteaks!

;-)

John L wrote:
> "Jakov Sosic" <jsosic@jsosic.homeunix.org> wrote in message news:6bmpb6-m93.ln1@pc-jsosic.srce.hr...
>> Hi!
>>
>> I wonder, what do you think, will Solaris really slowly die, and be
>> replaced with GNU-izied OpenSolaris? As far as I can see from
>> opensolaris developers and Sun personel, that is the intended way for
>> the future...
>>
>> What do you think about that? I'm sad to see it happening, but with Ian
>> Murdoch being the chief, it seems inevitable...
>>
>>
> 
> Sun needs urgently to reconsider its enterprise customers, and
> slow down the pace of change. Ironically, the linux vendors realised
> this years ago so, for instance, Red Hat has Fedora to trial the cutting
> edge stuff, and RHEL which changes at a more glacial pace.
> 
> You might say OpenSolaris leads Solaris but Solaris 10 still has
> too many changes in my view, both from Solaris 9 and between
> releases.
> 

Changes that break existing functionality are a really bad idea!  New 
features are a necessity if you hope to sell new copies to customers.

There is also the competition to consider.  If your O/S is not "better", 
in some sense, why should anyone buy it?  I know that Solaris is "free" 
but there is still the problem of making money.  Sun charges for service 
and support.

On 2009-04-20, solx <nospam@example.net> wrote:

> The problem for Sun is that it has been forced into the most competitive 
> market having to sell against HP and Dell in the x64 server and 
> workstation arena. Not only that it has an operating system Solaris 
> which can run on its competitor's computers. Sun hardware too expensive 
> then use HP, Dell or local company which uses Gigabyte, ASUS or 
> Supermicro motherboards. Sun needs to generate a revenue stream from 
> support on the x64 platform.

Sun x64 servers are not more expensive than the servers from Dell, HP or
IBM in the same segment. But problem is that probably Sun doesn't get
enough in return from x64 server arena...



-- 
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |

On 2009-04-20, Drazen Kacar <dave@fly.srk.fer.hr> wrote:

> Did you ever try to imagine what software development process looks like
> when you have millions of lines of code *and* you want to release stable
> and reliable product?

No?


-- 
|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
=================================================================
| start fighting cancer -> http://www.worldcommunitygrid.org/   |

Most existing users of Solaris (or probably any OS) IMO mainly care that an upgrade
doesn't break their apps or scripts.  That means any additional functionality in any
given executable (at a particular location) has to be extremely backwards-compatible.

Most new users probably want functionality closest to whatever they're already familiar with.
Exluding OSs that are way too different (Windows) leaves Linux and Mac OS X as the big numbers.

I've used every version of Solaris since forever, and various SunOS 3.x and 4.x, version 7,
Xenix, Apollos, OSF1, SVR2-4, and even a tiny bit of AIX and HP/UX over the years.
And I've tried OpenSolaris (i.e.the distro - Indiana).  Mostly it's not a major problem,
although I can certainly imagine some difficulties.  On the administrative side, I'd say there
are a few pieces missing before an Indiana-based distro would be supportable on
a large scale; particularly on the installation and packaging front.

There's another dichotomy that IMO relates partially to the existing-vs-new users one:
desktop vs server.  New mindshare comes from desktop, but a lot of the cutesy stuff
that makes getting into a new desktop environment easy and pleasant is irrelevant on
a server.  And servers are where the money is at, both in sales and suppport contracts.
Yes, big businesses will pay for support for desktops too.  But small businesses may not,
unless they're using their systems as both servers and desktops.

Further, some GNU versions of utilities are indeed better, but others are not.  And some
have features that many people like, like color for ls, but lack features that the Solaris
version has (like support for displaying info about ACLs and extended attributes).  In some
cases, what is happening is that those options and features of the GNU version that are not
incompatible with the Solaris version are being added (re-created, due to license difference)
to the Solaris version in /usr/bin, which the vanilla GNU version would be available in some
other location.  IMO, that's pretty reasonable, and anyone that thinks that all Solaris versions
should be replaced by GNU versions even if that would be incompatible, should be invited to either
experience for themselves the breakage that would cause (I'm being very nice, at first I was
thinking more along the lines of _just_die_) or they should roll their own distro (or pick one
someone else already did that to, like Nexenta or Belenix or whatever).

IMO, in the basics, the Solaris kernel is vastly better than the Linux kernel, since it was
written by engineers with design principles like stability of advertised interfaces in mind,
not by a bunch of (admittedly talented) wannabes throwing their desired features out there
and seeing which ones Linus would be ok with.  Emerging standards are supported, but not in a
way that breaks compatibility for existing apps.

However, on commodity hardware, Solaris has historically been way short of device drivers.
That's changing; and when FUSE support, decent audio via a port of OSS modified significantly
for Solaris's peculiar needs (think a Sun Ray server, where you've got not one but possibly
hundreds of desktops running off of it and using audio) get integrated, together with all
the other drivers that have been and are being added, I think it will be close enough for
most people, unless they want to run on really old, free, second-hand hardware.  For those
that do, I'd point out that old hardware probably isn't power-efficient, and that there's a lot
of new, low-cost stuff out there - netbooks, the Eee PC, and so on.  Give or take some WiFi
problems still being worked, from what I've read, OpenSolaris pretty much works on the Eee PC.
That's a pretty cheap box, IMO; and assuming that it lasts that long, would probably save its
own price in electricity over a few years compared to some old 2nd-hand box.

And the Milax distro is said to run on systems with as little as 256MB of RAM, not that I'd
want to...

Because the GPL is interpreted as not allowing linking code between GPL licensed modules
and modules that are under a license that in some ways is more restrictive (even if arguably
freer in others), the goodies that originated with the Solaris kernel (dtrace, zfs, zones being
the most well known but by no means the only ones) probably won't end up in the Linux kernel;
either they'll have to re-invent them, or do klugey stuff like zfs in FUSE.  But others with
less ideological approaches to licences (FreeBSD and Darwin/Mac OS X) have or are adopting
some of those (dtrace and zfs at least) from Solaris.  And while Solaris can't use Linux drivers
due to the license difference, the internals are different enough that it wouldn't be a huge
help anyway in most cases; and Solaris certainly has gotten its share of drivers from the
*BSDs, for example (still needing lots of changes, but given the common ancestry, less than
they'd have needed porting a Linux driver even if that were possible).

So the driver situation is already better, and before too long, should be much better.

There are at least half a dozen different distros based on OpenSolaris, some having done
some serious work of their own to address particular concerns, some of which in turn is
finding its way back into the base.

So how can talk of Solaris being all but dead be anything but FUD?

solx wrote:
> John L wrote:
>> "Jakov Sosic" <jsosic@jsosic.homeunix.org> wrote in message 
>> news:6bmpb6-m93.ln1@pc-jsosic.srce.hr...
>>> Hi!
>>>
>>> I wonder, what do you think, will Solaris really slowly die, and be
>>> replaced with GNU-izied OpenSolaris? As far as I can see from
>>> opensolaris developers and Sun personel, that is the intended way for
>>> the future...
>>>
>>> What do you think about that? I'm sad to see it happening, but with Ian
>>> Murdoch being the chief, it seems inevitable...
>>>
>>>
>>
>> Sun needs urgently to reconsider its enterprise customers, and
>> slow down the pace of change. Ironically, the linux vendors realised
>> this years ago so, for instance, Red Hat has Fedora to trial the cutting
>> edge stuff, and RHEL which changes at a more glacial pace.
>>
>> You might say OpenSolaris leads Solaris but Solaris 10 still has
>> too many changes in my view, both from Solaris 9 and between
>> releases.
>>
>> Sun could also learn from Microsoft, and ensure that before something
>> is released, it is checked for usability and consistency. Quick: what
>> is the name of that SMF service you need to start: is it named after
>> its functionality, its daemon, or something a bit like it?
>>
>> Sun's hardware range too has long been an incoherent mess, and
>> why do they keep mucking around with the names? The successor
>> to the 880 was the 890, which begot, what, the M5000?
>> Whereas HP Proliant 580 went from 580 G3 to 580 G4 to 580 G5.
> 
> The problem for Sun is that it has been forced into the most competitive 
> market having to sell against HP and Dell in the x64 server and 
> workstation arena. Not only that it has an operating system Solaris 
> which can run on its competitor's computers. Sun hardware too expensive 
> then use HP, Dell or local company which uses Gigabyte, ASUS or 
> Supermicro motherboards. Sun needs to generate a revenue stream from 
> support on the x64 platform.
> Sun should have in my opinion been pushing the 64bit capabilities of 
> Solaris. I found it very funny that while running a 64bit desktop for 
> years, it was only last year that my Windows colleagues migrated in XP 
> x64 and even then driver support was very patchy.
> The problem for Sun is that they are heavily reliant on the sales of 
> Sparc based hardware but they are not producing a T series workstation 
> for developers to work on. So developers either have to switch to 
> Solaris x86 or Windows or Linux not a particularly good position for Sun 
> to put its customers in. Some companies will switch to Windows as they 
> see it as an opportunity to get rid of Solaris, some switch to Linux and 
> a few might switch to Solaris/OpenSolaris. Where companies are reliant 
> on 3rd Party libraries paying for Solaris/OpenSolaris x86 support 
> requires effort but they probably already have a Linux port so the 
> decision could made on cost.
> For Sun and Solaris/OpenSolaris to survive, applications need to be 
> easily ported. Having tried to build a few applications, it has been a 
> long annoying process. The situation has definitely improved compared to 
> three years ago but it would be good to be able to download prebuilt 
> libraries that are part of Solaris/OpenSolaris and were selectable.
> I have used sunfreeware, blastwave  and opencsw but if I am building an 
> application I would have to libraries available as part of the core OS.
> It is a pity that the maintainers of these sites could not provide the 
> libraries though the Software Management facility as part of OpenSolaris.
> 
>>
>> It ought to be easy to buy Sun, and it isn't.
> 
> Sun has a lot of resellers which have traditionally provided presales 
> and postsales support.
>>
>> And what is the point of restricting information about bugs, patches
>> and so on? Who gains from Sun system administration getting harder?
>> IBM? HP? Certainly not Sun because if problems are harder to
>> solve, perceived reliability goes down and cost of ownership rises.
> They need to make money from support.
>>
>> And it is very nice that Sun has spent billions of dollars so we can run
>> MySql on Solaris 10 on laptops in our bedrooms but, really, that is not
>> where the money is.
> MySQL was a ill thoughtout purchase unless Sun is actually making money 
> from it.
> 
>>
Wow, Oracle are buying Sun for $7.4 billion. Does this mean 
Solaris/OpenSolaris is going to be dropped as Oracle has a Linux distro 
similar to RedHat or are they going to use Solaris/OpenSolaris and drop 
Linux?
The issue is what Oracle is going to do with the hardware business, I 
suspect that only the software is wanted by Oracle. Oracle could 
possibly sell the SPARC hardware business to Fujitsu and the x64 
business to HP or Dell. Maybe a really mad idea HP buys all the hardware 
business from Oracle.

solx wrote:

> Wow, Oracle are buying Sun for $7.4 billion. Does this mean 
> Solaris/OpenSolaris is going to be dropped as Oracle has a Linux distro 
> similar to RedHat or are they going to use Solaris/OpenSolaris and drop 
> Linux?

I believe just the opposite.  Oracle will now have proprietary hardware 
(yes I know about open sparc for processors, but I haven't seen anything 
mature from it) to go with their proprietary database.  Maybe it will 
encourage more people to adapt to Sparc.

> The issue is what Oracle is going to do with the hardware business, I 
> suspect that only the software is wanted by Oracle. Oracle could 
> possibly sell the SPARC hardware business to Fujitsu and the x64 
> business to HP or Dell. Maybe a really mad idea HP buys all the hardware 
> business from Oracle.
> 

Suns hardware does a good job with Oracle.

Paul Gress wrote:

>  I believe just the opposite.  Oracle will now have proprietary hardware 
>  (yes I know about open sparc for processors, but I haven't seen anything 
>  mature from it) to go with their proprietary database.  Maybe it will 
>  encourage more people to adapt to Sparc.

Mmmm, what's the next Oracle acquisition? AMD? :-)

-- 
 .-.   .-.    Yes, I am an agent of Satan, but my duties are largely
(_  \ /  _)   ceremonial.
     |
     |        dave@fly.srk.fer.hr

On Sat, 18 Apr 2009 21:30:14 +0200, Jakov Sosic <jsosic@jsosic.homeunix.org>
   wrote:
> Hi!
>
> I wonder, what do you think, will Solaris really slowly die, and be
> replaced with GNU-izied OpenSolaris? As far as I can see from
> opensolaris developers and Sun personel, that is the intended way for
> the future...

I'll add my eurocent-worth.

I started using UNIX at the end of the 80s. In our labs at uni, we had a
split between Apollo (later HP) and Sun workstations. Now the days of
the UNIX workstation are almost over. That leaves Solaris with only a
position on servers. This is very staid and conservative. And that road
leads to a slow death IMO.

Much of the early GNU development was done on SunOS. Now, my impression
is that virtually no third party software is developed first on Solaris.
That leaves Solaris in the position of a port OS, if there is enough
demand. There hasn't been much migration from Solaris SPARC to Solaris
x86 - Sun missed that boat, and Linux stepped in. 

I've no idea what Oracle will make of all this.

A bientot
Paul
-- 
Paul Floyd                 http://paulf.free.fr

On 2009-04-20 21:35:34 +0100, Paul Floyd <root@127.0.0.1> said:

> On Sat, 18 Apr 2009 21:30:14 +0200, Jakov Sosic <jsosic@jsosic.homeunix.org>
>    wrote:
>> Hi!
>> 
>> I wonder, what do you think, will Solaris really slowly die, and be
>> replaced with GNU-izied OpenSolaris? As far as I can see from
>> opensolaris developers and Sun personel, that is the intended way for
>> the future...
> 
> I'll add my eurocent-worth.
> 
> I started using UNIX at the end of the 80s. In our labs at uni, we had a
> split between Apollo (later HP) and Sun workstations. Now the days of
> the UNIX workstation are almost over. That leaves Solaris with only a

Only if you equate UNIX with X11 - Macs are pretty popular 
desktop/laptop boxes running UNIX.

-- 
Chris

In <slrngupn9a.t3.root@tryfan.orange.fr>, Paul Floyd wrote:
> Much of the early GNU development was done on SunOS. Now, my impression
> is that virtually no third party software is developed first on Solaris.
> That leaves Solaris in the position of a port OS, if there is enough
> demand. There hasn't been much migration from Solaris SPARC to Solaris
> x86 - Sun missed that boat, and Linux stepped in. 

This really depends what you mean by "third party software".  There are
plenty of companies targeting Solaris for their internal software -- my
employer being one of them -- and at least some of those -- again, my
employer being one -- making the shift from Solaris on SPARC to Solaris
on x64 as appropriate.

We're no Oracle, but we're good for a few thousand machines, and the
same is true of many of those other companies in the same position.

-- 
* Matt McLeod | mail: matt@boggle.org | blog: http://abortrephrase.com/ *
     --- People can do the work, so machines have time to think ---

On Mon, 20 Apr 2009 21:38:41 +0100, Chris Ridd <chrisridd@mac.com> wrote:
> On 2009-04-20 21:35:34 +0100, Paul Floyd <root@127.0.0.1> said:
>
>> On Sat, 18 Apr 2009 21:30:14 +0200, Jakov Sosic <jsosic@jsosic.homeunix.org>
>>    wrote:
>>> Hi!
>>> 
>>> I wonder, what do you think, will Solaris really slowly die, and be
>>> replaced with GNU-izied OpenSolaris? As far as I can see from
>>> opensolaris developers and Sun personel, that is the intended way for
>>> the future...
>> 
>> I'll add my eurocent-worth.
>> 
>> I started using UNIX at the end of the 80s. In our labs at uni, we had a
>> split between Apollo (later HP) and Sun workstations. Now the days of
>> the UNIX workstation are almost over. That leaves Solaris with only a
>
> Only if you equate UNIX with X11 - Macs are pretty popular 
> desktop/laptop boxes running UNIX.

I'd say that the Mac is mainly an Objective C platform, but also capable
of running UNIX and X11 apps. (I'm writing this at a MacBook Pro,
connected to an HP workstation running S10). Also the Mac seems to have
an odd market segment - not business, not server, not HPC. Still, I
quite like it.

A bientot
Paul
-- 
Paul Floyd                 http://paulf.free.fr

In article <slrngupn9a.t3.root@tryfan.orange.fr>,
	Paul Floyd <root@127.0.0.1> writes:
> On Sat, 18 Apr 2009 21:30:14 +0200, Jakov Sosic <jsosic@jsosic.homeunix.org>
>    wrote:
>> Hi!
>>
>> I wonder, what do you think, will Solaris really slowly die, and be
>> replaced with GNU-izied OpenSolaris? As far as I can see from
>> opensolaris developers and Sun personel, that is the intended way for
>> the future...
> 
> I'll add my eurocent-worth.
> 
> I started using UNIX at the end of the 80s. In our labs at uni, we had a
> split between Apollo (later HP) and Sun workstations. Now the days of
> the UNIX workstation are almost over. That leaves Solaris with only a
> position on servers. This is very staid and conservative. And that road
> leads to a slow death IMO.
> 
> Much of the early GNU development was done on SunOS. Now, my impression
> is that virtually no third party software is developed first on Solaris.
> That leaves Solaris in the position of a port OS, if there is enough
> demand. There hasn't been much migration from Solaris SPARC to Solaris
> x86 - Sun missed that boat, and Linux stepped in. 
> 
> I've no idea what Oracle will make of all this.

I will agree to this extent: while really interesting kernel work
(zfs, dtrace, zones, multiple instances of the IP stack, etc; as
contrasted with device drivers for all the cheapest random
commodity hardware that one can dig out of a dumpster)) is most
likely to be done for servers, most of the _mindshare_ that gets
apps onto a system comes with the workstations.  I've been saying
that for some years now, as well as saying that given that SPARC
had a record of delays and of delivering less per-thread performance,
x86 was Sun's obvious insurance against that.

I suspect an at least partial understanding of that is why Sun
bothered to even keep x86 workstations, and to put a lot of effort
into staying more current with GNOME in OpenSolaris, porting lots
of apps, etc.

Given that there's a real need not just for innovation that benefits
the desktop, but for innovation that benefits big SMP servers, clusters,
clouds ,and all sorts of other arrangements, I hope it's not too late...

rlhamil@smart.net (Richard L. Hamilton) writes:

>I suspect an at least partial understanding of that is why Sun
>bothered to even keep x86 workstations, and to put a lot of effort
>into staying more current with GNOME in OpenSolaris, porting lots
>of apps, etc.

It's one of the reasons why we also started to support laptops;
it's easy to install on most x86 hardware so it's easy to play with it.

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.

In article <49eebe1b$0$184$e4fe514c@news.xs4all.nl>,
	Casper H.S. Dik <Casper.Dik@Sun.COM> writes:
> rlhamil@smart.net (Richard L. Hamilton) writes:
> 
>>I suspect an at least partial understanding of that is why Sun
>>bothered to even keep x86 workstations, and to put a lot of effort
>>into staying more current with GNOME in OpenSolaris, porting lots
>>of apps, etc.
> 
> It's one of the reasons why we also started to support laptops;
> it's easy to install on most x86 hardware so it's easy to play with it.
> 
> Casper

Well, there's a point to be made for a distinctive, robust, and
capable workstation which will remain of value over an extended
lifespan.  But there are many laptops already available, and I
imagine that making them is a rather specialized area that Sun
wouldn't find much synergy in making themselves, as long as
[Open]Solaris runs on a reasonable variety of readily available
laptops, for those whose work (or other use) favors laptops, and
who would find an advantage in using a common OS on their laptop
and elsewhere.  Not to mention Solaris developers that might want
to work on the road.  (Now, if one heard less stories of Sun folks
doing presentations on laptops that _weren't_ running Solaris...)

And those who have watched for long presumably know and hopefully
appreciate that you personally have done your part there, with
the early power management prototypes for example (and having
apparently ended up at one point as the new laptop tester
person).  Not to mention the significant and substantial work on
fine-grained permissions, and no doubt quite a few other areas
perhaps less obvious, but also adding value.

Your resume must look _very_ impressive by now.  Hopefully you
won't be needing it anytime soon.  :-)  (much the same could be
said about a number of other names I've come to recognize among your
co-workers)  However things work out, best of luck to all!

On 2009-04-22 18:05:49 +0100, rlhamil@smart.net (Richard L. Hamilton) said:

> In article <49eebe1b$0$184$e4fe514c@news.xs4all.nl>,
> 	Casper H.S. Dik <Casper.Dik@Sun.COM> writes:
>> rlhamil@smart.net (Richard L. Hamilton) writes:
>> 
>>> I suspect an at least partial understanding of that is why Sun
>>> bothered to even keep x86 workstations, and to put a lot of effort
>>> into staying more current with GNOME in OpenSolaris, porting lots
>>> of apps, etc.
>> 
>> It's one of the reasons why we also started to support laptops;
>> it's easy to install on most x86 hardware so it's easy to play with it.
>> 
>> Casper
> 
> Well, there's a point to be made for a distinctive, robust, and
> capable workstation which will remain of value over an extended
> lifespan.  But there are many laptops already available, and I
> imagine that making them is a rather specialized area that Sun
> wouldn't find much synergy in making themselves, as long as

You've seen that Sun resell Toshiba laptops running OpenSolaris?

<http://www.opensolaris.com/toshibanotebook/>

Currently the arrangement is limited to US customers, but apparently 
soon it will include the UK.
-- 
Chris

On Wed, 22 Apr 2009 17:05:49 GMT, rlhamil@smart.net (Richard L. Hamilton) wrote:
referring to Casper H.S. Dik's work on Solaris:
> Your resume must look _very_ impressive by now.  Hopefully you
> won't be needing it anytime soon.  :-)  (much the same could be
> said about a number of other names I've come to recognize among your
> co-workers)  However things work out, best of luck to all!

+1 to that.  I've learned quite a lot of stuff from simply lurking here,
and have benefited greatly both from the posts and the work of Solaris
engineers both here and by using their work :)

In article <75929tF16ro95U1@mid.individual.net>,
Chris Ridd  <chrisridd@mac.com> wrote:
<SNIP!>
>You've seen that Sun resell Toshiba laptops running OpenSolaris?
>
><http://www.opensolaris.com/toshibanotebook/>
>
>Currently the arrangement is limited to US customers, but apparently 
>soon it will include the UK.

These are pretty sweet.  I'm typing this on one of the OpenSolaris configured
Portege R600 outside watching my children play.

-- 
Daniel L. McDonald  -  Solaris Security & Networking Engineering
Mail: danmcd@sun.com             |  * MY OPINIONS ARE NOT NECESSARILY SUN'S! *
35 Network Drive  Burlington, MA |"rising falling at force ten
http://blogs.sun.com/danmcd/     | we twist the world and ride the wind" - Rush

In article <nn7rb6-9l4.ln1@pc-jsosic.srce.hr>,
Jakov Sosic  <jsosic@gmail.com> wrote:
>On 2009-04-19, Paul Gress <pgress@pb.net> wrote:
>
>> No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
>> Solaris 11 will still have much of the Solaris we have today with a lot 
>> of the newer gnome packages.  It should still offer binary compatibility 
>> as in the past.  Opensolaris will not offer binary compatibility.  The 

                                     ^^^   HUH?

Well, isn't 11 supposed to be a version of opensolaris?

Then what -- have to recompile EVERYTHING?

   (I bet some people have lost the sources!)


>> only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
>> the Sparc platform is XSun.  Without this package, I cannot upgrade my 
>> Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
>> graphics card to replace the xvr1200.  As far as I know, there is no 
>> such animal.
>
>So you basically think that SX:CE will be Solaris Next, and not Indiana
>with all the new stuff?
>
>
>
>-- 
>|    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
>=================================================================
>| start fighting cancer -> http://www.worldcommunitygrid.org/   |

In article <49eb5294$0$5916$607ed4bc@cv.net>,
Paul Gress  <pgress@pb.net> wrote:
>Jakov Sosic wrote:
>> On 2009-04-19, Paul Gress <pgress@pb.net> wrote:
>> 
>>> No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
>>> Solaris 11 will still have much of the Solaris we have today with a lot 
>>> of the newer gnome packages.  It should still offer binary compatibility 
>>> as in the past.  Opensolaris will not offer binary compatibility.  The 
>>> only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
>>> the Sparc platform is XSun.  Without this package, I cannot upgrade my 
>>> Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
>>> graphics card to replace the xvr1200.  As far as I know, there is no 
>>> such animal.
>> 
>> So you basically think that SX:CE will be Solaris Next, and not Indiana
>> with all the new stuff?
>> 
>> 
>> 
>
>Yes, I also believe Sun announced this, but I'm not going to search for 
>it.  Solaris 11 will be based off of Opensolaris source, but not the 
>same.  It will also include closed binary for backwards compatibility. 
>They may even include all the gnu utilities, but not in the Path first.
>
>As for my Blade 2500, I'm taking the time to write it off in my mind, 

Write it off?  Are you saying that there's something about it
that won't work with coming solaris versions?

(Reason I ask: I might just get one!)z


>when it's time to replace it most likely it's going to be an X64 based 
>workstation.  Maybe an Ultra 27.  I haven't decided yet.  I'm also 
>looking at the XI Computer Workstations 
>(http://www.xicomputer.com/products/mtowerxeon.asp), they seem to be 
>better equipt then the Ultra 27, capable of much more ram, and dual 5500 
>series processors.  Sun should take a look at them.  Here we have a 
>small workstation company outperforming what Sun offers.
>
>Paul

In article <6MmdnRGchOP7rHbUnZ2dnUVZ8t6dnZ2d@pipex.net>,
solx  <nospam@example.net> wrote:
....


>Having survived the migration from SunOS 3.4 to Solaris 1 (BSD to SysV 
>migration) including the GNU utilities is just another migration process 
>as the OS evolves.

Yeah, great fun.  NOT.


>In respect of the SunBlade 2500, Sun don't make UltraSPARC workstations 
>so sadly you will need to think about replacing it with an x64 machine.
>UltraSPARC appears to be only a server based now.

Please explain a bit.

I've currently got a somewhat kaput blade100, and a friend of mine
thinks a (somewhat used?) blade2500 is what I should get --
$2K + shipping and tax with solaris 10 installed, maybe mirroring
too, etc.

When I get a computer (sun), I like to keep it for YEARS.


What danger, if any, going to a 2500?


Thanks,

David

In article <49eb43dd$0$512$5a6aecb4@news.aaisp.net.uk>,
Andrew Gabriel <andrew@cucumber.demon.co.uk> wrote:
>In article <6MmdnRGchOP7rHbUnZ2dnUVZ8t6dnZ2d@pipex.net>,
>	solx <nospam@example.net> writes:
>> In respect of the SunBlade 2500, Sun don't make UltraSPARC workstations 
>> so sadly you will need to think about replacing it with an x64 machine.
>> UltraSPARC appears to be only a server based now.
>
>There are no more Ultrasparc servers now either (except possibly
>48VDC Netras for the Telco industry). Sparc servers are T series
>(Niagra and decendents) or M series (sparc64 VI and later), in
>addition to the x64 servers and workstations of course.
>
>-- 
>Andrew Gabriel
>[email address is not usable -- followup in the newsgroup]

Anyone know a good place where all these (sparc) chips are laid out
in a table or something, with dates, ete, and pros and cons,
and what OS they might NOT work under?


So I can better follow these discussions.

Thanks.

(Hmmm -- I will see if wikipedia has anything on sun-chips)


David

David Combs wrote:
> In article <nn7rb6-9l4.ln1@pc-jsosic.srce.hr>,
> Jakov Sosic  <jsosic@gmail.com> wrote:
>> On 2009-04-19, Paul Gress <pgress@pb.net> wrote:
>>
>>> No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
>>> Solaris 11 will still have much of the Solaris we have today with a lot 
>>> of the newer gnome packages.  It should still offer binary compatibility 
>>> as in the past.  Opensolaris will not offer binary compatibility.  The 
> 
>                                      ^^^   HUH?
> 
> Well, isn't 11 supposed to be a version of opensolaris?
> 
> Then what -- have to recompile EVERYTHING?
> 
>    (I bet some people have lost the sources!)
> 

Solaris 11 will be sold with the proprietary run time libraries to run 
older Motif, CDE and Openwin apps, Opensolaris will not, but may have a 
separate download for this capability.

> 
>>> only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
>>> the Sparc platform is XSun.  Without this package, I cannot upgrade my 
>>> Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
>>> graphics card to replace the xvr1200.  As far as I know, there is no 
>>> such animal.
>> So you basically think that SX:CE will be Solaris Next, and not Indiana
>> with all the new stuff?
>>
>>
>>
>> -- 
>> |    Jakov Sosic    |    ICQ: 28410271    |   PGP: 0x965CAE2D   |
>> =================================================================
>> | start fighting cancer -> http://www.worldcommunitygrid.org/   |
> 
> 

David Combs wrote:
> In article <49eb5294$0$5916$607ed4bc@cv.net>,
> Paul Gress  <pgress@pb.net> wrote:
>> Jakov Sosic wrote:
>>> On 2009-04-19, Paul Gress <pgress@pb.net> wrote:
>>>
>>>> No I don't see it that way.  It is that way for Opensolaris.  SXCE, or 
>>>> Solaris 11 will still have much of the Solaris we have today with a lot 
>>>> of the newer gnome packages.  It should still offer binary compatibility 
>>>> as in the past.  Opensolaris will not offer binary compatibility.  The 
>>>> only thing Solaris 11 will be, as far as I'm concerned, deficient, for 
>>>> the Sparc platform is XSun.  Without this package, I cannot upgrade my 
>>>> Blade 2500 with xvr1200 graphics.  Unless there is an equavilent 3D open 
>>>> graphics card to replace the xvr1200.  As far as I know, there is no 
>>>> such animal.
>>> So you basically think that SX:CE will be Solaris Next, and not Indiana
>>> with all the new stuff?
>>>
>>>
>>>
>> Yes, I also believe Sun announced this, but I'm not going to search for 
>> it.  Solaris 11 will be based off of Opensolaris source, but not the 
>> same.  It will also include closed binary for backwards compatibility. 
>> They may even include all the gnu utilities, but not in the Path first.
>>
>> As for my Blade 2500, I'm taking the time to write it off in my mind, 
> 
> Write it off?  Are you saying that there's something about it
> that won't work with coming solaris versions?
> 
> (Reason I ask: I might just get one!)z
> 

Yes, I run an application called Pro-engineer, a Mechanical CAD program. 
  It requires OpenGL to display it's 3D graphics properly.  OpenGL 3D 
graphics uses XSun.  My graphics card, XVR600 (looks like I made an 
error, the Blade 2500 cannot use the XVR1200) works with XSun.  Sun is 
planning to EOF XSun, being replaced with Xorg.  The XVR600 and XVR300 
Graphics card has no driver for OpenGL with Xorg.


> 
>> when it's time to replace it most likely it's going to be an X64 based 
>> workstation.  Maybe an Ultra 27.  I haven't decided yet.  I'm also 
>> looking at the XI Computer Workstations 
>> (http://www.xicomputer.com/products/mtowerxeon.asp), they seem to be 
>> better equipt then the Ultra 27, capable of much more ram, and dual 5500 
>> series processors.  Sun should take a look at them.  Here we have a 
>> small workstation company outperforming what Sun offers.
>>
>> Paul
> 
> 

Paul Gress wrote:
> David Combs wrote:
>> In article <49eb5294$0$5916$607ed4bc@cv.net>,
>> Paul Gress  <pgress@pb.net> wrote:
>>> Jakov Sosic wrote:
>>>> On 2009-04-19, Paul Gress <pgress@pb.net> wrote:
>>>>
>>>>> No I don't see it that way.  It is that way for Opensolaris.  SXCE, 
>>>>> or Solaris 11 will still have much of the Solaris we have today 
>>>>> with a lot of the newer gnome packages.  It should still offer 
>>>>> binary compatibility as in the past.  Opensolaris will not offer 
>>>>> binary compatibility.  The only thing Solaris 11 will be, as far as 
>>>>> I'm concerned, deficient, for the Sparc platform is XSun.  Without 
>>>>> this package, I cannot upgrade my Blade 2500 with xvr1200 
>>>>> graphics.  Unless there is an equavilent 3D open graphics card to 
>>>>> replace the xvr1200.  As far as I know, there is no such animal.
>>>> So you basically think that SX:CE will be Solaris Next, and not Indiana
>>>> with all the new stuff?
>>>>
>>>>
>>>>
>>> Yes, I also believe Sun announced this, but I'm not going to search 
>>> for it.  Solaris 11 will be based off of Opensolaris source, but not 
>>> the same.  It will also include closed binary for backwards 
>>> compatibility. They may even include all the gnu utilities, but not 
>>> in the Path first.
>>>
>>> As for my Blade 2500, I'm taking the time to write it off in my mind, 
>>
>> Write it off?  Are you saying that there's something about it
>> that won't work with coming solaris versions?
>>
>> (Reason I ask: I might just get one!)z
>>
> 
> Yes, I run an application called Pro-engineer, a Mechanical CAD program. 
>  It requires OpenGL to display it's 3D graphics properly.  OpenGL 3D 
> graphics uses XSun.  My graphics card, XVR600 (looks like I made an 
> error, the Blade 2500 cannot use the XVR1200) works with XSun.  Sun is 
> planning to EOF XSun, being replaced with Xorg.  The XVR600 and XVR300 
> Graphics card has no driver for OpenGL with Xorg.

ProE is supported on x64 as well a Sparc.  One of my clients uses Dell 
(I think) M6400 laptops as their ProE stations.  How things have changed!

-- 
Ian Collins

Ian Collins wrote:
> 
> ProE is supported on x64 as well a Sparc.  One of my clients uses Dell 
> (I think) M6400 laptops as their ProE stations.  How things have changed!
> 

Yes I know, I already have it running on my Clevo dual core Laptop with 
Nvidia Graphics.  But, in about 3 months I'm going to purchase a new X64 
workstation, mainly because there is no upgrade path for Sparc in a 
workstation.  Sun is now falling behind on X64 workstations.  There are 
no Intel Xeon 5500 series chips available for a workstation.  I may be 
forced to purchase a competitors product.

On 2009-05-23 05:33:29 +0100, Paul Gress <pgress@pb.net> said:

> Ian Collins wrote:
>> 
>> ProE is supported on x64 as well a Sparc.  One of my clients uses Dell 
>> (I think) M6400 laptops as their ProE stations.  How things have 
>> changed!
>> 
> 
> Yes I know, I already have it running on my Clevo dual core Laptop with 
> Nvidia Graphics.  But, in about 3 months I'm going to purchase a new 
> X64 workstation, mainly because there is no upgrade path for Sparc in a 
> workstation.  Sun is now falling behind on X64 workstations.  There are 
> no Intel Xeon 5500 series chips available for a workstation.  I may be 
> forced to purchase a competitors product.

Aren't they server chips, not really meant for workstations?
-- 
Chris

Chris Ridd wrote:
> On 2009-05-23 05:33:29 +0100, Paul Gress <pgress@pb.net> said:
> 
>> Ian Collins wrote:
>>>
>>> ProE is supported on x64 as well a Sparc.  One of my clients uses 
>>> Dell (I think) M6400 laptops as their ProE stations.  How things have 
>>> changed!
>>>
>>
>> Yes I know, I already have it running on my Clevo dual core Laptop 
>> with Nvidia Graphics.  But, in about 3 months I'm going to purchase a 
>> new X64 workstation, mainly because there is no upgrade path for Sparc 
>> in a workstation.  Sun is now falling behind on X64 workstations.  
>> There are no Intel Xeon 5500 series chips available for a 
>> workstation.  I may be forced to purchase a competitors product.
> 
> Aren't they server chips, not really meant for workstations?

They're well enough priced (and power friendly) to use in a decent dual 
processor workstation.

-- 
Ian Collins

Chris Ridd wrote:
> On 2009-05-23 05:33:29 +0100, Paul Gress <pgress@pb.net> said:
> 
>> Ian Collins wrote:
>>>
>>> ProE is supported on x64 as well a Sparc.  One of my clients uses 
>>> Dell (I think) M6400 laptops as their ProE stations.  How things have 
>>> changed!
>>>
>>
>> Yes I know, I already have it running on my Clevo dual core Laptop 
>> with Nvidia Graphics.  But, in about 3 months I'm going to purchase a 
>> new X64 workstation, mainly because there is no upgrade path for Sparc 
>> in a workstation.  Sun is now falling behind on X64 workstations.  
>> There are no Intel Xeon 5500 series chips available for a 
>> workstation.  I may be forced to purchase a competitors product.
> 
> Aren't they server chips, not really meant for workstations?


If you look at one of my previous posts, the Xi� MTower� 2P64X 
Workstation (http://www.xicomputer.com/products/mtowerxeon.asp) is 
awesome.  If I purchase one of those computers, I'll be set for the next 
3 - 5 years.  I can add more memory and a second processor giving me 8 
cores.  The Sun Ultra 27 doesn't come close.  The Sun Ultra 27 maxes out 
at 12 G of memory, which is where I want to start.  fro Xi, they max out 
at 64G - 100G depending which motherboard is chosen.  It basically, in 
my opinion, has breathing room.

Don't get me wrong, I like Sun, I have W2100z, Blade 2500, Ultra 2, 
Blade 100 Workstations.  I'm using SXCE on my Laptop and W2100z 
computers.  I plan to put Opensolaris on my Ultra 2.

I believe the future of Workstations is multiple cores.  Manufactures of 
Games and CAD software are starting to program for multi-cores now.

In article <4a177c99$0$5389$607ed4bc@cv.net>,
	Paul Gress <pgress@pb.net> writes:
> Ian Collins wrote:
>> 
>> ProE is supported on x64 as well a Sparc.  One of my clients uses Dell 
>> (I think) M6400 laptops as their ProE stations.  How things have changed!
> 
> Yes I know, I already have it running on my Clevo dual core Laptop with 
> Nvidia Graphics.  But, in about 3 months I'm going to purchase a new X64 
> workstation, mainly because there is no upgrade path for Sparc in a 
> workstation.  Sun is now falling behind on X64 workstations.  There are 
> no Intel Xeon 5500 series chips available for a workstation.  I may be 
> forced to purchase a competitors product.

What are you looking for in the 5500 chips, verses 3500 (e.g. Ultra 27)?

Might want to have a look at:
http://blogs.sun.com/jnerl/entry/configuring_and_optimizing_intel_xeon

-- 
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

In article <4a177c99$0$5389$607ed4bc@cv.net>,
Paul Gress  <pgress@pb.net> wrote:
>Ian Collins wrote:
>> 
>> ProE is supported on x64 as well a Sparc.  One of my clients uses Dell 
>> (I think) M6400 laptops as their ProE stations.  How things have changed!
>> 
>
>Yes I know, I already have it running on my Clevo dual core Laptop with 
>Nvidia Graphics.  But, in about 3 months I'm going to purchase a new X64 
>workstation, mainly because there is no upgrade path for Sparc in a 
>workstation.  Sun is now falling behind on X64 workstations.  There are 
>no Intel Xeon 5500 series chips available for a workstation.  I may be 
>forced to purchase a competitors product.

My suggestion:  COMPLAIN   (about no more sparc-workstations).

Complain to EVERYONE: ceo, pres, etc of Sun, AND of Oracle.

Tell 'em you're just one of lots of people (it that's true)
who will be FORCED onto NON-sun equipment by this "no upgrade path"
for sparc workstations.

MY problem with it: you can NO LONGER get sun SERVICE on 
an INEXPENSIVE sparc-based computer.

WHY NOT?

Because of Sun's (STUPID) "end of life" rule.

So, you go look at SUN's remanufactured page, see something
you like, nice price, etc, and ASSUME that you'll be able
to get FIVE years of service contracts on it.

NOPE!  Sun tells me that the "end of life" STILL APPLIES.

That is, you buy a remanufactured-by-Sun computer (buy it from sun
itself), and discover that you can purchase a service contract
for NO LONGER THAN TWO MORE YEARS, and after that, it's "you're
on your own, buddy!".

So if you're not yourself a hardware hacker, you might as well
throw it out after those two years have passed.  :-(

Not real good, huh?


David

In article <4a183741$0$5383$607ed4bc@cv.net>,
Paul Gress  <pgress@pb.net> wrote:
....
>
>Don't get me wrong, I like Sun, I have W2100z, Blade 2500, Ultra 2, 
>Blade 100 Workstations.  I'm using SXCE on my Laptop and W2100z 
>computers.  I plan to put Opensolaris on my Ultra 2.
....

How to you get the sun's maintained?

Here in the NYC area, I've had a hard time finding someone
who services Sun computers willing to service mine.

So, how do YOU do it?


David

David Combs wrote:
> In article <4a177c99$0$5389$607ed4bc@cv.net>,
> Paul Gress  <pgress@pb.net> wrote:
>> Ian Collins wrote:
>>> ProE is supported on x64 as well a Sparc.  One of my clients uses Dell 
>>> (I think) M6400 laptops as their ProE stations.  How things have changed!
>>>
>> Yes I know, I already have it running on my Clevo dual core Laptop with 
>> Nvidia Graphics.  But, in about 3 months I'm going to purchase a new X64 
>> workstation, mainly because there is no upgrade path for Sparc in a 
>> workstation.  Sun is now falling behind on X64 workstations.  There are 
>> no Intel Xeon 5500 series chips available for a workstation.  I may be 
>> forced to purchase a competitors product.
> 
> My suggestion:  COMPLAIN   (about no more sparc-workstations).
> 
> Complain to EVERYONE: ceo, pres, etc of Sun, AND of Oracle.
> 
> Tell 'em you're just one of lots of people (it that's true)
> who will be FORCED onto NON-sun equipment by this "no upgrade path"
> for sparc workstations.
> 
> MY problem with it: you can NO LONGER get sun SERVICE on 
> an INEXPENSIVE sparc-based computer.
> 
> WHY NOT?
> 
> Because of Sun's (STUPID) "end of life" rule.
> 
> So, you go look at SUN's remanufactured page, see something
> you like, nice price, etc, and ASSUME that you'll be able
> to get FIVE years of service contracts on it.
> 
> NOPE!  Sun tells me that the "end of life" STILL APPLIES.
> 
> That is, you buy a remanufactured-by-Sun computer (buy it from sun
> itself), and discover that you can purchase a service contract
> for NO LONGER THAN TWO MORE YEARS, and after that, it's "you're
> on your own, buddy!".
> 
> So if you're not yourself a hardware hacker, you might as well
> throw it out after those two years have passed.  :-(
> 
> Not real good, huh?
> 

It's not all bad!  Most failures are things like disk drives, tape 
drives and fans; e.g. moving parts.  Replacing disk drives is hardly 
rocket science.  Fans aren't rocket science either!

   Tape drives do present a problem!  They need maintenance by people 
who know how; I don't think there's any substitute.

I've known people to buy obsolete hardware, stack it in the warehouse, 
and use it to replace failed hardware.

On May 27, 4:44 pm, dkco...@panix.com (David Combs) wrote:

> MY problem with it: you can NO LONGER get sun SERVICE on
> an INEXPENSIVE sparc-based computer.
>
> WHY NOT?
>
> Because of Sun's (STUPID) "end of life" rule.


It's not stupid from their perspective.

They want to "encourage" you to upgrade, where you will
not only shell out more money for a new service contract,
but for new hardware as well.

If they didn't EOL this stuff, some people would *never*
buy new equipment!

:-)

ThanksButNo wrote:
> On May 27, 4:44 pm, dkco...@panix.com (David Combs) wrote:
> 
>> MY problem with it: you can NO LONGER get sun SERVICE on
>> an INEXPENSIVE sparc-based computer.
>>
>> WHY NOT?
>>
>> Because of Sun's (STUPID) "end of life" rule.
> 
> 
> It's not stupid from their perspective.
> 
> They want to "encourage" you to upgrade, where you will
> not only shell out more money for a new service contract,
> but for new hardware as well.
> 
> If they didn't EOL this stuff, some people would *never*
> buy new equipment!
> 
> :-)

Were talking about Workstations, where there is no upgrade path for 
Sparc and on the X86 line I complained Sun is not keeping up with 
smaller competitors.  I want to upgrade my Sun Blade 2500, I have to go 
to an X86 line, I looked at the Ultra 27, it can only have a maximum 
memory of 12 gig ECC Ram.  My W2100z could at least be equipt with 32 
gig of ECC Ram.  So my complaint is on Suns current line of X86 
Workstations, I may have to go to a competitor that offers much more 
configurable Workstations ranging from approx 64 gig to 100 gig of ECC 
Ram and multiple Intel Xeon 5500 series chips.

I'm just disappointed with Sun for not offering state of the art 
workstations, where much smaller companies can.

ThanksButNo wrote:
> On May 27, 4:44 pm, dkco...@panix.com (David Combs) wrote:
> 
>> MY problem with it: you can NO LONGER get sun SERVICE on
>> an INEXPENSIVE sparc-based computer.
>>
>> WHY NOT?
>>
>> Because of Sun's (STUPID) "end of life" rule.
> 
> 
> It's not stupid from their perspective.
> 
> They want to "encourage" you to upgrade, where you will
> not only shell out more money for a new service contract,
> but for new hardware as well.
> 
> If they didn't EOL this stuff, some people would *never*
> buy new equipment!
> 
> :-)

Were talking about Workstations, where there is no upgrade path for 
Sparc and on the X86 line I complained Sun is not keeping up with 
smaller competitors.  I want to upgrade my Sun Blade 2500, I have to go 
to an X86 line, I looked at the Ultra 27, it can only have a maximum 
memory of 12 gig ECC Ram.  My W2100z could at least be equipt with 32 
gig of ECC Ram.  So my complaint is on Suns current line of X86 
Workstations, I may have to go to a competitor that offers much more 
configurable Workstations ranging from approx 64 gig to 100 gig of ECC 
Ram and multiple Intel Xeon 5500 series chips.

I'm just disappointed with Sun for not offering state of the art 
workstations, where much smaller companies can.

Paul Gress wrote:
> ThanksButNo wrote:
>> On May 27, 4:44 pm, dkco...@panix.com (David Combs) wrote:
>>
>>> MY problem with it: you can NO LONGER get sun SERVICE on
>>> an INEXPENSIVE sparc-based computer.
>>>
>>> WHY NOT?
>>>
>>> Because of Sun's (STUPID) "end of life" rule.
>>
>>
>> It's not stupid from their perspective.
>>
>> They want to "encourage" you to upgrade, where you will
>> not only shell out more money for a new service contract,
>> but for new hardware as well.
>>
>> If they didn't EOL this stuff, some people would *never*
>> buy new equipment!
>>
>> :-)
> 
> Were talking about Workstations, where there is no upgrade path for 
> Sparc and on the X86 line I complained Sun is not keeping up with 
> smaller competitors.  I want to upgrade my Sun Blade 2500, I have to go 
> to an X86 line, I looked at the Ultra 27, it can only have a maximum 
> memory of 12 gig ECC Ram.  My W2100z could at least be equipt with 32 
> gig of ECC Ram.  So my complaint is on Suns current line of X86 
> Workstations, I may have to go to a competitor that offers much more 
> configurable Workstations ranging from approx 64 gig to 100 gig of ECC 
> Ram and multiple Intel Xeon 5500 series chips.
> 
> I'm just disappointed with Sun for not offering state of the art 
> workstations, where much smaller companies can.

What you are specifying looks a lot more like a mainframe than a 
workstation!  I've spent forty years in IT and have NEVER seen or heard 
of anything with that much RAM.

Richard B. Gilbert wrote:

> What you are specifying looks a lot more like a mainframe than a 
> workstation!  I've spent forty years in IT and have NEVER seen or heard 
> of anything with that much RAM.
> 

The link below is for the configuration page.  It shows it's available 
with up to 50 gig of ECC Ram.  But if you go almost all the way to the 
bottom and look at Motherboard selections, a google on those 
Motherboards are configurable to up to 100 gig of ECC Ram.

Xi� MTower� 2P64X Workstation
http://www.xicomputer.com/products/Configure_prof.asp?model=mtowerxeon&configid=


Paul

Paul Gress wrote:
> Richard B. Gilbert wrote:
> 
>> What you are specifying looks a lot more like a mainframe than a 
>> workstation!  I've spent forty years in IT and have NEVER seen or 
>> heard of anything with that much RAM.
>>
> 
> The link below is for the configuration page.  It shows it's available 
> with up to 50 gig of ECC Ram.  But if you go almost all the way to the 
> bottom and look at Motherboard selections, a google on those 
> Motherboards are configurable to up to 100 gig of ECC Ram.

If you can find 8GB DIMs!

-- 
Ian Collins

On May 29, 4:29 pm, Ian Collins <ian-n...@hotmail.com> wrote:
> Paul Gress wrote:
> > Richard B. Gilbert wrote:
>
> >> What you are specifying looks a lot more like a mainframe than a
> >> workstation!  I've spent forty years in IT and have NEVER seen or
> >> heard of anything with that much RAM.
>
> > The link below is for the configuration page.  It shows it's available
> > with up to 50 gig of ECC Ram.  But if you go almost all the way to the
> > bottom and look at Motherboard selections, a google on those
> > Motherboards are configurable to up to 100 gig of ECC Ram.
>
> If you can find 8GB DIMs!

These days, there is very little that would surprise me!

I can imagine a 128G solid state "ram" that plugs into
a USB 4.0 port and runs circles around anything we have
today.  And high school kids walking around everywhere
with 'em hanging off their key-chains, and losing 'em,
and mom says, "That's Ok, we'll get some more next time
we stop at the 99 cent store."  No worries about the
data, thanks to pervasive broadband wi-fi auto-backup
systems connected on every cell-phone tower.

And we'll be codgers whining about, "I remember when all
we had were 8 inch floppies, and ... zzzZZ ZZzzz ..."

<8-(

Ian Collins <ian-news@hotmail.com> writes:
>Paul Gress wrote:
>> Richard B. Gilbert wrote:
>> 
>>> What you are specifying looks a lot more like a mainframe than a 
>>> workstation!  I've spent forty years in IT and have NEVER seen or 
>>> heard of anything with that much RAM.
>>>
>> 
>> The link below is for the configuration page.  It shows it's available 
>> with up to 50 gig of ECC Ram.  But if you go almost all the way to the 
>> bottom and look at Motherboard selections, a google on those 
>> Motherboards are configurable to up to 100 gig of ECC Ram.

>If you can find 8GB DIMs!

Finding them is easy. 

Affording them is something else.  ($900 a stick). 

On Fri, 29 May 2009 18:46:26 -0700 (PDT), ThanksButNo
   <no.no.thanks@gmail.com> wrote:
>
> These days, there is very little that would surprise me!
>
> I can imagine a 128G solid state "ram" that plugs into
> a USB 4.0 port and runs circles around anything we have
> today.  And high school kids walking around everywhere
> with 'em hanging off their key-chains, and losing 'em,
> and mom says, "That's Ok, we'll get some more next time
> we stop at the 99 cent store."  No worries about the
> data, thanks to pervasive broadband wi-fi auto-backup
> systems connected on every cell-phone tower.
>
> And we'll be codgers whining about, "I remember when all
> we had were 8 inch floppies, and ... zzzZZ ZZzzz ..."

DRAM and Flash densities are definitily running out of steam. Both use
charge stored in a capacitor, and as the number of atoms (and electrons)
starts to become quite small, then they don't work too well. That means
not much scope for more 2D shrinking. The next think might be to add
more poly layers (so the devices become 3D), subject to heat dissipation
issues. And then there are the possibilities of completely new
architectures - optical, quantum, magnetic or whatever.

In any case, I can't see the same factor happening again between the 1st
PC RAM that I bought (2Mbyte sticks) and the last (1Gbyte sticks) with
DRAM.

A+
Paul
-- 
Paul Floyd                 http://paulf.free.fr

* Andrew Gabriel:

> What are you looking for in the 5500 chips, verses 3500 (e.g. Ultra 27)?

The XEON 3500 is single processor only (its a relabelled Core i7), the 
XEON 5500 are dual processor capable.

Benjamin

* Chris Ridd:

> Aren't they server chips, not really meant for workstations?

Nope. Like all XEON 5000 series processors they were made for dual 
processor servers and workstations.

Benjamin

* Paul Gress:

> If you look at one of my previous posts, the Xi� MTower� 2P64X 
> Workstation (http://www.xicomputer.com/products/mtowerxeon.asp) is 
> awesome.  

Well, its a standard off-the-shelf system made from generic parts (XI 
computer is no manufacturer, it is just an assembler). All they do is 
buying standard parts and putting in in a gamer PC case. This is hardly 
comparable to a Sun x64 workstation which has components made to Suns 
specifications, runs through a lot of tests and is certified for a lot 
of professional applications. Assemblers usually don't even do proper 
EMC tests. And since most standard components have a very short product 
life if for example the system board fails after say one year you can't 
be sure to get an identical one any more. I would never buy a 
workstation from an assembler, only from one of the big manufacturers 
(Dell, HP, Sun, Lenovo, FSC).

My next workstation will be a HP z800:
<http://h10010.www1.hp.com/wwpc/au/en/sm/WF05a/12454-12454-296719-296721-296721-3718645.html>

Two XEON 5500 and up to 192GB RAM. Sadly it's still not available in Europe.

Benjamin

* Paul Gress:

> Were talking about Workstations, where there is no upgrade path for 
> Sparc and on the X86 line I complained Sun is not keeping up with 
> smaller competitors.  

Sun for a long time didn't keep up even with the big competitors in the 
workstation market. For dual processor workstations, all Sun had was the 
Ultra 40 (1st gen Opteron Socket940) and later the Ultra 40 M2 (Opteron 
Socket F), while the competitor offered XEON 5100/5200 (dual core) and 
XEON 5300/5400 (quadcore) workstations which in most applications blew 
any dual Opteron machine away. Sun offered these XEONs only in servers 
but not in workstations.

> I'm just disappointed with Sun for not offering state of the art 
> workstations, where much smaller companies can.

Sun probably doesn't see much money in the workstation market any more. 
Especially since with x64 the competition is strong.

Benjamin

* Richard B. Gilbert:

> What you are specifying looks a lot more like a mainframe than a 
> workstation!  I've spent forty years in IT and have NEVER seen or heard 
> of anything with that much RAM.

In certain areas these amounts of memory are quite common. For example, 
my old xw8400 workstation I bought 2007 already had 32GB RAM, and as 
soon as DDR3 registered memory falls into a somewhat sane price range 
I'll probably go for 128GB when I get my next workstation.

Benjamin

On Sun, 31 May 2009, Benjamin Gawert wrote:

> In certain areas these amounts of memory are quite common. For example, my old
> xw8400 workstation I bought 2007 already had 32GB RAM, and as soon as DDR3
> registered memory falls into a somewhat sane price range I'll probably go for
> 128GB when I get my next workstation.

Blimey!  What on earth do you do that is so memory intensive?!

-- 
Rich Teer, SCSA, SCNA, SCSECA

URLs: http://www.rite-group.com/rich
      http://www.linkedin.com/in/richteer

On May 31, 9:23 am, Rich Teer <rich.t...@rite-group.com> wrote:
> On Sun, 31 May 2009, Benjamin Gawert wrote:
> > In certain areas these amounts of memory are quite common. For example, my old
> > xw8400 workstation I bought 2007 already had 32GB RAM, and as soon as DDR3
> > registered memory falls into a somewhat sane price range I'll probably go for
> > 128GB when I get my next workstation.
>
> Blimey!  What on earth do you do that is so memory intensive?!
>

Put your entire database in RAM -- goodbye disk latency!

8-D

In article <bee61d22-df2b-4054-9e08-773ab0b3654f@f16g2000vbf.googlegroups.com>,
	ThanksButNo <no.no.thanks@gmail.com> writes:
> On May 31, 9:23 am, Rich Teer <rich.t...@rite-group.com> wrote:
>> On Sun, 31 May 2009, Benjamin Gawert wrote:
>> > In certain areas these amounts of memory are quite common. For example, my old
>> > xw8400 workstation I bought 2007 already had 32GB RAM, and as soon as DDR3
>> > registered memory falls into a somewhat sane price range I'll probably go for
>> > 128GB when I get my next workstation.
>>
>> Blimey!  What on earth do you do that is so memory intensive?!
> 
> Put your entire database in RAM -- goodbye disk latency!

On a server, sure (but consider Enterprise SSD's instead).
But on a workstation?

-- 
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

Rich Teer wrote:
> On Sun, 31 May 2009, Benjamin Gawert wrote:
> 
>> In certain areas these amounts of memory are quite common. For example, my old
>> xw8400 workstation I bought 2007 already had 32GB RAM, and as soon as DDR3
>> registered memory falls into a somewhat sane price range I'll probably go for
>> 128GB when I get my next workstation.
> 
> Blimey!  What on earth do you do that is so memory intensive?!

Hello world in Java?

-- 
Ian Collins

Andrew Gabriel wrote:
> In article <bee61d22-df2b-4054-9e08-773ab0b3654f@f16g2000vbf.googlegroups.com>,
> 	ThanksButNo <no.no.thanks@gmail.com> writes:
>> On May 31, 9:23 am, Rich Teer <rich.t...@rite-group.com> wrote:
>>> On Sun, 31 May 2009, Benjamin Gawert wrote:
>>>> In certain areas these amounts of memory are quite common. For example, my old
>>>> xw8400 workstation I bought 2007 already had 32GB RAM, and as soon as DDR3
>>>> registered memory falls into a somewhat sane price range I'll probably go for
>>>> 128GB when I get my next workstation.
>>> Blimey!  What on earth do you do that is so memory intensive?!
>> Put your entire database in RAM -- goodbye disk latency!
> 
> On a server, sure (but consider Enterprise SSD's instead).
> But on a workstation?
> 

I do CAD, pro-engineer specifically.  Because I'm an engineer, I've been 
known to have large CAD assemblies, multiple models open, which consumes 
large amounts of memory.  To make things worse, I often use Gimp to edit 
images at the same time, basically photo rendering my models and 
assemblies.  To say the least, I quite often start paging out to the 
hard disk with 8 gig of ram installed.  So yes, I think 12 gig is just 
the starting point for workstations today.  This is where Sun peaks out at.

Paul

Benjamin Gawert wrote:
> * Paul Gress:
> 
>> If you look at one of my previous posts, the Xi� MTower� 2P64X 
>> Workstation (http://www.xicomputer.com/products/mtowerxeon.asp) is 
>> awesome.  
> 
> Well, its a standard off-the-shelf system made from generic parts (XI 
> computer is no manufacturer, it is just an assembler). All they do is 
> buying standard parts and putting in in a gamer PC case. This is hardly 
> comparable to a Sun x64 workstation which has components made to Suns 
> specifications, runs through a lot of tests and is certified for a lot 
> of professional applications. Assemblers usually don't even do proper 
> EMC tests. And since most standard components have a very short product 
> life if for example the system board fails after say one year you can't 
> be sure to get an identical one any more. I would never buy a 
> workstation from an assembler, only from one of the big manufacturers 
> (Dell, HP, Sun, Lenovo, FSC).
> 
> My next workstation will be a HP z800:
> <http://h10010.www1.hp.com/wwpc/au/en/sm/WF05a/12454-12454-296719-296721-296721-3718645.html> 
> 
> 
> Two XEON 5500 and up to 192GB RAM. Sadly it's still not available in 
> Europe.
> 
> Benjamin


I looked at that workstation and it's awesome.  I agree with what you 
say about small assembliers.  I didn't know HP had such an offering, thanks.


Paul

Paul Gress wrote:
> Andrew Gabriel wrote:
>> In article 
>> <bee61d22-df2b-4054-9e08-773ab0b3654f@f16g2000vbf.googlegroups.com>,
>>     ThanksButNo <no.no.thanks@gmail.com> writes:
>>> On May 31, 9:23 am, Rich Teer <rich.t...@rite-group.com> wrote:
>>>> On Sun, 31 May 2009, Benjamin Gawert wrote:
>>>>> In certain areas these amounts of memory are quite common. For 
>>>>> example, my old
>>>>> xw8400 workstation I bought 2007 already had 32GB RAM, and as soon 
>>>>> as DDR3
>>>>> registered memory falls into a somewhat sane price range I'll 
>>>>> probably go for
>>>>> 128GB when I get my next workstation.
>>>> Blimey!  What on earth do you do that is so memory intensive?!
>>> Put your entire database in RAM -- goodbye disk latency!
>>
>> On a server, sure (but consider Enterprise SSD's instead).
>> But on a workstation?
>>
> 
> I do CAD, pro-engineer specifically.  Because I'm an engineer, I've been 
> known to have large CAD assemblies, multiple models open, which consumes 
> large amounts of memory.  To make things worse, I often use Gimp to edit 
> images at the same time, basically photo rendering my models and 
> assemblies.  To say the least, I quite often start paging out to the 
> hard disk with 8 gig of ram installed.  So yes, I think 12 gig is just 
> the starting point for workstations today.  This is where Sun peaks out at.
> 
> Paul

Can Sun really afford to launch a family of workstations that support, 
say, 96 or 128GB of RAM?  I suspect it's a niche market, meaning low 
volume.  That, in turn, means expensive!

Richard B. Gilbert wrote:

> 
> Can Sun really afford to launch a family of workstations that support, 
> say, 96 or 128GB of RAM?  I suspect it's a niche market, meaning low 
> volume.  That, in turn, means expensive!

Sun use to have awesome workstations.  I remember when I fist got my 
Ultra 2.  Yes it was only a dual 400 mhz Sparc, but was faster and had 
better graphics then a PC.  I've since then purchased a Blade 100 and a 
Blade 2500, which was slower then PC's, but I was sold on Sparc 
stability.  I purchased a W2100z, this has 8 gig of ram which can be 
loaded to 32 gig, still whaqt I would call a workstation.  Sun is now I 
would say producing power user computers, not necessarily workstations 
any more for X86/64.

If I required servers, I'd be having a different discussion.

So now it looks as if Sun is going out of the workstation business 
entirely.  I hope this doesn't affect my CAD application which currently 
runs on Solaris x64 (Solaris 10 and SXCE NV114).

Paul

Richard B. Gilbert wrote:
> 
> Can Sun really afford to launch a family of workstations that support, 
> say, 96 or 128GB of RAM?  I suspect it's a niche market, meaning low 
> volume.  That, in turn, means expensive!

Look at the prices of the Supermicro Xeon 5500 boards, you'll be 
pleasantly surprised!

-- 
Ian Collins

In article <4a23470f$0$5912$607ed4bc@cv.net>,
	Paul Gress <pgress@pb.net> writes:
> 
> So now it looks as if Sun is going out of the workstation business 
> entirely.  I hope this doesn't affect my CAD application which currently 
> runs on Solaris x64 (Solaris 10 and SXCE NV114).

Out of interest, what is the CAD application?

-- 
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

On Mon, 01 Jun 2009 19:27:45 +1200
Ian Collins <ian-news@hotmail.com> wrote:

> Richard B. Gilbert wrote:
> > 
> > Can Sun really afford to launch a family of workstations that
> > support, say, 96 or 128GB of RAM?  I suspect it's a niche market,
> > meaning low volume.  That, in turn, means expensive!
> 
> Look at the prices of the Supermicro Xeon 5500 boards, you'll be 
> pleasantly surprised!

That is the point - what value can Sun add to Xeon 5500 processors,
Supermicro boards, RAM chips and SATA disks? A different case design?

The only thing that Sun could try and do is sell more of them, so that
they can pocket the discounts. Sun no longer sells "workstations" for
the same reasons HP no longer sells them, the main one being that
"real" workstations no longer exist.

All the hardware savvy that made the difference between a workstation
and a PC has now been integrated in the CPUs and support ICs. Sun cannot
design a substantially faster processor than Intel or AMD. Sun cannot
design a better motherboard than the likes of Supermicro, Tyan,
Gigabyte or Asus. Sun cannot design a better case than Antec, etc. 

In this decade, Sun did design a totally different processor - something
that cannot be said from HP, Dell or other even Intel. Sun added
substantial innovations to their OS - something that cannot be said from
Microsoft (they tried and failed), HP or IBM (who didn't even try). 

Sun didn't succeed, but you cannot fault them for not trying. After
all, would we be having this discussion if Sun had become like HP, a
technologically largely empty marketing organisation?

-- 
Stefaan A Eeckels
-- 
"Object-oriented programming is an exceptionally bad idea which
 could only have originated in California." --Edsger Dijkstra

Andrew Gabriel wrote:
> In article <4a23470f$0$5912$607ed4bc@cv.net>,
> 	Paul Gress <pgress@pb.net> writes:
>> So now it looks as if Sun is going out of the workstation business 
>> entirely.  I hope this doesn't affect my CAD application which currently 
>> runs on Solaris x64 (Solaris 10 and SXCE NV114).
> 
> Out of interest, what is the CAD application?
> 

Computer Aided Design

Basically my CAD program is for Mechanical Design.  The way you design 
on computer today is to create what is called a solid model, which is a 
3D representation of the product your designing.  You then create all 
your drawings after the model is finished.

Paul

On Mon, 1 Jun 2009, Andrew Gabriel wrote:

> > So now it looks as if Sun is going out of the workstation business 
> > entirely.  I hope this doesn't affect my CAD application which currently 
> > runs on Solaris x64 (Solaris 10 and SXCE NV114).
> 
> Out of interest, what is the CAD application?

Earlier in this thread, Paul mentioned Pro-Engineer.  I assume that's the
CAD software he speaks of.

-- 
Rich Teer, SCSA, SCNA, SCSECA

URLs: http://www.rite-group.com/rich
      http://www.linkedin.com/in/richteer

In article <4a23f38b$0$25617$607ed4bc@cv.net>,
	Paul Gress <pgress@pb.net> writes:
> Andrew Gabriel wrote:
>> In article <4a23470f$0$5912$607ed4bc@cv.net>,
>> 	Paul Gress <pgress@pb.net> writes:
>>> So now it looks as if Sun is going out of the workstation business 
>>> entirely.  I hope this doesn't affect my CAD application which currently 
>>> runs on Solaris x64 (Solaris 10 and SXCE NV114).
>> 
>> Out of interest, what is the CAD application?
> 
> Computer Aided Design
> 
> Basically my CAD program is for Mechanical Design.  The way you design 
> on computer today is to create what is called a solid model, which is a 
> 3D representation of the product your designing.  You then create all 
> your drawings after the model is finished.

I meant, what's the product name?

Does it directly use OpenGL 3D graphics support?

-- 
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

Andrew Gabriel wrote:
> In article <4a23f38b$0$25617$607ed4bc@cv.net>,
> 	Paul Gress <pgress@pb.net> writes:
>> Andrew Gabriel wrote:
>>> In article <4a23470f$0$5912$607ed4bc@cv.net>,
>>> 	Paul Gress <pgress@pb.net> writes:
>>>> So now it looks as if Sun is going out of the workstation business 
>>>> entirely.  I hope this doesn't affect my CAD application which currently 
>>>> runs on Solaris x64 (Solaris 10 and SXCE NV114).
>>> Out of interest, what is the CAD application?
>> Computer Aided Design
>>
>> Basically my CAD program is for Mechanical Design.  The way you design 
>> on computer today is to create what is called a solid model, which is a 
>> 3D representation of the product your designing.  You then create all 
>> your drawings after the model is finished.
> 
> I meant, what's the product name?
> 
> Does it directly use OpenGL 3D graphics support?
> 

My CAD product name is Pro-Engineer Wildfire 4.

Yes, without OpenGL it would be painfully slow.  The next Graphic card I 
plan to use is the Nvidia Quadro FX5800.  I only choose my computers now 
to olny use Nvidia, my laptop has the Quadro FX1600m, another fast 3D 
mobile graphics card.  When you do build 3D models you constantly spin 
it around in 3D.  It's basically the same 3D intensive Graphics as 
Gaming.  The Nvidia cards designed for gaming are the GeForce cards and 
for CAD are the FX cards.  I could build an assembly (many 3D models 
attached together) with say 300 parts, the CAD software keeps it all in 
RAM.  This is why having lots of RAM is important for CAD work.

Paul

Stefaan A Eeckels wrote:

> 
> That is the point - what value can Sun add to Xeon 5500 processors,
> Supermicro boards, RAM chips and SATA disks? A different case design?
> 
> The only thing that Sun could try and do is sell more of them, so that
> they can pocket the discounts. Sun no longer sells "workstations" for
> the same reasons HP no longer sells them, the main one being that
> "real" workstations no longer exist.
> 
> All the hardware savvy that made the difference between a workstation
> and a PC has now been integrated in the CPUs and support ICs. Sun cannot
> design a substantially faster processor than Intel or AMD. Sun cannot
> design a better motherboard than the likes of Supermicro, Tyan,
> Gigabyte or Asus. Sun cannot design a better case than Antec, etc. 
> 
> In this decade, Sun did design a totally different processor - something
> that cannot be said from HP, Dell or other even Intel. Sun added
> substantial innovations to their OS - something that cannot be said from
> Microsoft (they tried and failed), HP or IBM (who didn't even try). 
> 
> Sun didn't succeed, but you cannot fault them for not trying. After
> all, would we be having this discussion if Sun had become like HP, a
> technologically largely empty marketing organisation?
> 

You make some very good points.  What company makes workstations any 
more today, and then if you think you found one, what differentiates it 
from other workstation manufactures.

OK, some of my thoughts.

I looked at the Sun X2270 server, it has all the makings for a good 
workstation, only deficient in PCI express slots (only has one) and no 
provisions for an internal CD/DVD drive.  Basically I could use this if 
it were repackaged in another case with a CD/DVD drive (to use the USB). 
  I basically don't need to use the PCI expansion slots as long as it 
has USB.  USB is easy to expand.

So, Sun has the workings/beginnings for a good workstation, they just 
don't do it.

I might try to do it myself, but to go all through the design work for 
just one isn't worth it.  I couldn't do multiple ones to sell as I 
believe you need to be a certified reseller or VAR.


Paul

* Rich Teer:

> Blimey!  What on earth do you do that is so memory intensive?!

I did (and still do) certain simulations that use big data sets, and 
when I bought my xw8400 32GB already imposed to be a limit (but at that 
time only 4GB FBDIMMs were available so more wasn't possible).

I have also seen professional video processing suites for HD video that 
use 32GB RAM.

Benjamin

* Richard B. Gilbert:

> Can Sun really afford to launch a family of workstations that support, 
> say, 96 or 128GB of RAM? 

Yes, they can, as every other big player like HP or Dell can do that for 
many years now. Especially since memory is so cheap.

> I suspect it's a niche market, meaning low 
> volume.  That, in turn, means expensive!

Well, as to a low volume market and expensive systems, if you look at 
all the other offerings that are there you will see that this isn't the 
case.

Benjamin

* Paul Gress:

> So now it looks as if Sun is going out of the workstation business 
> entirely. 

They probably are for some time now.

> I hope this doesn't affect my CAD application which currently 
> runs on Solaris x64 (Solaris 10 and SXCE NV114).

It will, maybe not necessarily in the long term, but Pro/E like other 
similar programs will at some point just drop Solaris support.

Benjamin

* Paul Gress:

> My CAD product name is Pro-Engineer Wildfire 4.
> 
> Yes, without OpenGL it would be painfully slow.  The next Graphic card I 
> plan to use is the Nvidia Quadro FX5800. 

A Quadro FX 5800 for Pro/E is wasted money. I had a Quadro FX 5600, but 
even that for Pro/E it was wasted. If all you need is Pro/E, a cheaper 
midrange card is a much more sensible choice.

Benjamin

* Stefaan A Eeckels:

> That is the point - what value can Sun add to Xeon 5500 processors,
> Supermicro boards, RAM chips and SATA disks? A different case design?

Sun can add the same value as HP or Dell or Lenovo can offer, and which 
is the reason that in every serious workstation application no-one even 
touches a system made from generic off-the-shelf parts: offering a 
complete system which is extensively tested (something which assemblers 
don't do), comes with appropriate service options and is certified for 
most professional applications.

> The only thing that Sun could try and do is sell more of them, so that
> they can pocket the discounts. Sun no longer sells "workstations" for
> the same reasons HP no longer sells them, the main one being that
> "real" workstations no longer exist.

"Workstations" do exist, just because they have a x64 CPU doesn't make 
them less of a workstation.

> All the hardware savvy that made the difference between a workstation
> and a PC has now been integrated in the CPUs and support ICs. Sun cannot
> design a substantially faster processor than Intel or AMD. Sun cannot
> design a better motherboard than the likes of Supermicro, Tyan,
> Gigabyte or Asus.

Well, Dell and HP can (they use designs based on Supermicro mainboards).

> Sun cannot design a better case than Antec, etc. 

Why not? Everyone can design better cases than Antec.

> In this decade, Sun did design a totally different processor - something
> that cannot be said from HP, Dell or other even Intel. 

Well, it can be (and has been) done. intel for example had many new 
designs in the last decade (Netburst, Core 2, Nehalem, Larrabee). HP 
designed the PA-8800 and PA-8900 etc.

Dell never ever designed processors. This is above their scope.

> Sun added
> substantial innovations to their OS - something that cannot be said from
> Microsoft (they tried and failed),

MS did add a lot of new things to their OS (hint: MS has more operating 
systems than Vista).

> HP or IBM (who didn't even try). 

HP still improves HP-UX, and IBM puts new things in AIX and zOS.

> Sun didn't succeed, but you cannot fault them for not trying. After
> all, would we be having this discussion if Sun had become like HP, a
> technologically largely empty marketing organisation?

Sun didn't succed because their marketing was notoriously incompetent. 
HP on the other side succeeds in the workstation market because they 
have the right products for reasonable price, offer good service (I have 
had my nightmares with Sun service in the past) and does better 
marketing. Go figure.

Benjamin

Benjamin Gawert wrote:
> * Paul Gress:
> 
>> So now it looks as if Sun is going out of the workstation business 
>> entirely. 
> 
> They probably are for some time now.
> 
>> I hope this doesn't affect my CAD application which currently runs on 
>> Solaris x64 (Solaris 10 and SXCE NV114).
> 
> It will, maybe not necessarily in the long term, but Pro/E like other 
> similar programs will at some point just drop Solaris support.
> 
>

At that point, I'll just drop maintenance and stagnate or start looking 
for another Solaris CAD application.  I will NOT, move to Windows.

Benjamin Gawert wrote:

> 
> It will, maybe not necessarily in the long term, but Pro/E like other 
> similar programs will at some point just drop Solaris support.
> 

I also must say, Solaris has outlasted all other Unixes.  Pro/E use to 
run on SGI Irix, HPUX, and Linux and I believe AIX, but I can't verify 
that one yet.  Currently (WF4), it runs on HPUX, Solaris Sparc/X64 and 
of course Windows.  The next release (WF5), now in pretesting stage only 
runs on Solaris X64 and Windows XP/Vista.  Solaris is still hanging in 
there and will at least be there for the next release.  I haven't yet 
researched WF6.

Benjamin Gawert wrote:
> * Paul Gress:
> 
>> My CAD product name is Pro-Engineer Wildfire 4.
>>
>> Yes, without OpenGL it would be painfully slow.  The next Graphic card 
>> I plan to use is the Nvidia Quadro FX5800. 
> 
> A Quadro FX 5800 for Pro/E is wasted money. I had a Quadro FX 5600, but 
> even that for Pro/E it was wasted. If all you need is Pro/E, a cheaper 
> midrange card is a much more sensible choice.
> 

You may be right, on my laptop the Quadro FX1600m seems pretty good.  I 
haven't yet caused it to loose frames.

Paul

On Jun 2, 6:53=A0am, Benjamin Gawert <bgaw...@gmx.de> wrote:

> Sun didn't succed because their marketing was notoriously incompetent.
> HP on the other side succeeds in the workstation market because they
> have the right products for reasonable price, offer good service (I have
> had my nightmares with Sun service in the past) and does better
> marketing. Go figure.

I think that this (and many of the other articles in this thread)
utterly misunderstood what has happened to the workstation market.

20 years ago, there was a thriving Unix workstation market.
Workstations were much more powerful than PCs and cost correspondingly
more (a PC was maybe $2000, a workstation was upwards of $10,000).

From around 1995, this distinction went away: PCs became very similar
to workstations, and eventually identical.  Windows also ate the Unix
desktop market.  This killed workstations as a viable market: there
was very limited profit to be made from the hardware, and the tiny
size of the software marked meant that it was not really possible to
invest the money needed for graphics drivers and so on, unless you
could somehow people to work for free for you (Linux).

The only way it is possible to survive making Unix workstations now is
by doing two things:
1. already have a large Windows PC business, and rebadge some of your
highest-end PCs as workstations;
2. either run Windows on them, or make use of the work done for free
by Linux people for you.

Lo and behold that is what Dell, HP and so forth are doing.  Sun, on
the other hand, have never had a large Windows PC market and are
probably not interested in having one.

Why is this so surprising?  The only surprising thing is how long Sun
held out in the workstation market (the reason for which is reasonably
obvious: they needed to provide a development & test platform for
servers).

Paul Gress wrote:
> Stefaan A Eeckels wrote:
> 
>>
>> That is the point - what value can Sun add to Xeon 5500 processors,
>> Supermicro boards, RAM chips and SATA disks? A different case design?
>>
>> The only thing that Sun could try and do is sell more of them, so that
>> they can pocket the discounts. Sun no longer sells "workstations" for
>> the same reasons HP no longer sells them, the main one being that
>> "real" workstations no longer exist.
>>
>> All the hardware savvy that made the difference between a workstation
>> and a PC has now been integrated in the CPUs and support ICs. Sun cannot
>> design a substantially faster processor than Intel or AMD. Sun cannot
>> design a better motherboard than the likes of Supermicro, Tyan,
>> Gigabyte or Asus. Sun cannot design a better case than Antec, etc.
>> In this decade, Sun did design a totally different processor - something
>> that cannot be said from HP, Dell or other even Intel. Sun added
>> substantial innovations to their OS - something that cannot be said from
>> Microsoft (they tried and failed), HP or IBM (who didn't even try).
>> Sun didn't succeed, but you cannot fault them for not trying. After
>> all, would we be having this discussion if Sun had become like HP, a
>> technologically largely empty marketing organisation?
>>
> 
> You make some very good points.  What company makes workstations any 
> more today, and then if you think you found one, what differentiates it 
> from other workstation manufactures.
> 
> OK, some of my thoughts.
> 
> I looked at the Sun X2270 server, it has all the makings for a good 
> workstation, only deficient in PCI express slots (only has one) and no 
> provisions for an internal CD/DVD drive.  Basically I could use this if 
> it were repackaged in another case with a CD/DVD drive (to use the USB). 
>  I basically don't need to use the PCI expansion slots as long as it has 
> USB.  USB is easy to expand.
> 
> So, Sun has the workings/beginnings for a good workstation, they just 
> don't do it.
> 
> I might try to do it myself, but to go all through the design work for 
> just one isn't worth it.  I couldn't do multiple ones to sell as I 
> believe you need to be a certified reseller or VAR.
> 
> 

To use the Sun name, yes!  You could still build the "GressStation" 
using Sun parts plus your own custom stuff.

Could you sell enough to make a profit?

Benjamin Gawert wrote:
> * Richard B. Gilbert:
> 
>> Can Sun really afford to launch a family of workstations that support, 
>> say, 96 or 128GB of RAM? 
> 
> Yes, they can, as every other big player like HP or Dell can do that for 
> many years now. Especially since memory is so cheap.
> 
>> I suspect it's a niche market, meaning low volume.  That, in turn, 
>> means expensive!
> 
> Well, as to a low volume market and expensive systems, if you look at 
> all the other offerings that are there you will see that this isn't the 
> case.
> 

There must be some reason why someone isn't doing it.  If there were 
money to be made, someone would do it!

Why don't YOU build and sell custom CAD/CAM workstations with 256 GB of 
RAM?  Might as well put quad processors in it as well; if you're going 
to build Frankenstein's Monster, you might as well go whole hog!

Two graphics heads. . . .  Check!
Chilled water for cooling. . . . Check!
1000 HP Motor-Generator for 5VDC. . . .  Check!
50KW substation for power. . . . Check!

On 2009-06-02 12:09:39 +0100, Tim Bradshaw <tfb+google@tfeb.org> said:

> On Jun 2, 6:53�am, Benjamin Gawert <bgaw...@gmx.de> wrote:
> 
>> Sun didn't succed because their marketing was notoriously incompetent.
>> HP on the other side succeeds in the workstation market because they
>> have the right products for reasonable price, offer good service (I have
>> had my nightmares with Sun service in the past) and does better
>> marketing. Go figure.
> 
> I think that this (and many of the other articles in this thread)
> utterly misunderstood what has happened to the workstation market.
> 
> 20 years ago, there was a thriving Unix workstation market.
> Workstations were much more powerful than PCs and cost correspondingly
> more (a PC was maybe $2000, a workstation was upwards of $10,000).
> 
> From around 1995, this distinction went away: PCs became very similar
> to workstations, and eventually identical.  Windows also ate the Unix
> desktop market.  This killed workstations as a viable market: there
> was very limited profit to be made from the hardware, and the tiny
> size of the software marked meant that it was not really possible to
> invest the money needed for graphics drivers and so on, unless you
> could somehow people to work for free for you (Linux).
> 
> The only way it is possible to survive making Unix workstations now is
> by doing two things:
> 1. already have a large Windows PC business, and rebadge some of your
> highest-end PCs as workstations;
> 2. either run Windows on them, or make use of the work done for free
> by Linux people for you.

Or:

3. be Apple.

:-)

Though to be fair they're probably not generally sold as "Unix" boxes, 
rather as Macs.

-- 
Chris

* Paul Gress:

>> It will, maybe not necessarily in the long term, but Pro/E like other 
>> similar programs will at some point just drop Solaris support.
> 
> At that point, I'll just drop maintenance and stagnate or start looking 
> for another Solaris CAD application. 

Which as other ISVs (the few that haven't dropped Solaris already, that 
means) are doing the same this will probably get very hard.

> I will NOT, move to Windows.

Well, this could very likely mean that you will lock yourself out of the 
majority of commercial CAD applications.

BTW: all my major workstations run under Windows (Vista x64). Not only 
it is rock solid, for my applications the performance is much better 
than on Linux.

Benjamin

* Richard B. Gilbert:

>>> Can Sun really afford to launch a family of workstations that 
>>> support, say, 96 or 128GB of RAM? 
>>
>> Yes, they can, as every other big player like HP or Dell can do that 
>> for many years now. Especially since memory is so cheap.
>>
>>> I suspect it's a niche market, meaning low volume.  That, in turn, 
>>> means expensive!
>>
>> Well, as to a low volume market and expensive systems, if you look at 
>> all the other offerings that are there you will see that this isn't 
>> the case.
>>
> 
> There must be some reason why someone isn't doing it.  If there were 
> money to be made, someone would do it!

What makes you think no-one does it? Did you actually read my post? I 
already said that HP does it all the time as does Dell, Lenovo, FSC.

> Why don't YOU build and sell custom CAD/CAM workstations with 256 GB of 
> RAM?  Might as well put quad processors in it as well; if you're going 
> to build Frankenstein's Monster, you might as well go whole hog!

Besides that I see no reason to try to compete in a market where big 
names like HP are very active (building workstations), of course my 
workstations have quadcore processors since they became available ;-)

Benjamin

* Benjamin Gawert:

>> There must be some reason why someone isn't doing it.  If there were 
>> money to be made, someone would do it!
> 
> What makes you think no-one does it? Did you actually read my post? I 
> already said that HP does it all the time as does Dell, Lenovo, FSC.

Just to show you that this is not really a small niche market, just an 
example what HP offers for this market:

HP z800 (two quadcore Nehalem XEONs, up to 192GB RAM, 6 SATA2 and 8 SAS 
channels, up to two Quadro FX 5800 with 4GB memory)
<http://h20195.www2.hp.com/v2/GetPDF.aspx/c01713609.pdf>

This is the predecessor, the xw8600 (two quadcore XEON 5400, up to 128GB 
RAM, 6x SATA2 and 8x SAS):
<http://h18000.www1.hp.com/products/quickspecs/12849_ca/12849_ca.pdf>

Similar systems are available from Dell (Precision T7500) and Lenovo 
(ThinkStation D20).

Benjamin

Benjamin Gawert wrote:

> 
> BTW: all my major workstations run under Windows (Vista x64). Not only 
> it is rock solid, for my applications the performance is much better 
> than on Linux.
> 

I have been using computers since 1990.  Heck I even liked Windows then. 
  But when I started seeing how unfair Microsoft was with competition I 
decided to vote with my wallet and started using Interactive Unix.  If 
you sell an operating system, you cannot deliberately lock out potential 
competition by putting locks internally in your operating system.

So since about 1992 I was on Unix.  I still will not use Windows as a 
major OS, but only on a as needed basis.

Paul

Paul Gress wrote:
> Benjamin Gawert wrote:
> 
>>
>> BTW: all my major workstations run under Windows (Vista x64). Not only 
>> it is rock solid, for my applications the performance is much better 
>> than on Linux.
>>
> 
> I have been using computers since 1990.  Heck I even liked Windows then. 
>  But when I started seeing how unfair Microsoft was with competition I 
> decided to vote with my wallet and started using Interactive Unix.  If 
> you sell an operating system, you cannot deliberately lock out potential 
> competition by putting locks internally in your operating system.
> 
> So since about 1992 I was on Unix.  I still will not use Windows as a 
> major OS, but only on a as needed basis.
> 

I use Windows/XP, Solaris 8, 9, & 10, RHEL V4, and OpenVMS on various 
machines.  Each O/S has things it does better.  I use the system best 
suited to the task at hand.

On Tue, 02 Jun 2009 06:53:19 +0100
Benjamin Gawert <bgawert@gmx.de> wrote:

> * Stefaan A Eeckels:
> 
> > That is the point - what value can Sun add to Xeon 5500 processors,
> > Supermicro boards, RAM chips and SATA disks? A different case
> > design?
> 
> Sun can add the same value as HP or Dell or Lenovo can offer, and
> which is the reason that in every serious workstation application
> no-one even touches a system made from generic off-the-shelf parts:
> offering a complete system which is extensively tested (something
> which assemblers don't do), comes with appropriate service options
> and is certified for most professional applications.

"Certification" is - especially for managers who like to cover
their asses - an added value that does sell high-priced kit. It's a
bit like people buying gold speaker cables because they believe it
sounds better. But this is marketing, not technology, and that is
exactly my point. Sun is still an old-school technology company that
(somewhat pathetically) hopes to sell stuff on technical merit. Point in
case is that they have soldiered on with technological innovations such
as ZFS when they should have known that it could/would upset their
server customers. 

> > The only thing that Sun could try and do is sell more of them, so
> > that they can pocket the discounts. Sun no longer sells
> > "workstations" for the same reasons HP no longer sells them, the
> > main one being that "real" workstations no longer exist.
> 
> "Workstations" do exist, just because they have a x64 CPU doesn't
> make them less of a workstation.

"Workstation" today is marketing speak for "souped-up PC". I can build
one from off-the-shelf parts. The only thing it won't have is HP's
moniker. 

> > All the hardware savvy that made the difference between a
> > workstation and a PC has now been integrated in the CPUs and
> > support ICs. Sun cannot design a substantially faster processor
> > than Intel or AMD. Sun cannot design a better motherboard than the
> > likes of Supermicro, Tyan, Gigabyte or Asus.
> 
> Well, Dell and HP can (they use designs based on Supermicro
> mainboards).

Exactly. Just as my daughter's Compaq is based on an Asus motherboard.
HP can no longer design a CPU, can no longer design a motherboard, and
probably would struggle designing a PC case. It's all done by
subcontractors to HP specs.

> > Sun cannot design a better case than Antec, etc. 
> 
> Why not? Everyone can design better cases than Antec.

Oh - who's currently in favour? Lian-Li, Chieftec? What I meant is that
it is neigh impossible to sit down and design something substantially
better than an off-the-shelf part, and then sell it at a profit unless
one goes for the mass-market. 

> > In this decade, Sun did design a totally different processor -
> > something that cannot be said from HP, Dell or other even Intel. 
> 
> Well, it can be (and has been) done. intel for example had many new 
> designs in the last decade (Netburst, Core 2, Nehalem, Larrabee). HP 
> designed the PA-8800 and PA-8900 etc.

These are not new, they are merely refinements of existing designs. 

Neither HP nor Intel took a gamble on a heavily multi-threaded CPU with
lousy single-thread performance. Call it ill-advised, call it a
marketing disaster - but at least is was a fresh look at processor
design.

> Dell never ever designed processors. This is above their scope.

Given HP's incredible savvy in replacing PA-RISC with Itanium, that's
probably a good thing.

> > Sun added
> > substantial innovations to their OS - something that cannot be said
> > from Microsoft (they tried and failed),
> 
> MS did add a lot of new things to their OS (hint: MS has more
> operating systems than Vista).

Yes, marvelous innovations like finally dropping the GUI from Windows
Server 2008. Hell, it took them 15 years to shake off the diktat of
their marketing droids (and Citrix to get a working remote desktop :).
If any OS is in need of a better file system it's Windows - and guess
what, that's exactly what Microsoft failed to deliver on.

> > HP or IBM (who didn't even try). 
> 
> HP still improves HP-UX, and IBM puts new things in AIX and zOS.

PH-UX is dying and you know it. HP would dearly love their existing
customers to switch to Windows so that they can close down their HP-UX
"cost" centre. Re-selling Windows is far less risky and makes more
marketing sense. 

The transformation that turned HP from a revered technology company to
a soulless marketing outfit has been one of the saddest events of the
last decade. Yes, they still exist, their marketing is first-rate, they
sell a lot of PCs, and their x86 servers are no better and no worse than
those of Dell (discounting the need to pay for iLO which cost them our
business). Unfortunately, their "HP invent" slogan is nothing more
than the abbreviation of 

"HP invent... yet another incompatible ink cartridge".

-- 
Stefaan A Eeckels
-- 
And as crazy as this sounds, people tend to be able to manage systems
better if they have a good internal mental model of how the system 
works.                                              --Logan Shaw

In article <78l4noF1n6pkqU1@mid.individual.net>,
Benjamin Gawert  <bgawert@gmx.de> wrote:
>Just to show you that this is not really a small niche market, just an 
>example what HP offers for this market:

[snipped]

>Similar systems are available from Dell (Precision T7500) and Lenovo 
>(ThinkStation D20).

HP and Dell purchased VoodooPC and Alienware, respectively.
<URL:http://www.theregister.co.uk/2006/09/29/hp_buys_voodoopc/>
<URL:http://www.theregister.co.uk/2006/03/23/dell_buys_alienware/>

I think there's still a market for high-end stationary workstations
for developers, content creators and gamers, but the geniuses in Sun
who only months ago anounced their first mobile x64 workstation systems
either disagree or can't make Sun's ERP system support them.

The geniuses prefer we use SunRays. IBM's marketing wonks offer
a similar server/thin-client solution:
<URL:http://www-03.ibm.com/systems/bladecenter/hardware/workstation/>

John
groenveld@acm.org

* Stefaan A Eeckels:

> "Certification" is - especially for managers who like to cover
> their asses - an added value that does sell high-priced kit. It's a
> bit like people buying gold speaker cables because they believe it
> sounds better. But this is marketing, not technology, and that is
> exactly my point. 

Well, if you ever tried to get support from the ISV of your software 
earn your money with you quickly see that certification is in fact no 
marketing thing. Certifications are the requirements for most 
professional applications to be able to get any support.

Also, when a HP or Dell workstation which comes with support and is 
fully tested doesn't cost more as the same system self-build by standard 
components then I can't see how this can be "high priced". Especially 
since even entry level single processor workstations from Dell, HP and 
Lenovo etc are alsready certified.

BTW: Most Sun workstations like other RISC workstations were certified 
for certain application. Go figure.

> Sun is still an old-school technology company that
> (somewhat pathetically) hopes to sell stuff on technical merit. 

This might be the case, but technical merit doesn't make the need of 
customer orientation and good marketing go away.

> "Workstation" today is marketing speak for "souped-up PC". 

The same could be said about a Ultra 45.

> I can build
> one from off-the-shelf parts. The only thing it won't have is HP's
> moniker. 

The same is valid for a Sun workstation. Several SPARC mainboards were 
available.

>> Well, Dell and HP can (they use designs based on Supermicro
>> mainboards).
> 
> Exactly. Just as my daughter's Compaq is based on an Asus motherboard.

Your daughters Compaq is a consumer PC made from the cheapest components 
available.

> HP can no longer design a CPU, 

Wrong. Not too long ago they designed the PA-8800 and PA-8900. HP is 
also involved in Itanium development.

> can no longer design a motherboard,

Nonsense. Besides that the mobo of the new z800 is a HP design, HP 
designs all system boards for their servers (ProLiants, Integrity) 
except for the very cheap entry level ProLiant 100 series.

> and
> probably would struggle designing a PC case. 

All of their workstation and server cases are designed by HP.

> It's all done by
> subcontractors to HP specs.

All that is done by subcontractors is production, because that is what 
they can do cheaper.

HP still does a lot of development. Only their consumer PCs are designed 
and made by different sources as all they need to be is cheap. One 
should however not make the mistake to mix consumer grade Presarios with 
their business products.

>>> Sun cannot design a better case than Antec, etc. 
>> Why not? Everyone can design better cases than Antec.
> 
> Oh - who's currently in favour? Lian-Li, Chieftec? 

None of them. They all make cases that attract gamers and PC freaks for 
building their next overclocked gaming rig.

> What I meant is that
> it is neigh impossible to sit down and design something substantially
> better than an off-the-shelf part, and then sell it at a profit unless
> one goes for the mass-market. 

Again, the cases that Dell, HP, Lenovo and also Apple use are 
substantially better, and despite their systems being not really more 
expensive than a self built system (except Apple of course) they can 
sell all that at a profit.

>>> In this decade, Sun did design a totally different processor -
>>> something that cannot be said from HP, Dell or other even Intel. 
>> Well, it can be (and has been) done. intel for example had many new 
>> designs in the last decade (Netburst, Core 2, Nehalem, Larrabee). HP 
>> designed the PA-8800 and PA-8900 etc.
> 
> These are not new, they are merely refinements of existing designs. 

As is the SPARC.

> Neither HP nor Intel took a gamble on a heavily multi-threaded CPU with
> lousy single-thread performance. Call it ill-advised, call it a
> marketing disaster - but at least is was a fresh look at processor
> design.

Maybe, but the UltraSPARC T1 was nothing completely new. It like other 
was just another evolution of the SPARC architecture, just with a 
different design goal (multithreaded).

>> MS did add a lot of new things to their OS (hint: MS has more
>> operating systems than Vista).
> 
> Yes, marvelous innovations like finally dropping the GUI from Windows
> Server 2008. Hell, it took them 15 years to shake off the diktat of
> their marketing droids (and Citrix to get a working remote desktop :).
> If any OS is in need of a better file system it's Windows - and guess
> what, that's exactly what Microsoft failed to deliver on.

This might seem to be the case if you are full of MS hate but reality is 
that NTFS while not being as advanced as ZFS of course is quite good and 
much better than most file systems Linux has to offer.

As to innovations, Windows uses internal a lot of new things that are 
different from other OSes.

While most UNIXes and Linux share a lot of same things (in fact, Sun 
seems to be the only innovative party in the UNIX world), Windows is a 
completely different approach. A little less hate and bit more objective 
view would probably help to realize that.

>> HP still improves HP-UX, and IBM puts new things in AIX and zOS.
> 
> PH-UX is dying and you know it.

It is suffering (same as Solaris SPARC is suffering), heck, even AIX is 
suffering, but this doesn't change the fact that HP and IBM still 
continue development of their UNIX variants.

>  HP would dearly love their existing
> customers to switch to Windows so that they can close down their HP-UX
> "cost" centre. Re-selling Windows is far less risky and makes more
> marketing sense. 

Thats nonsense. In the business where HP-UX is, Windows isn't the 
favourite platform. Besides, the real problem is the too expensive 
Itanium platform and less the operating system.

And for x64, HP not only supports Windows and Linux but also Solaris on 
their ProLiant servers.

Benjamin

* Paul Gress:

> I have been using computers since 1990.  Heck I even liked Windows then. 
>  But when I started seeing how unfair Microsoft was with competition I 
> decided to vote with my wallet and started using Interactive Unix.  If 
> you sell an operating system, you cannot deliberately lock out potential 
> competition by putting locks internally in your operating system.

Well, that's what the trials against Microsoft in the US and in Europe 
are for, to put this straight.

And besides that: Windows today has not much in common with Windows of 
1990 (Win 3.x) or even 1998 (Win9x).

> So since about 1992 I was on Unix.  I still will not use Windows as a 
> major OS, but only on a as needed basis.

I have to agree with Richard here, the best OS is the one that does the 
job. All OSes have pros and cons, in the end it should be choosen after 
the task and not after ideology, especially when its used to earn a living.

Benjamin

In article <78mibiF1mhjhpU1@mid.individual.net>,
	Benjamin Gawert <bgawert@gmx.de> writes:
> * Stefaan A Eeckels:
> 
>> HP can no longer design a CPU, 
> 
> Wrong. Not too long ago they designed the PA-8800 and PA-8900. HP is 

That's some years ago now. HP disbanded their processor development
about 5 years ago. (Well, in theory the group was all transfered to
Intel as part of HP's payment to Intel to continue devloping Itanium,
although in practice the majority left for other companies, including
Sun.) HP have had no CPU developement capability since then.

> also involved in Itanium development.

Well, it had to pay Intel $3B to continue developing Itanium
because sales of the processors alone won't cover it, so I guess that
could be said to be quite a substantial involvement.

> Maybe, but the UltraSPARC T1 was nothing completely new. It like other 
> was just another evolution of the SPARC architecture, just with a 
> different design goal (multithreaded).

It sure was backwards compatible. Sun knows better than to make an
Itanium style mistake. However, it was a completely new design, and
a new way of thinking at how to solve the clock speed ceiling which
all processor manfuacturers had hit (increasing clock speed further
was only getting miniscule performance gains, but with vastly higher
costs and power consumption -- something which had stopped scaling
in the way it had been in the previous 2-3 decades).

>>> HP still improves HP-UX, and IBM puts new things in AIX and zOS.
>> 
>> PH-UX is dying and you know it.
> 
> It is suffering (same as Solaris SPARC is suffering), heck, even AIX is 
> suffering, but this doesn't change the fact that HP and IBM still 
> continue development of their UNIX variants.

IBM seem to have put in a burst of effort over the last few years,
(since Solaris 10 released) catching up on things like WPARS, after
not seeming to be very interested in doing much to AIX for the years
before.

HP has been really stung by having driven their customers up the
Intanium dead-end. They did this because they didn't have enough
sales to fund the continued development of HP-PA, and needed to
move to a commodity chip. The trouble is that they lost a load of
customers in the process, Itanium never became a commodity chip,
and it's numbers are lower than HP-PA was, making it even less
viable. The customer base wouldn't stand another shift, so they're
sunk. I think the only possible escape route they ever had in this
path would have been to jump ship to AMD64 when it appeared, which
would have opened up a good possibility for HP-UX on Nehalem today,
but they were too emotionally attached to Itamium for that to have
been considered. Add to this that HP-UX forked from an anchient
SysV which makes putting modern features in it in a performant
way much more expensive than with the more modern unixs, and that
HP have always bought significant subsystems for HP-UX from 3rd
parties so they don't have the same in-house expertise that Sun
and IBM do (same position they're in with their processor now),
and they really are in a much more difficult/challenging position
than other unix vendors.

>>  HP would dearly love their existing
>> customers to switch to Windows so that they can close down their HP-UX
>> "cost" centre. Re-selling Windows is far less risky and makes more
>> marketing sense. 
> 
> Thats nonsense. In the business where HP-UX is, Windows isn't the 
> favourite platform. Besides, the real problem is the too expensive 
> Itanium platform and less the operating system.
> 
> And for x64, HP not only supports Windows and Linux but also Solaris on 
> their ProLiant servers.
> 
> Benjamin

-- 
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

On May 30, 7:55 am, Paul Floyd <r...@127.0.0.1> wrote:
> On Fri, 29 May 2009 18:46:26 -0700 (PDT), ThanksButNo
>    <no.no.tha...@gmail.com> wrote:
>
> > These days, there is very little that would surprise me!
>
> > I can imagine a 128G solid state "ram" that plugs into
> > a USB 4.0 port and runs circles around anything we have
> > today.  And high school kids walking around everywhere
> > with 'em hanging off their key-chains, and losing 'em,
> > and mom says, "That's Ok, we'll get some more next time
> > we stop at the 99 cent store."  No worries about the
> > data, thanks to pervasive broadband wi-fi auto-backup
> > systems connected on every cell-phone tower.
>
> > And we'll be codgers whining about, "I remember when all
> > we had were 8 inch floppies, and ... zzzZZ ZZzzz ..."
>
> DRAM and Flash densities are definitily running out of steam. Both use
> charge stored in a capacitor, and as the number of atoms (and electrons)
> starts to become quite small, then they don't work too well. That means
> not much scope for more 2D shrinking. The next think might be to add
> more poly layers (so the devices become 3D), subject to heat dissipation
> issues. And then there are the possibilities of completely new
> architectures - optical, quantum, magnetic or whatever.
>
> In any case, I can't see the same factor happening again between the 1st
> PC RAM that I bought (2Mbyte sticks) and the last (1Gbyte sticks) with
> DRAM.

Funny thing is, historically, there has always been a very
good technical reason why Moore's Law can no longer happen,
and all the experts in the trade tell us all that computers
are finally as fast now as they possibly can ever be, then some
jerk goes and ruins everything by finding a way around that
technical reason and Moore's Law continues rumbling forward
without even so much as pausing for breath.

I still can't get over modem rates having surpassed 9600 baud.

So -- maybe there's a good technical reason why we can't store
terabytes on a keychain-mounted thumb-drive -- today -- but I
wouldn't bet against somebody, somewhere, somehow, finding some
way to do it anyway -- tomorrow!

:-)

On Tue, 02 Jun 2009 08:14:11 -0400, "Richard B. Gilbert" <rgilbert88@comcast.net> wrote:
> Why don't YOU build and sell custom CAD/CAM workstations with 256 GB
> of RAM?  Might as well put quad processors in it as well; if you're
> going to build Frankenstein's Monster, you might as well go whole hog!
>
> Two graphics heads. . . .  Check!
> Chilled water for cooling. . . . Check!
> 1000 HP Motor-Generator for 5VDC. . . .  Check!
> 50KW substation for power. . . . Check!

There's probably a market for high-end workstations.  I'm not sure how
big it is, but those who work with large CAD, architects designing huge
3-D views of their work, and graphics artists commonly have 10x the raw
power I am using for network-related stuff right there, on their desk :)

On Jun 2, 1:34=A0pm, Chris Ridd <chrisr...@mac.com> wrote:

> 3. be Apple.

Yes, missed that one.  But Apple must make a lot of use of the work
done for free by people - most of the userland (outside the GUI) is
pretty much either GNU or *BSD based.

On Jun 3, 9:03=A0am, and...@cucumber.demon.co.uk (Andrew Gabriel) wrote:

> IBM seem to have put in a burst of effort over the last few years,
> (since Solaris 10 released) catching up on things like WPARS, after
> not seeming to be very interested in doing much to AIX for the years
> before.
>
> HP has been really stung by having driven their customers up the
> Intanium dead-end. They did this because they didn't have enough
> sales to fund the continued development of HP-PA, and needed to
> move to a commodity chip. The trouble is that they lost a load of
> customers in the process, Itanium never became a commodity chip,
> and it's numbers are lower than HP-PA was, making it even less
> viable. The customer base wouldn't stand another shift, so they're
> sunk. I think the only possible escape route they ever had in this
> path would have been to jump ship to AMD64 when it appeared, which
> would have opened up a good possibility for HP-UX on Nehalem today,
> but they were too emotionally attached to Itamium for that to have
> been considered. Add to this that HP-UX forked from an anchient
> SysV which makes putting modern features in it in a performant
> way much more expensive than with the more modern unixs, and that
> HP have always bought significant subsystems for HP-UX from 3rd
> parties so they don't have the same in-house expertise that Sun
> and IBM do (same position they're in with their processor now),
> and they really are in a much more difficult/challenging position
> than other unix vendors.

I think this is a good summary.  My impression (where I work we have
all three flavours, though I'm almost entirely on the Solaris bit) is
that HP-UX is slowly fading away - it was behind the other two 10
years ago and it is now really a long way behind.  it also doesn't
matter to HP, of course, who make money elsewhere.  Solaris is streets
ahead as an OS but terribly deficient in management tools and the
virtualisation story is an enormous mess.  AIX is horrible, but has
good management tools and they wrote the book on virtualisation.

On Wed, 3 Jun 2009 09:28:03 -0700 (PDT), ThanksButNo
   <no.no.thanks@gmail.com> wrote:
> On May 30, 7:55 am, Paul Floyd <r...@127.0.0.1> wrote:
>>
>> DRAM and Flash densities are definitily running out of steam. Both use
>> charge stored in a capacitor, and as the number of atoms (and electrons)
>> starts to become quite small, then they don't work too well. That means
>> not much scope for more 2D shrinking. The next think might be to add
>> more poly layers (so the devices become 3D), subject to heat dissipation
>> issues. And then there are the possibilities of completely new
>> architectures - optical, quantum, magnetic or whatever.
>>
>> In any case, I can't see the same factor happening again between the 1st
>> PC RAM that I bought (2Mbyte sticks) and the last (1Gbyte sticks) with
>> DRAM.
>
> Funny thing is, historically, there has always been a very
> good technical reason why Moore's Law can no longer happen,
> and all the experts in the trade tell us all that computers
> are finally as fast now as they possibly can ever be, then some
> jerk goes and ruins everything by finding a way around that
> technical reason and Moore's Law continues rumbling forward
> without even so much as pausing for breath.
>
> I still can't get over modem rates having surpassed 9600 baud.
>
> So -- maybe there's a good technical reason why we can't store
> terabytes on a keychain-mounted thumb-drive -- today -- but I
> wouldn't bet against somebody, somewhere, somehow, finding some
> way to do it anyway -- tomorrow!

I suggest that you try reading what I said, to the end.

For a 65nm process to gain a factor of 512 in density, then it's going
to have to shrink to about 3nm. The atomic radius of silicon is about
0.1nm. That means features that are about 30 atoms wide. Even if that is
possible (overcoming the problems of lithography, process variation,
changes to the behaviour due to non-linearities), then I sorely doubt
that it would be economically viable. Back when processes were well over
100nm, most semiconductor manufacturers could afford fabs and the
associated design to get the process working. Now we're down in the
30-40nm range, and there are 3 left running.

Since you brought up Moore's Law, this refers to a doubling of
transistor counts every 1-2 years. For 2D processes, all good things
come to an end.  At the end of the day, there are hard limits that nobody
can get around - the speed of light, the size of atoms, the laws of
thermodynamics.

Like I said, there might be something new one day (3D, quantum, optics),
and perhaps improved design (like ILP gave 'More Moore' to CPU speeds
for a while).  Of course, if you know better and have a way to shrink process
dimensions below, say, 10nm, then I suggest that you get down to the
patent office sharpish.

A bientot
Paul
-- 
Paul Floyd                 http://paulf.free.fr

On 2009-06-03 19:19:06 +0100, Tim Bradshaw <tfb+google@tfeb.org> said:

> On Jun 2, 1:34�pm, Chris Ridd <chrisr...@mac.com> wrote:
> 
>> 3. be Apple.
> 
> Yes, missed that one.  But Apple must make a lot of use of the work
> done for free by people - most of the userland (outside the GUI) is
> pretty much either GNU or *BSD based.

I suppose so, though they *have* given stuff back to the "community" - 
the one-way street thing is a bit of a myth.

They certainly do all their kernel and system framework development in-house.
-- 
Chris

On Jun 3, 12:33 pm, Paul Floyd <r...@127.0.0.1> wrote:
> On Wed, 3 Jun 2009 09:28:03 -0700 (PDT), ThanksButNo
>    <no.no.tha...@gmail.com> wrote:
>
> > On May 30, 7:55 am, Paul Floyd <r...@127.0.0.1> wrote:
>
> >> DRAM and Flash densities are definitily running out of steam. Both use
> >> charge stored in a capacitor, and as the number of atoms (and electrons)
> >> starts to become quite small, then they don't work too well. That means
> >> not much scope for more 2D shrinking. The next think might be to add
> >> more poly layers (so the devices become 3D), subject to heat dissipation
> >> issues. And then there are the possibilities of completely new
> >> architectures - optical, quantum, magnetic or whatever.
>
> >> In any case, I can't see the same factor happening again between the 1st
> >> PC RAM that I bought (2Mbyte sticks) and the last (1Gbyte sticks) with
> >> DRAM.
>
> > Funny thing is, historically, there has always been a very
> > good technical reason why Moore's Law can no longer happen,
> > and all the experts in the trade tell us all that computers
> > are finally as fast now as they possibly can ever be, then some
> > jerk goes and ruins everything by finding a way around that
> > technical reason and Moore's Law continues rumbling forward
> > without even so much as pausing for breath.
>
> > I still can't get over modem rates having surpassed 9600 baud.
>
> > So -- maybe there's a good technical reason why we can't store
> > terabytes on a keychain-mounted thumb-drive -- today -- but I
> > wouldn't bet against somebody, somewhere, somehow, finding some
> > way to do it anyway -- tomorrow!
>
> I suggest that you try reading what I said, to the end.

What makes you think I didn't read it, or that anything
you said is germane to my point?

Which was, there's no telling how big, how fast, how much
the technology is going to improve in the future.  Probably
sooner than any of us can predict.

Sure, maybe it won't use flash technology -- OR hard disks,
OR dram OR silicon -- But the point is, it'll be bigger and
faster and better than we have now.  And us old folks will
shake our heads in wonder, and say, "Gee, I remember back
when ..."


> Since you brought up Moore's Law, this refers to a doubling of
> transistor counts every 1-2 years.

Quoting the Wikipedia:

"Almost every measure of the capabilities of digital electronic
devices is strongly linked to Moore's law: processing speed, memory
capacity, sensors and even the number and size of pixels in digital
cameras.[6] All of these are improving at (roughly) exponential rates
as well."

> In any case, I can't see the same factor happening again between the 1st
> PC RAM that I bought (2Mbyte sticks) and the last (1Gbyte sticks) with
> DRAM.

Quoting Clarke's First Law:
"When a distinguished but elderly scientist states that something is
possible, he is almost certainly right. When he states that something
is impossible, he is very probably wrong."
    Arthur C. Clarke

No worries! Sit back and let the inventors invent.
Who knows what they'll come up with next.  Just
wait and see, and be prepared to be amazed!

8-)

Wouldn't it be HELPFUL, SALES- and PROFIT-ENHANCING for some
(sun)patriotic employee to edit this thread AND PROVIDE IT
TO EVERYONE IN TOP-LEVEL MANAGEMENT?

The business-press is always advising top-management to
go out and get FEEDBACK -- FROM THEIR OWN CUSTOMERS.

   As further boost (as excuse to do something): to NOT acquaint
   top management with some of the threads here (such as this one)
   could lead to accusations of not "going that extra mile" that
   jobs *these difficult days* seem to require.


CHEERS!

David

PS: whatever lengths you have to go to get sun back into
the workstation business, is fine by me!

David Combs wrote:
> Wouldn't it be HELPFUL, SALES- and PROFIT-ENHANCING for some
> (sun)patriotic employee to edit this thread AND PROVIDE IT
> TO EVERYONE IN TOP-LEVEL MANAGEMENT?
> 
> The business-press is always advising top-management to
> go out and get FEEDBACK -- FROM THEIR OWN CUSTOMERS.
> 
>    As further boost (as excuse to do something): to NOT acquaint
>    top management with some of the threads here (such as this one)
>    could lead to accusations of not "going that extra mile" that
>    jobs *these difficult days* seem to require.
> 

Of course there is the risk that you will be fired for wasting time 
reading newsgroups instead of doing what you were paid for!

On Jun 3, 8:07 pm, "Richard B. Gilbert" <rgilber...@comcast.net>
wrote:

>
> Of course there is the risk that you will be fired for wasting time
> reading newsgroups instead of doing what you were paid for!

That's why I don't use my real name!!!

8-D

On Wed, 3 Jun 2009 13:53:28 -0700 (PDT), ThanksButNo
   <no.no.thanks@gmail.com> wrote:
>>
>> I suggest that you try reading what I said, to the end.
>
> What makes you think I didn't read it, or that anything
> you said is germane to my point?

Mainly because your reply indicated to me that you had not understood
what I said.

> Which was, there's no telling how big, how fast, how much
> the technology is going to improve in the future.  Probably
> sooner than any of us can predict.

"the technology". Very precise.

> Sure, maybe it won't use flash technology -- OR hard disks,
> OR dram OR silicon -- But the point is, it'll be bigger and
> faster and better than we have now.  And us old folks will
> shake our heads in wonder, and say, "Gee, I remember back
> when ..."

Again this is more ore less what I originally wrote: DRAM and FLASH are
running out of steam, but there may be other technologies on the way.

>> Since you brought up Moore's Law, this refers to a doubling of
>> transistor counts every 1-2 years.
>
> Quoting the Wikipedia:
>
> "Almost every measure of the capabilities of digital electronic
> devices is strongly linked to Moore's law: processing speed, memory
> capacity, sensors and even the number and size of pixels in digital
> cameras.[6] All of these are improving at (roughly) exponential rates
> as well."

Great. A wikipedia quote. Can't argue with that. Saying essentiially
what I said as well. Except for pixel size of course. That's bullshit.
It is fundamentally limited by the wavelength of visible light. I can
see there being exponential growth until the pixel size reaches the
optical resolution limit, and the sensor size growing up to fill the
optical backplane. After that, it would just be marketing.

>> In any case, I can't see the same factor happening again between the 1st
>> PC RAM that I bought (2Mbyte sticks) and the last (1Gbyte sticks) with
>> DRAM.

> Quoting Clarke's First Law:
> "When a distinguished but elderly scientist states that something is
> possible, he is almost certainly right. When he states that something
> is impossible, he is very probably wrong."
>     Arthur C. Clarke

Great. More quotes. Must be right. 

So I'll say it for a third time, maybe it'll sink in. DRAM and FLASH are
running out of steam, but there may be other technologies on the way.

Time will tell of course.

A bientot
Paul
-- 
Paul Floyd                 http://paulf.free.fr

On Jun 4, 12:42 pm, Paul Floyd <r...@127.0.0.1> wrote:

> So I'll say it for a third time, maybe it'll sink in. DRAM and FLASH are
> running out of steam, but there may be other technologies on the way.

Did anybody say you were wrong?

\:-\

In article <C8GdnVrlwoAm-rjXnZ2dnUVZ_vudnZ2d@giganews.com>,
Richard B. Gilbert <rgilbert88@comcast.net> wrote:
>Paul Gress wrote:
>> Benjamin Gawert wrote:
>> 
>>>
>>> BTW: all my major workstations run under Windows (Vista x64). Not only 
>>> it is rock solid, for my applications the performance is much better 
>>> than on Linux.
>>>
>> 
>> I have been using computers since 1990.  Heck I even liked Windows then. 
>>  But when I started seeing how unfair Microsoft was with competition I 
>> decided to vote with my wallet and started using Interactive Unix.  If 
>> you sell an operating system, you cannot deliberately lock out potential 
>> competition by putting locks internally in your operating system.
>> 
>> So since about 1992 I was on Unix.  I still will not use Windows as a 
>> major OS, but only on a as needed basis.
>> 
>
>I use Windows/XP, Solaris 8, 9, & 10, RHEL V4, and OpenVMS on various 
>machines.  Each O/S has things it does better.  I use the system best 
>suited to the task at hand.
>
>

1: Please, what does windows do better than the others?

2: what is RHEL?


Thanks,

David

David Combs wrote:
> In article <C8GdnVrlwoAm-rjXnZ2dnUVZ_vudnZ2d@giganews.com>,
> Richard B. Gilbert <rgilbert88@comcast.net> wrote:
>> Paul Gress wrote:
>>> Benjamin Gawert wrote:
>>>
>>>> BTW: all my major workstations run under Windows (Vista x64). Not only 
>>>> it is rock solid, for my applications the performance is much better 
>>>> than on Linux.
>>>>
>>> I have been using computers since 1990.  Heck I even liked Windows then. 
>>>  But when I started seeing how unfair Microsoft was with competition I 
>>> decided to vote with my wallet and started using Interactive Unix.  If 
>>> you sell an operating system, you cannot deliberately lock out potential 
>>> competition by putting locks internally in your operating system.
>>>
>>> So since about 1992 I was on Unix.  I still will not use Windows as a 
>>> major OS, but only on a as needed basis.
>>>
>> I use Windows/XP, Solaris 8, 9, & 10, RHEL V4, and OpenVMS on various 
>> machines.  Each O/S has things it does better.  I use the system best 
>> suited to the task at hand.
> 
> 1: Please, what does windows do better than the others?

Run windows only applications.

-- 
Ian Collins

In article <601d2d26-eb66-4255-91f0-a13fd64f1152@l28g2000vba.googlegroups.com>,
Tim Bradshaw  <tfb+google@tfeb.org> wrote:
>...
....
> ...  Windows also ate the Unix
>desktop market.  This killed workstations as a viable market ...

Well, Scott McNealy's "we're going to own the desktop" (meaning pc's,
everything) due to Sun being first to have graphic-displays (bouncing
balls, etc) -- might that challange have had something to do with it?


David

On Jun 26, 1:02=A0pm, dkco...@panix.com (David Combs) wrote:

>
> 2: what is RHEL?
>

Red Hat Enterprise Linux

/:-/

On Fri, 26 Jun 2009 20:02:51 +0000 (UTC), dkcombs@panix.com
(David Combs) wrote:

>2: what is RHEL?

Red Hat Enterprise Linux
-- 
   )  Kees
  (
c[_]    "... After all, all he did was string together a lot of old, well-known quotations."
     -- H. L. Mencken, on Shakespeare [#512]

David Combs <dkcombs@panix.com> wrote:
> 2: what is RHEL?

Red Hat Enterprise Linux

rick jones
-- 
I don't interest myself in "why". I think more often in terms of
"when", sometimes "where"; always "how much."  - Joubert
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

David Combs wrote:
> In article <C8GdnVrlwoAm-rjXnZ2dnUVZ_vudnZ2d@giganews.com>,
> Richard B. Gilbert <rgilbert88@comcast.net> wrote:
>> Paul Gress wrote:
>>> Benjamin Gawert wrote:
>>>
>>>> BTW: all my major workstations run under Windows (Vista x64). Not only 
>>>> it is rock solid, for my applications the performance is much better 
>>>> than on Linux.
>>>>
>>> I have been using computers since 1990.  Heck I even liked Windows then. 
>>>  But when I started seeing how unfair Microsoft was with competition I 
>>> decided to vote with my wallet and started using Interactive Unix.  If 
>>> you sell an operating system, you cannot deliberately lock out potential 
>>> competition by putting locks internally in your operating system.
>>>
>>> So since about 1992 I was on Unix.  I still will not use Windows as a 
>>> major OS, but only on a as needed basis.
>>>
>> I use Windows/XP, Solaris 8, 9, & 10, RHEL V4, and OpenVMS on various 
>> machines.  Each O/S has things it does better.  I use the system best 
>> suited to the task at hand.
>>
>>
> 
> 1: Please, what does windows do better than the others?
> 
The user interface.  Windows had serious reliability problems many years 
ago but I've used W/2K and W/XP for many years now with NO problems.

> 2: what is RHEL?
Red Hat Enterprise Linux

quoting Richard B. Gilbert (Fri, 26 Jun 2009 21:18:07 -0400):

> The user interface

Are you kidding me? That candy click-a-d'click interface? Even OS-X does
a far better job. Gnome/Compiz has a much larger appeal to me. And also
I like the fact that I can choose minimalistic desktops too. The windows
interface does not suck, but to say it's doing a much better job that
others...?

-- 
Dick Hoogendijk -- PGP/GnuPG key: 01D2433D
+http://nagual.nl/ | SunOS 10u7 05/09 ZFS+

Dick Hoogendijk <dick@nagual.nl> wrote:
> quoting Richard B. Gilbert (Fri, 26 Jun 2009 21:18:07 -0400):
> 
>> The user interface
> 
> Are you kidding me? That candy click-a-d'click interface? Even OS-X does
> a far better job. Gnome/Compiz has a much larger appeal to me. And also
> I like the fact that I can choose minimalistic desktops too. The windows
> interface does not suck, but to say it's doing a much better job that
> others...?
> 

Indeed. It was the Windows user interface that first drove me to Linux where
windows did what *I* told them to, not what some drug addled ergonomics
specialist in Microsoft decided would be best.

-- 
'The Voyage of the Beagle' by Darwin is available for free from
Project Gutenberg http://www.gutenberg.org/wiki/Main_Page

Dick Hoogendijk wrote:
> quoting Richard B. Gilbert (Fri, 26 Jun 2009 21:18:07 -0400):
> 
>> The user interface
> 
> Are you kidding me? That candy click-a-d'click interface? Even OS-X does
> a far better job. Gnome/Compiz has a much larger appeal to me. And also
> I like the fact that I can choose minimalistic desktops too. The windows
> interface does not suck, but to say it's doing a much better job that
> others...?
> 

OS-X SHOULD do it better!  Apple had a GUI when the rest of the world 
was still chipping flint!

On 2009-06-27 09:16:24 +0100, Dick Hoogendijk <dick@nagual.nl> said:

> Are you kidding me? That candy click-a-d'click interface? Even OS-X does
> a far better job. Gnome/Compiz has a much larger appeal to me. And also
> I like the fact that I can choose minimalistic desktops too. The windows
> interface does not suck, but to say it's doing a much better job that
> others...?

Other than OS X, it clearly is, witness the fact that all the other 
pretenders to be desktop OSs (again, with the exception of OS X) are 
clearly simply windows knock-offs, occasionally with some extra candy.  
Yes, you can find alternative environments, especially if you are 
willing to put them together yourself, but you could always do that.

On Sat, 27 Jun 2009, Tim Bradshaw wrote:

> Other than OS X, it clearly is, witness the fact that all the other pretenders
> to be desktop OSs (again, with the exception of OS X) are clearly simply
> windows knock-offs, occasionally with some extra candy.  Yes, you can find

I respectfully disagree.  Yes, most other GUIs are essentially Windoze knock-offs
(with, as you say, the notable exception of MacOS), but I assert that's not
because Windoze is deemed to be superior.  Rather, it is becuase the vendors
of those other GUIs are trying to capture some of the Windoze market and are
using the visual familiarity as part of the lowing the barrier to entry.

-- 
Rich Teer, SCSA, SCNA, SCSECA

URLs: http://www.rite-group.com/rich
      http://www.linkedin.com/in/richteer

On 2009-06-27 16:10:35 +0100, Rich Teer <rich.teer@rite-group.com> said:

> I respectfully disagree.  Yes, most other GUIs are essentially Windoze 
> knock-offs
> (with, as you say, the notable exception of MacOS), but I assert that's not
> because Windoze is deemed to be superior.  Rather, it is becuase the vendors
> of those other GUIs are trying to capture some of the Windoze market and are
> using the visual familiarity as part of the lowing the barrier to entry.

I think this comes down to semantics.  Clearly the Windows GUI is not 
the best GUI one could imagine (as an existence proof: the OS X GUI is 
better).  However, it is better than the GUIs that the other desktop 
platforms actually *provide* (let's just ignore the mac for now), which 
are generally near-clones with some extra flash but much worse 
integration.  I agree with you as to the reasons for this, though I'd 
add that I think that Windows has become so dominant that people now 
find it difficult to imagine anything really different. So, no, Windows 
is not very good as a desktop OS: it's just better than the rest.

The real tragedy here, of course, is that a huge amount of effort has 
been wasted on various platforms to provide them with a swishy 
windows-knock-off GUI.  I think I understand why that is - the 
developers of these platforms want to use them on their desktops.  But 
they are almost the *only* people who will ever want to do that: 
Windows Has Won on the desktop, and people who don't like it will use 
Macs and Ubuntu, in that order.  So all that effort is wasted (though, 
of course, people are free to spend their time how they like, so long 
as it is their time).

quoting Tim Bradshaw (Sat, 27 Jun 2009 16:45:06 +0100):
> So, no, Windows is not very good as a desktop OS: it's just better
> than the rest.

Without providing any proof of my statements I disagree with you. Not
only is windows -not- a better desktop OS, KDE/Gnome are even far
better.

> Windows Has Won on the desktop, and people who don't like it will use
> Macs and Ubuntu, in that order. So all that effort is wasted (though,
> of course, people are free to spend their time how they like, so long
> as it is their time).

Profecy is something from ancient books. Profecies have a tendency not
to come true. Windows is loosing the desktop quite rapidly at the
moment. Who can say what will be the situation in 2 years from now.
I see you can; well.. I can't :-/

-- 
Dick Hoogendijk -- PGP/GnuPG key: 01D2433D
+http://nagual.nl/ | SunOS 10u7 05/09 ZFS+

On 2009-06-27 17:50:54 +0100, Dick Hoogendijk <dick@nagual.nl> said:

>  Windows is loosing the desktop quite rapidly at the
> moment.

That depends on what you mean by "rapidly". OS X has done it a fair 
amount of damage I suspect, but it's unlikely to ever be dominant  (I'm 
an OS X user, BTW)  The rest seems to sit at around 1% or so (maybe 2 
if you are optimistic).

Tim Bradshaw wrote:
> On 2009-06-27 17:50:54 +0100, Dick Hoogendijk <dick@nagual.nl> said:
> 
>>  Windows is loosing the desktop quite rapidly at the
>> moment.
> 
> That depends on what you mean by "rapidly". OS X has done it a fair 
> amount of damage I suspect, but it's unlikely to ever be dominant  (I'm 
> an OS X user, BTW)  The rest seems to sit at around 1% or so (maybe 2 if 
> you are optimistic).
> 
IE, Firefox?  If nothing else, serious competition drives improvement.

-- 
Ian Collins

On 2009-06-28 00:55:40 +0100, Ian Collins <ian-news@hotmail.com> said:

> IE, Firefox?  If nothing else, serious competition drives improvement.

Of course.  But I think this example kind of makes my point: Firefox 
was not just some IE knock-off when it came out, it was actually 
radically different and better in many ways.  That's not true for the 
default desktops which ship with any of the Unix/Linux desktop OSs 
unfortunately, apart from OS X.