I just installed IE8 to check out a user report of a problem with this 
triumph of Lisp, Cells, qooxdoo, and jsMath: http://teamalgebra.com

Installing IE8 over IE7 took about twenty minutes. I think a complete 
Ubuntu install is about half that....

They've run up the white flag, Microsoft, haven't they? Not really 
trying any more, are they? Sad to see, really, not that I hold any 
fondness for the monstrosity. It's just that old phenomenon of sympathy 
setting in when the mighty beast is seen to be breathing its last.

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 2010-08-18 11:12:53 +0100, Kenneth Tilton said:

> They've run up the white flag, Microsoft, haven't they?

I think it's what happens to monopolists: they get big by being a 
monopolist, and because they have no competition they get slow and 
bureaucratic.  Then the world changes, they can't keep up, and they 
either fade away or go through a trauma and reinvent themselves (the 
latter is what happened to IBM).

Of course the change in the world is not Linux desktops/laptopt (or any 
other dekstops/laptops), it's the irrelevence of that whole notion.  
Even at my advanced age and with an unfashionable phone I find I do an 
increasing amount via it because it's just there.  People a generation 
younger than me are obvously much further down the path away from 
desktops.

On 2010-08-18 06:51:43 -0400, Tim Bradshaw said:

> Of course the change in the world is not Linux desktops/laptopt (or any 
> other dekstops/laptops), it's the irrelevence of that whole notion.  
> Even at my advanced age and with an unfashionable phone I find I do an 
> increasing amount via it because it's just there.  People a generation 
> younger than me are obvously much further down the path away from 
> desktops.

Which makes one think that the future of lisp is HTML5 web apps that 
run on mobile browsers, since one can't be assured that a lisp will be 
available for native development[1] on all mobile platforms (though 
kawa is available on Android).

Kenny, have you tested your algebra tutor on mobile safari?

[1] yes, I'm referring to Apple's iOS language restrictions here.

warmest regards,

Ralph

-- 
Raffael Cavallaro

On 18 Aug, 11:12, Kenneth Tilton <kentil...@gmail.com> wrote:
> I just installed IE8 to check out a user report of a problem with this
> triumph of Lisp, Cells, qooxdoo, and jsMath:http://teamalgebra.com
>
> Installing IE8 over IE7 took about twenty minutes. I think a complete
> Ubuntu install is about half that....

can my mum use Ubuntu?

> They've run up the white flag, Microsoft, haven't they? Not really
> trying any more, are they? Sad to see, really, not that I hold any
> fondness for the monstrosity. It's just that old phenomenon of sympathy
> setting in when the mighty beast is seen to be breathing its last.

On 2010-08-18 14:26:52 +0100, Nick Keighley said:

> can my mum use Ubuntu?

Probably, it seems to be pretty easy to drive.

On 2010-08-18 09:26:52 -0400, Nick Keighley said:

> can my mum use Ubuntu?

Can she (safely) use Windows?

warmest regards,

Ralph

-- 
Raffael Cavallaro

 KT> Installing IE8 over IE7 took about twenty minutes. I think a complete
 KT> Ubuntu install is about half that....

Installing OS a a new drive and updating an application on existing OS is 
not the same thing. (Apples to oranges...)

When installing new OS it can write mostly linearly, up to 100 MB/sec on HDD 
drives.

Most likely filesystem of your Windows installation is quite fragmented. 
Additionally, it might need to update files scattered over HDD.
As a result it needs to write to many places on HDD. But HDD's head has very 
limited speed, it needs something like 10 ms to move from one place to 
another, thus it is limited to doing like 100 operations per second.
Random write performance is poor for HDDs -- something like 1 MB/sec.

Additionally, IE installer writes lots of shit to registry, and registry is 
kind of DB which provides some durability/crash protection, so it generates 
lots of random IO too.
(This registry thing has positive sides too -- e.g. applicaton update can 
easily update one key.  It is often a problem for Linux package updates.)

Could Microsoft optimize this? Yes, I think they could use transactional 
mode of FS and registry operations (available starting from Vista) and 
optimize filesystem/registry implementation to minimize HDD seeks.

But I don't think it is a good idea to do it now when days of HDDs are 
counted and SSDs (which do not have problems with random writes) are going 
to replace them as system drives, at least. HDDs might still be useful for 
multimedia storage, but for applications demanding random I/O SSDs are far 
superior.

On 2010-08-18 15:37:48 +0100, Captain Obvious said:

> Could Microsoft optimize this? Yes, I think they could use 
> transactional mode of FS and registry operations (available starting 
> from Vista) and optimize filesystem/registry implementation to minimize 
> HDD seeks.

I think that makes Kenny's point.  They could indeed have optimised 
this, and done so a long time ago.  Other people with tiny fractions of 
their budget *have* fixed this kind of filesystem-performance issue.  
But because MS are actually a dead company now, they end up spending 
some fantastic amount on n-different fixes none of which work.

On 8/18/2010 8:43 AM, Raffael Cavallaro wrote:
> On 2010-08-18 06:51:43 -0400, Tim Bradshaw said:
>
>> Of course the change in the world is not Linux desktops/laptopt (or
>> any other dekstops/laptops), it's the irrelevence of that whole
>> notion. Even at my advanced age and with an unfashionable phone I find
>> I do an increasing amount via it because it's just there. People a
>> generation younger than me are obvously much further down the path
>> away from desktops.
>
> Which makes one think that the future of lisp is HTML5 web apps that run
> on mobile browsers, since one can't be assured that a lisp will be
> available for native development[1] on all mobile platforms (though kawa
> is available on Android).
>
> Kenny, have you tested your algebra tutor on mobile safari?

I just have a BlackBerry. Last I looked it was loading jsMath. I am not 
optimistic. :)

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 8/18/2010 9:26 AM, Nick Keighley wrote:
> On 18 Aug, 11:12, Kenneth Tilton<kentil...@gmail.com>  wrote:
>> I just installed IE8 to check out a user report of a problem with this
>> triumph of Lisp, Cells, qooxdoo, and jsMath:http://teamalgebra.com
>>
>> Installing IE8 over IE7 took about twenty minutes. I think a complete
>> Ubuntu install is about half that....
>
> can my mum use Ubuntu?

I think so, and indeed I think she would have a better shot with Ubuntu 
than Windows. This is coming from a Linux hater who has had to work 
intensely with Linux lately and am thrilled I did not have to mess with 
the usual Linux maintenance nonsense.

kt

>
>> They've run up the white flag, Microsoft, haven't they? Not really
>> trying any more, are they? Sad to see, really, not that I hold any
>> fondness for the monstrosity. It's just that old phenomenon of sympathy
>> setting in when the mighty beast is seen to be breathing its last.
>


-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 8/18/2010 10:37 AM, Captain Obvious wrote:
> KT> Installing IE8 over IE7 took about twenty minutes. I think a complete
> KT> Ubuntu install is about half that....
>
> Installing OS a a new drive and updating an application on existing OS
> is not the same thing. (Apples to oranges...)
>
> When installing new OS it can write mostly linearly, up to 100 MB/sec on
> HDD drives.
>
> Most likely filesystem of your Windows installation is quite fragmented.
> Additionally, it might need to update files scattered over HDD.
> As a result it needs to write to many places on HDD. But HDD's head has
> very limited speed, it needs something like 10 ms to move from one place
> to another, thus it is limited to doing like 100 operations per second.
> Random write performance is poor for HDDs -- something like 1 MB/sec.
>
> Additionally, IE installer writes lots of shit to registry, and registry
> is kind of DB which provides some durability/crash protection, so it
> generates lots of random IO too.
> (This registry thing has positive sides too -- e.g. applicaton update
> can easily update one key. It is often a problem for Linux package
> updates.)
>
> Could Microsoft optimize this? Yes, I think they could use transactional
> mode of FS and registry operations (available starting from Vista) and
> optimize filesystem/registry implementation to minimize HDD seeks.
>
> But I don't think it is a good idea to do it now when days of HDDs are
> counted and SSDs (which do not have problems with random writes) are
> going to replace them as system drives, at least. HDDs might still be
> useful for multimedia storage, but for applications demanding random I/O
> SSDs are far superior.
>
>

Please, if I am kind enough not to comment on your moronic posts, then 
you can return the favor by not commenting on mine.

thx, kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 8/18/2010 12:00 PM, Tim Bradshaw wrote:

> budget *have* fixed this kind of filesystem-performance issue. But
> because MS are actually a dead company now, they end up spending some
> fantastic amount on n-different fixes none of which work.

....and eventually running up the white flag on educational software:

   http://www.microsoft.com/learningspace/encarta_eol.aspx

There goes my exit strategy!

kenny


-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

Nick Keighley <nick_keighley_nospam@hotmail.com> writes:

> On 18 Aug, 11:12, Kenneth Tilton <kentil...@gmail.com> wrote:
>> I just installed IE8 to check out a user report of a problem with this
>> triumph of Lisp, Cells, qooxdoo, and jsMath:http://teamalgebra.com
>>
>> Installing IE8 over IE7 took about twenty minutes. I think a complete
>> Ubuntu install is about half that....
>
> can my mum use Ubuntu?
>

Not sure about yours, but mine at 76 found it fine. In fact, it was a
lot easier getting her using Ubuntu than it was getting her to use a VCR
in the early 90s (and she didn't even have to set the clock!). Even
better, I don't have to spend a weekend each month fixing things,
ensuring the AV is up to date and cleaning out spyware.

Now, if I can just convince her to stop turning off her mobile phone "to
save the battery" .....

-- 
theophilusx (at) rapttech dot com dot au

 TB> I think that makes Kenny's point.  They could indeed have optimised
 TB> this, and done so a long time ago.  Other people with tiny fractions of
 TB> their budget *have* fixed this kind of filesystem-performance issue.

Like who?

 TB> But because MS are actually a dead company now, they end up spending
 TB> some fantastic amount on n-different fixes none of which work.

Actually there are many significant improvements in Vista and Win7. You just 
did not care to learn more about that.

 KT> Please, if I am kind enough not to comment on your moronic posts, then
 KT> you can return the favor by not commenting on mine.

Unlike you, I have real experience optimizing applications  (all kinds of 
thereof) so I know what I'm talking about.
But I'm sorry if you have problems spewing FUD when confronted with facts. 

On Wed, 18 Aug 2010 17:37:48 +0300, "Captain Obvious"
<udodenko@users.sourceforge.net> wrote:

>Could Microsoft optimize this? Yes, I think they could use transactional 
>mode of FS and registry operations (available starting from Vista) and 
>optimize filesystem/registry implementation to minimize HDD seeks.
>
>But I don't think it is a good idea to do it now when days of HDDs are 
>counted and SSDs (which do not have problems with random writes) are going 
>to replace them as system drives, at least. HDDs might still be useful for 
>multimedia storage, but for applications demanding random I/O SSDs are far 
>superior.

SSDs are *not* going to generally replace HDDs anytime soon ... they
are great for long term storage, but their wear issues swiftly create
problems when they are used for often changing and/or temporary files
or for OS swapping.  

Windows Vista and 7 allow you to use a SS drive for the swap file - on
a busy server it burns out the drive very quickly - sometimes in just
a few months.

Wear-leveling strategies mitigate the problem somewhat, but SSD bits
can only be written to about 10,000 times before they fail.  HDD bits
can be written to millions of times before they fail.  Mechanical
problems generally kill HDDs long before their recording surfaces
become unreliable.

YMMV but I believe the circuitry needs to become much more durable
before SSDs become a real contender to replace HDDs.

George

 GN> SSDs are *not* going to generally replace HDDs anytime soon ... they
 GN> are great for long term storage,

Here's a quote from Intel datasheet for their 34nm MLC (cheaper) SSDs:
"The drive will have a minimum of 5 years of useful life under typical 
client workloads with up to 20 GB of host writes per day."

I think that is kind of enough for sane applications.

 GN> but their wear issues swiftly create
 GN> problems when they are used for often changing and/or temporary files
 GN> or for OS swapping.

These are not sane applications.

If OS is constantly swapping, maybe one needs to buy more RAM -- it is 
pretty cheap nowadays.
Swap is for emergencies, which are (and should be) rare, by definition.

And there is not need to write temporary files on disk -- Windows API has 
options to mark files as temporary, and then it doesn't write files to disk 
while it has enough RAM.
On Linux you can use tmpfs for /tmp and other directories with temporary 
files, many distributions do that.

So I think if some app requires shitloads of writes, it is better to fix 
this app in one way or another.

 GN> Windows Vista and 7 allow you to use a SS drive for the swap file - on
 GN> a busy server it burns out the drive very quickly - sometimes in just
 GN> a few months.

Your busy server powered by Windows Vista is constantly swapping?
You're doing it wrong, then. On so many levels...

For more demanding customers Intel is offering SLC drives:

-----
Write endurance
32 GB drive supports 1 petabyte of lifetime random writes and 64 GB drive 
supports 2
petabyte of lifetime random writes.
----

So you can overwrite it fully 8 times a day for 10 years. That's enough, I 
think.


 GN>   HDD bits can be written to millions of times before they fail. 
Mechanical
 GN> problems generally kill HDDs long before their recording surfaces
 GN> become unreliable.

There is anecdoting evidence that mainstream HDDs die after a month or so of 
very heavy use (e.g. in hands of Gentoo users who rebuild whole system each 
day).
I personally do not care what fails -- recording surface or mechanics.

At least SSD will tell you: "I cannot write anymore", while HDDs (typically) 
will die with all your data.

 GN> YMMV but I believe the circuitry needs to become much more durable
 GN> before SSDs become a real contender to replace HDDs.

I think they are already durable enough, according to specs from Intel, at 
least.
If anything needs to be done it is making OSs/filesystem/applications less 
dumb -- not writing to disk when it is not necessary. 

On 2010-08-19 08:03:00 +0100, Captain Obvious said:

> Like who?

I'm not really up-to-date on filesystem work outside a very small area. 
 But Sun's ZFS for instance is a good example of how you might deal 
with this sort of thing: what it does is essentially have a write-once 
approach, so as (the blocks of) files are modified they are in fact 
copied, which has a lot of good results for performance.   In the 
typical Sun way it has one very *bad* result, which is that if you want 
to verifiably erase a file you basically can't, because the normal 
"open it, write a lot of random junk repeatedly, delete file" does not 
work.  That's quite a bad problem given Sun's market.  It's interesting 
that the ZFS approach is quite similar in some ways to what copying GCs 
do.

But ZFS is far from the only example, and probably not even the best.  
Even boring old FFS seems to be able to do pretty well nowadays, and 
there are probably a bunch of Linux filesystems which do better.  The 
dedicated filesystem people - NetApp and Veritas among others - also do 
interesting things, though I am not up to date on them,
> 
> Actually there are many significant improvements in Vista and Win7. You 
> just did not care to learn more about that.

You're seem to think I'm some kind kind of anti-windows loony, which 
would be a mistake 15 or 20 years ago, perhaps).  I'm quite sure that 
there are lots of good things in 7 at least - Vista seemed to have a 
sufficiency  of problems that I'd be more hesitant about it.  But how 
much did they spend for these things compared to their competition 
(Apple, not Linux)?  And they're not even *in* the important markets as 
best I can tell.

In fact, the interesting thing isn't that MS are becoming irrelevant, 
it's what went wrong at Nokia, who seemed like the obvious next people 
in about 2001.

> KT> Please, if I am kind enough not to comment on your moronic posts, then
> KT> you can return the favor by not commenting on mine.

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

 KT>>> Please, if I am kind enough not to comment on your moronic posts,
 KT>>> then you can return the favor by not commenting on mine.

Well, you DO comment, don't you?

It is another thing that your comments do not contain anything interesting, 
only insults.

Sure, it is easier to insult people when you do not have anything to say on 
the matter.

Can you please find your balls and reply as a man?
If you think that anything I wrote is "moronic" then show it up -- or STFU.

Got it, bitch?

Just add me to killfile if you do not want to comment.

I see no reason to not comment on your posts -- if I have something to say I 
will. 

On 8/19/2010 10:31 AM, Captain Obvious wrote:
> Got it, bitch?

Charming.

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

 ??>> Like who?

 TB> I'm not really up-to-date on filesystem work outside a very small area.

I'm not either, but as I understand all filesystems having "cool features" 
are still experimental.
And when they'll get mainstream these cool features won't be relevant 
anymore -- people will switch to SSDs.

 TB> But Sun's ZFS for instance is a good example of how you might deal
 TB> with this sort of thing: what it does is essentially have a write-once
 TB> approach, so as (the blocks of) files are modified they are in fact
 TB> copied, which has a lot of good results for performance.   In the
 TB> typical Sun way it has one very *bad* result,

I think it isn't the only bad result -- decisions like this are always 
trade-offs.

I would expect read performance to be worse because file blocks will get 
scattered all over the place, so you'll get random I/O where you got 
continuous with tradiotional file systems.
I'm lazy to find better references, but Wikipedia says:
> ZFS has no defragmentation utility. Usage of COW with often changed files 
> leads to high fragmentation.

Same thing about WAFL:

> As the name suggests Write Anywhere File Layout automatically fragments 
> data using temporal locality to write metadata alongside user data. This 
> fragmentation does not adversely affect files that are sequentially 
> written to or randomly read from, but does affect sequential read after 
> random write. Data ONTAP has the reallocate command as of 7G to perform 
> scheduled and manual defragmentation. Prior to 7G, the wafl scan 
> reallocate command would need to be invoked from an advanced privilege 
> level and could not be scheduled.

So it is a common problem.

Then storage overhead might be higher -- as you say, old pieces are still 
hanging there.

Then there is fragility.
Maybe NTFS isn't the best fs w.r.t file recovery, but it isn't worst 
either -- it has relatively simple design and so there are both whole-disk 
and file recovery products out there.
(Not to mention the fact that Windows' chkdisk can recover some errors 
caused by hardware problems.)

As for ZFS... I didn't try it myself, but I've heard lots of "I've lost all 
my files on ZFS and I have no idea how to recover them" stories.
It is anecdotal evidence, but I think it says something.

Then I've heard that ZFS is picky about memory -- it is not adviced to use 
ZFS on 32-bit machines and on machines with less than 512 MB of RAM; and you 
need to tune memory usage manually for good results.
OTOH NTFS on Win XP can work with any memory size starting from 128 MB.

NTFS is not the best filesystem but it is good as a mainstream filesystem --  
not bad performance, decent features (ACLs, alternative data streams, links, 
plugins (reparse points), transaction support in new versions...), 
relatively good recoverability and crash-resilence. So it's no wonder that 
Mircosoft doesn't replace it with some radical new design.

 ??>> Actually there are many significant improvements in Vista and Win7.
 ??>> You just did not care to learn more about that.

 TB> You're seem to think I'm some kind kind of anti-windows loony, which
 TB> would be a mistake 15 or 20 years ago, perhaps).

No, I've just mentioned a fact -- they've got lot of interesting stuff in 
Vista, so your statement about Microsoft only fixing errors is not true.
Even if you don't like it won't hurt to skim an article:

http://technet.microsoft.com/en-us/magazine/2007.02.vistakernel.aspx
http://technet.microsoft.com/en-us/magazine/2007.03.vistakernel.aspx

I particularly liked memory priorities/SuperFetch -- finally cache won't 
compete with application data.


 TB>   I'm quite sure that there are lots of good things in 7 at least -
 TB> Vista seemed to have a sufficiency  of problems that I'd be more
 TB> hesitant about it.

OS might have cool technologies and problems at same time.

 TB> But how much did they spend for these things compared to their
 TB> competition (Apple, not Linux)?

I have no idea.
To start with, I don't think Mac OS X has all features which Windows Vista/7 
has.

If you're going to compare total spending on development, that would be 
apples-to-oranges. (Or apples to windows?)

Windows has to run on thousands of different hardware configurations, 
support thousands of old devices and programs.
Compatibility is Microsoft's bread and butter. People say a lot of shit 
about this, but they do in fact spend LOTS of resources on compatibility and 
it pays -- it is a familiar OS with lots of familiar programs -- and it is 
exactly what users need.

 TB>   And they're not even *in* the important markets as best I can tell.
 TB> In fact, the interesting thing isn't that MS are becoming irrelevant,
 TB> it's what went wrong at Nokia, who seemed like the obvious next people
 TB> in about 2001.

I bet there will always be a need for desktop and server operating systems 
and people will demand best ones there.

You can browse the web, send mails and so on from pretty much any device.

But then you'll go to use desktop for things mobile device cannot do. 

On 2010-08-19 16:25:50 +0100, Captain Obvious said:

> I'm not either, but as I understand all filesystems having "cool 
> features" are still experimental.
> And when they'll get mainstream these cool features won't be relevant 
> anymore -- people will switch to SSDs.

I don't have the energy to get into an argument about this.  NTFS was 
deficient when designed (defragmenting, WTF?), and is an antique now.  
Modern filesystems (where "modern" includes many designed before NTFS) 
do much better.  SSDs are not the panacea you think they are.

But this is all beside the point: if it takes 20 minutes to upgrade 
from IE7 to IE8 on modern hardware, that's just a joke, as Kenny said, 
and there is no excuse which can possibly justify it.

On 8/19/2010 12:16 PM, Tim Bradshaw wrote:
> On 2010-08-19 16:25:50 +0100, Captain Obvious said:
>
>> I'm not either, but as I understand all filesystems having "cool
>> features" are still experimental.
>> And when they'll get mainstream these cool features won't be relevant
>> anymore -- people will switch to SSDs.
>
> I don't have the energy to get into an argument about this. NTFS was
> deficient when designed (defragmenting, WTF?), and is an antique now.
> Modern filesystems (where "modern" includes many designed before NTFS)
> do much better. SSDs are not the panacea you think they are.
>
> But this is all beside the point: if it takes 20 minutes to upgrade from
> IE7 to IE8 on modern hardware, that's just a joke, as Kenny said, and
> there is no excuse which can possibly justify it.
>

I can think of two much more interesting (than hardware) questions:

(1) anyone out there still at IE7 who (a) dares to upgrade to IE8 and 
(b) can see if they have the same experience?

(2) simple point of information: does anyone know wtf? was being done 
during the upgrade from IE7 to IE8? Mind you I'll still be pissing on MS 
after getting this information no matter what the answer, I'll just be 
doing it with more specificity.

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

Tim Bradshaw <tfb@tfeb.org> writes:

> On 2010-08-19 16:25:50 +0100, Captain Obvious said:
>
>> I'm not either, but as I understand all filesystems having "cool features"
>> are still experimental.
>> And when they'll get mainstream these cool features won't be relevant anymore
>> -- people will switch to SSDs.
>
> I don't have the energy to get into an argument about this.  NTFS was deficient
> when designed (defragmenting, WTF?), and is an antique now.  Modern filesystems
> (where "modern" includes many designed before NTFS) do much better.  SSDs are
> not the panacea you think they are.
>
> But this is all beside the point: if it takes 20 minutes to upgrade from IE7 to
> IE8 on modern hardware, that's just a joke, as Kenny said, and there is no
> excuse which can possibly justify it.
>

Thank god someone with the ability to recognise this point. All this
bullshit about how Windows (or anyone else) has improved and stupid
irrelevant points about how they could do it better and how the next
generation of hardware will make this probolem go away or that it is
possible to make it work better if you have spent 20 years tweaking
technology and have loads of knowledge and experience is irrelevant. 

The bottom line is that taking 20 minutes to install a web browser in
this day an age is just ridiculous. End of story.

Tim

-- 
tcross (at) rapttech dot com dot au

On 8/19/2010 3:41 PM, Tim X wrote:
>
> The bottom line is that taking 20 minutes to install a web browser in
> this day an age is just ridiculous. End of story.
I think the time is a good indication that something more than the 
'browser' is getting updated ;)

I use FF,IE (don't ask what version, I don't even bother), Chrome
(and most likely Safari/webkit too as a result of the window itunes) for 
various things on Win 7

Don't ask me why. Don't have the energy to writeup
-Antony

On 8/19/2010 6:41 PM, Tim X wrote:
> Tim Bradshaw<tfb@tfeb.org>  writes:
>
>> On 2010-08-19 16:25:50 +0100, Captain Obvious said:
>>
>>> I'm not either, but as I understand all filesystems having "cool features"
>>> are still experimental.
>>> And when they'll get mainstream these cool features won't be relevant anymore
>>> -- people will switch to SSDs.
>>
>> I don't have the energy to get into an argument about this.  NTFS was deficient
>> when designed (defragmenting, WTF?), and is an antique now.  Modern filesystems
>> (where "modern" includes many designed before NTFS) do much better.  SSDs are
>> not the panacea you think they are.
>>
>> But this is all beside the point: if it takes 20 minutes to upgrade from IE7 to
>> IE8 on modern hardware, that's just a joke, as Kenny said, and there is no
>> excuse which can possibly justify it.
>>
>
> Thank god someone with the ability to recognise this point. All this
> bullshit about how Windows (or anyone else) has improved and stupid
> irrelevant points about how they could do it better and how the next
> generation of hardware will make this probolem go away or that it is
> possible to make it work better if you have spent 20 years tweaking
> technology and have loads of knowledge and experience is irrelevant.
>
> The bottom line is that taking 20 minutes to install a web browser in
> this day an age is just ridiculous. End of story.

I heard somewhere that if you take a sick person to a surgeon, the 
surgeon will always find a way to cut them to cure them.

Seems to me a Windows tuning specialist had a knee-jerk reaction to my 
problem report and has offered input as welcome as that of a specialist 
in custom tailoring who sees a conference on obesity and dashes in to help.

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 2010-08-19 23:41:54 +0100, Tim X said:

> The bottom line is that taking 20 minutes to install a web browser in
> this day an age is just ridiculous. End of story.

Yes, that was what I was trying to say before I got dragged off into 
arguing about filesystems.  Technical excuses and reasons don't matter: 
that long is too long, and that's all there is to it.

 ??>> The bottom line is that taking 20 minutes to install a web browser in
 ??>> this day an age is just ridiculous. End of story.

 TB> Yes, that was what I was trying to say before I got dragged off into
 TB> arguing about filesystems.  Technical excuses and reasons don't matter:
 TB> that long is too long, and that's all there is to it.

Here's a message about Solaris 10:

http://www.unix.com/solaris/125693-solaris-10-patching-14-hours.html

"system took 14 hours to install ~320 patches!!."

This, obviously, means that Solaris 10 is slow as molasses.

The bottom line is that taking 14 hours to install patches in this day an 
age is just ridiculous. End of story.

They've run up the white flag, Sun Microsystems, haven't they? Not really
trying any more, are they? Sad to see, really, not that I hold any
fondness for the monstrosity. It's just that old phenomenon of sympathy
setting in when the mighty beast is seen to be breathing its last.

....

I could go on bitching about it in same style as you guys.
The bottom line is that ALL operating systems have some deficiencies.
Taking some isolated, anecdotal case as a cornerstone of your reasoning 
won't produce a meaningful discussion.

I'm not a "Windows tuning specialist", by the way.
I've been working for years developing software for Windows in C++. 
(Different things, from 3D graphics and games to backup software.)
But for the last 4 years I'm mostly dealing with Lisp web servers running on 
Linux-powered servers.
So I know a bit of both, and ad hominem attacks were totally pointless.

Software performance evaluation/optimization is one of my main areas of 
interest -- it is just interesting to me.
So I was doing it with a lot of things -- 3D graphics, disk I/O, databases, 
memory caches, virtual memory/swapping, algorithms, machine code.
So I can talk about this stuff. I don't want to brag, but people who know me 
often ask for performance-related advices.

But of course it is easier to ignore all details and just jump to conclusion 
"Microsoft sucks/is dead" based on some vague non-investigated anecdotal 
case.
Ignorance is strength. 

On 2010-08-20 11:20:18 +0100, Captain Obvious said:

> This, obviously, means that Solaris 10 is slow as molasses.
> 
> The bottom line is that taking 14 hours to install patches in this day 
> an age is just ridiculous. End of story.
> 
> They've run up the white flag, Sun Microsystems, haven't they? Not really
> trying any more, are they? Sad to see, really, not that I hold any
> fondness for the monstrosity. It's just that old phenomenon of sympathy
> setting in when the mighty beast is seen to be breathing its last.

Absolutely.  Patching on Solaris is a sick joke (it's a big part of 
what I do for my living in fact, and I not only know the performance is 
terrible, I know *why*, and I've explained this to various people who 
are clearly in denial about it).  Boot times are also a bit 
embarrassing.  And I think we can probably draw the same conclusions 
about Solaris as we did about Microsoft (note a lot of the other 
factors are present as well: more than 5 years since FCS of 10, lots of 
other overengineered misdesign in the system).

> So I know a bit of both, and ad hominem attacks were totally pointless.

Do you think I made one?

I'm done with this now, it's got boring.

 ??>> They've run up the white flag, Sun Microsystems, haven't they? Not
 ??>> really trying any more, are they? Sad to see, really, not that I hold
 ??>> any fondness for the monstrosity. It's just that old phenomenon of
 ??>> sympathy setting in when the mighty beast is seen to be breathing its
 ??>> last.

 TB> Absolutely.

No, they did not. As I understand OpenSolaris had its own package system 
(IPS), writen from scratch and without these problems.
So I'd say they were still trying to fix it and improve it. If we're 
discussing long term perspective rather than a current state it matters.
(It's another thing that OpenSolaris is now officially dead, as well as Sun 
Micro in general...)

Discussing current state is like beating a dead horse. Obviously all 
operating systems and all complex software have some problems.
It is a well known fact and picking individual anecdotal cases doesn't make 
a lot of sense.

 TB> in denial about it).  Boot times are also a bit embarrassing.  And I
 TB> think we can probably draw the same conclusions about Solaris as we did
 TB> about Microsoft

I'm afraid you can pick pretty much any software vendor and paint it grim.

 ??>> So I know a bit of both, and ad hominem attacks were totally
 ??>> pointless.

 TB> Do you think I made one?

No, Kenny and Tim X did, in some way.

But you've completely dismissed my arguments and reiterated your points 
which were already refuted.
That's almost as bad.

 TB> I'm done with this now, it's got boring.

It was boring from the start, I just wanted to stir up a bit this 
anti-microsoft circlejerk. Didn't work very well... 

On 2010-08-20 12:33:18 +0100, Captain Obvious said:

> No, they did not. As I understand OpenSolaris had its own package 
> system (IPS), writen from scratch and without these problems.
> So I'd say they were still trying to fix it and improve it. If we're 
> discussing long term perspective rather than a current state it matters.
> (It's another thing that OpenSolaris is now officially dead, as well as 
> Sun Micro in general...)
> 
> Discussing current state is like beating a dead horse. Obviously all 
> operating systems and all complex software have some problems.
> It is a well known fact and picking individual anecdotal cases doesn't 
> make a lot of sense.

The irony that I'm sitting here getting bored while waiting for some 
patch bundle to apply to   a Solaris machine makes a response 
irresitable.  I don't expect you to understand the following, but I 
enjoyed writing it.

So, let's see.

Do you think that the view that MS and Sun (under whatever ownership) 
are slowly dying is somehow contentious?  Because if you do I think you 
have really seriously not been keeping up over the last couple of years.

Now, Solaris.

Patching performance has been hurting customers for something like 15 
years.  You might think that an organisation which wasn't suffering 
from terminal sclerosis might have managed to fix the problem in that 
time, but no.

So, what actually is the performance problem, and how hard is it to 
fix?  Maybe its some intractable thing which really requires massive, 
slow redesign.

Well, it turns out that the problem is this: Solaris machines have a 
"contents" file which maps from filenames to packages (with some extra 
information like permissiosn, checksums etc).  It's a text file, with 
one line per file.  On a full install of 10 it's around 20MB.  Patches 
need to modify this file to update the entries for the files the 
modify.  A big patch (probably the biggest current patch) might modify 
6000 files.

Patches do this by *reading and writing the whole file for each line 
they change*.  Read that againmthen do the sums.  It's about 240GB of 
I/O, and that's for a single (albeit bad) patch.  Applying a patch 
bundle of one or two hundred patches can do over a terabyte of I/O, to 
update a system which will be well under 10GB in size.

And that is not all.  Systems with whole-root zones will do this for 
every zone.  And people have systems with upwards of 10 whole-root 
zones.  I probably don't need to tell you that those systems just don't 
get patched, because they essentially can't be.

So you'd think they'd want to fix this wouldn't you?  Well, once you 
get past the denial you hear two stories:

1. "We can't fix it, because we'd need to use a database to avoid this 
problem, and that would mean an incompatible change and we can't do 
that."

2. "IPS"

So what about (1): there are two easy answers to this.  Either use a 
database and run an extract from it at the end to generate the contents 
file, or (much better) use the existing contents file and update it by 
writing changes into a log and then rolling the log into the file in 
one fell swoop.  The second approach is really not hard to implement, 
neither approach is 15 years of hard to implement.

And for (2): when people answer questions about bugs like this you know 
they're trying to avoid the truth.  "We will write a new and different 
one" is almost never the right answer.  Let's imagine Solaris 11, with 
IPS, ships in 2012Q1.  People might start deploying it in critical 
applications in 2012Q3.  Maybe the existing Solaris 10 installed base 
might go away in 2015-2016 (that is probably optimistic).  That's at 
least 5 more years of pain.  By all means keep working on IPS (which is 
needed for other reasons), but just fix patchadd already.

So, maybe this bug will take 20 years to "fix" where "fix" means 
"ignore until we replace it with something else".  That's not a good 
indicator, at the very best.

That "something else" of course will turn out to be Linux.  And for 
someone with more years invested in Suns & Solaris than this bug will 
take to fix, that's a bitter truth.

On 8/20/2010 6:20 AM, Captain Obvious wrote:
> But of course it is easier to ignore all details and just jump to
> conclusion "Microsoft sucks/is dead" based on some vague
> non-investigated anecdotal case.

Yep! I have real work to do, you know:

    http://www.stuckonalgebra.com/index.html

A pure Lisp Web application with just enough JS glue to make a JS 
framework transparent, in this case qooxdoo:

    http://wiki.github.com/kennytilton/qooxlisp/

Die, Microsoft! Die!

Tim Bradshaw <tfb@tfeb.org> writes:

> On 2010-08-20 12:33:18 +0100, Captain Obvious said:
>
>> No, they did not. As I understand OpenSolaris had its own package
>> system (IPS), writen from scratch and without these problems.
>> So I'd say they were still trying to fix it and improve it. If we're
>> discussing long term perspective rather than a current state it
>> matters.
>> (It's another thing that OpenSolaris is now officially dead, as well
>> as Sun Micro in general...)
>>
>> Discussing current state is like beating a dead horse. Obviously all
>> operating systems and all complex software have some problems.
>> It is a well known fact and picking individual anecdotal cases
>> doesn't make a lot of sense.
>
> The irony that I'm sitting here getting bored while waiting for some
> patch bundle to apply to   a Solaris machine makes a response
> irresitable.  I don't expect you to understand the following, but I
> enjoyed writing it.
>
> So, let's see.
>
> Do you think that the view that MS and Sun (under whatever ownership)
> are slowly dying is somehow contentious?  Because if you do I think
> you have really seriously not been keeping up over the last couple of
> years.
>
> Now, Solaris.
>
> Patching performance has been hurting customers for something like 15
> years.  You might think that an organisation which wasn't suffering
> from terminal sclerosis might have managed to fix the problem in that
> time, but no.
>
> So, what actually is the performance problem, and how hard is it to
> fix?  Maybe its some intractable thing which really requires massive,
> slow redesign.
>
> Well, it turns out that the problem is this: Solaris machines have a
> "contents" file which maps from filenames to packages (with some extra
> information like permissiosn, checksums etc).  It's a text file, with
> one line per file.  On a full install of 10 it's around 20MB.  Patches
> need to modify this file to update the entries for the files the
> modify.  A big patch (probably the biggest current patch) might modify
> 6000 files.
>
> Patches do this by *reading and writing the whole file for each line
> they change*.  Read that againmthen do the sums.  It's about 240GB of
> I/O, and that's for a single (albeit bad) patch.  Applying a patch
> bundle of one or two hundred patches can do over a terabyte of I/O, to
> update a system which will be well under 10GB in size.

This is really silly.  They could easily append change lines, and
later have the administrator run an update procedure that would
rewrite the file with all the changes applied at once.

-- 
__Pascal Bourguignon__                     http://www.informatimago.com/

On Thu, 19 Aug 2010 13:27:21 +0300, "Captain Obvious"
<udodenko@users.sourceforge.net> wrote:

> GN> Windows Vista and 7 allow you to use a SS drive for the swap file - on
> GN> a busy server it burns out the drive very quickly - sometimes in just
> GN> a few months.
>
>Your busy server powered by Windows Vista is constantly swapping?
>You're doing it wrong, then. On so many levels...

I'm not going to be drawn into a debate on the merits of Windows as a
server.  People do it, so deal with it.  And I'm not going to debate
virtual servers either - they have their place but they are not the
be-all and end-all of server-side software.

In general there is no way to prevent Windows from swapping - it
doesn't matter how much RAM you have.  Windows uses relocatable DLLs
as opposed to position independent ones, so a busy system frequently
has many copies of the same DLLs loaded simultaneously.  DLLs that are
not mapped at their image address use the page file as backing store.
Programs and DLLs fortunate enough to be mapped at their image
addresses are mapped directly from their disk files.

Windows gets upset if you disable swapping entirely and the
auto-tuning large cache manager in the server versions essentially
does what it pleases to adapt to dynamically changing loads.  There is
a fair amount of manual tuning that can be done, but it is very
difficult to dial in a server running a mix of applications ... so it
is almost always left to the cache manager.

George

On Aug 18, 3:51=C2=A0am, Tim Bradshaw <t...@tfeb.org> wrote:
> On 2010-08-18 11:12:53 +0100, Kenneth Tilton said:
>
> > They've run up the white flag, Microsoft, haven't they?
>
> I think it's what happens to monopolists: they get big by being a
> monopolist, and because they have no competition they get slow and
> bureaucratic.

lol. what load of crap.

many tech geekers don't understand what the word monopoly means.

monopoly is preventing others from competition.
monopoly is not saturation of a market.

when tech geekers speaks of Microsoft monopoly, they mean saturation
of a market.
e.g. GNU is a monopoly, linux is a monopoly, google is a monopoly.

btw, what kenny is babbling about? i installed ie8 from ie7 in maybe 5
min about a year ago?

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On Aug 18, 9:20=A0am, Kenneth Tilton <kentil...@gmail.com> wrote:
> On 8/18/2010 9:26 AM, Nick Keighley wrote:
>
> > On 18 Aug, 11:12, Kenneth Tilton<kentil...@gmail.com> =A0wrote:
> >> I just installed IE8 to check out a user report of a problem with this
> >> triumph of Lisp, Cells, qooxdoo, and jsMath:http://teamalgebra.com
>
> >> Installing IE8 over IE7 took about twenty minutes. I think a complete
> >> Ubuntu install is about half that....
>
> > can my mum use Ubuntu?
>
> I think so, and indeed I think she would have a better shot with Ubuntu
> than Windows. This is coming from a Linux hater who has had to work
> intensely with Linux lately and am thrilled I did not have to mess with
> the usual Linux maintenance nonsense.
>
> kt

o rly?

where can i buy ubuntu? from slashdot?

 Xah

On 2010-08-20 16:50:14 +0100, Pascal J. Bourguignon said:

> This is really silly.  They could easily append change lines, and
> later have the administrator run an update procedure that would
> rewrite the file with all the changes applied at once.

Well, perhaps they couldn't do that because the file might well 
effectively be a published API, so you can't have extra things in it.  
But something very close to that is the 
write-the-changes-to-a-log-which-you-roll-in thing (this "log" can, of 
course, be a sed script or something like that - something quite 
primitive).

But yes, it is really silly.

On Aug 19, 2:18=C2=A0am, George Neuner <gneun...@comcast.net> wrote:
> On Wed, 18 Aug 2010 17:37:48 +0300, "Captain Obvious"
>
> <udode...@users.sourceforge.net> wrote:
> >Could Microsoft optimize this? Yes, I think they could use transactional
> >mode of FS and registry operations (available starting from Vista) and
> >optimize filesystem/registry implementation to minimize HDD seeks.
>
> >But I don't think it is a good idea to do it now when days of HDDs are
> >counted and SSDs (which do not have problems with random writes) are goi=
ng
> >to replace them as system drives, at least. HDDs might still be useful f=
or
> >multimedia storage, but for applications demanding random I/O SSDs are f=
ar
> >superior.

here's my take

=E2=80=A2 Solid-State Drive Comes Of Age!
  http://xahlee.org/comp/solid_state_drives.html

plain text version follows
--------------------------------------------------
Solid-State Drive Comes Of Age!

Xah Lee, 2010-03-17

Solid-state drive has come of age!

When i started using a computer, in 1990, the standard hard drive
capacity is 20 mega bytes. Its physical size is larger than a big
hand, weights like a small brick. It costs some 5 hundred dollars.

Today, a cheap $25 usb stick has 32 giga byte capacity. Its size is
about 1 finger, and weights like 2 pencils.

The capacity is 800 times larger, price is 20 times cheaper, physical
size is perhaps 30 times smaller, weight is perhaps 50 times smaller.
Access time and thru-put, is some 50 or more times faster.

Solid-state drive has come of age! So, looking in the near future,
hard drive is about to go extinct. Many small laptops already don't
use hard drives. They have solid-state drives, i.e. a storage device
that does not have mechanical moving parts. So, when would majority of
consumer laptops use solid state drives? I'd say in 4 years. Also,
optical disks, such as DVD, and recently Blu-ray Disc, doesn't seem to
have a bright long-term future. They just seem too unwieldy and slow.

Advantages Over Hard Disk

Solid-state drives (SSDs) has quite a lot significant advantages over
hard disk drives (HDDs). Here's a summary from Wikipedia:

Faster start-up because no spin-up is required.
Faster random access, because there's no read-write head, as in
turntable
No more file fragmentation problem.
Silent operation due to the lack of moving parts.
Low power consumption and little heat.
High mechanical reliability, because no moving parts.
Ability to endure extreme shock, high altitude, vibration and extremes
of temperature.
Lower weight and size.
Flash SSDs have twice the data density of HDDs.
Failures occur less frequently while writing/erasing data.
Wikipedia also gave a list of disadvantages, but they are not about
the inherent merits of the technology, but rather current state of
affairs. e.g. obviously, hard drives are still popular and economical,
especially for large capacity needs. I'd say, in 4 years, majority of
laptops will be using SSD, and perhaps in 10 years, data centers will
be using SSD for their multi exa bytes of storage needs. (kilo, mega,
giga, tera, peta, exa)

SRAM, DRAM, Flash Memory

The Wikipedia article on SSD mentions SRAM, DRAM, Flash Memory. So,
what are they? Here's a simple explanation.

Memory vs Disk

Broadly speaking, =E2=80=9CMemory=E2=80=9D and =E2=80=9CDisk/Drive=E2=80=9D=
 are all data storage
devices. Those that are fast, are typically called Memory, and
typically has smaller capacity. Those that are slow, usually with
large capacity, are called Disks or Drives.

Volatile vs Non-volatile

Storage tech can be volatile or non-volatile. Volatile are those needs
electricity to maintain their state. (i.e. they need batteries)
Otherwise, it's non-volatile. Paper, records, CD, DVD, hard disks, are
non-volatile. Non-volatile tech usually can have large capacity.

Flash Memory =E2=87=92 Non-volatile

Solid state storage devices, can use volatile or non-volatile techs.
The most common non-volatile tech, is called Flash memory. It is used
in USB disks.

SRAM and DRAM =E2=87=92 Volatile

Now, there are 2 types of commonly used volatile memory. Static random
access memory (SRAM) and Dynamic random access memory (DRAM). Both are
volatile. The difference is that DRAM requires constant electric
recharge, or refresh, for its memory to stay. Both SRAM and DRAM, need
batteries, but DRAM need constant refresh. A bit technical here.

Now, a Solid-state Drive can use non-volatile tech e.g. Flash Memory,
or it can use volatile tech e.g. SRAM or DRAM. But basically, when the
capacity needs to be large, it is non-volatile, e.g. Flash Memory.

Problem With Hard Drives

For a efficiency and elegance nerd, hard drive can't really be liked.

You have seen a phonograph, right? You know, a (plastic) disk that
spins, with a head/needle that reads the data. The heart of the hard
drive tech is like that. The spinning disk(s) is called platter, and
the needle is called the read/write head. Ultimately, such device
can't scale (grow).

To make it faster, ultimately it needs to spin faster, but spin rate
has physics limitations. To make the capacity larger, you have to
either increase the physical size of the disk, or increase the density
(like shrinking the grooves on a record). Both reaches physics
limitations quickly. Another way to make it faster is to have multiple
disks and multiple heads. Having multiple disks is mechanical
complexity, with physics restraints, grows in physical dimension fast.
Making it to use multiple heads per disk layer has many engineering
problems, best described as =E2=80=9Chead collision=E2=80=9D. Optical drive=
s, such as
CD, DVD, are all like this. They can't grow.

Today, SSD is in your cellphone, iPods.

Tech Advances

Jukeboxes vs iPod

speaking of iPod... that reminds me of jukeboxes, and the incredible
technological advance i lived thru, in the past 20 years.

Remember Jukebox? In the 1990, when you want to play a whole
collection of music, you use jukebox. It is sized like a refrigerator.
Costs thousands of dollars. Today, you have iPod. Sized smaller than
your hand, with capacity that's a thousand times more, not to mention
music fidelity.

Tech Solving The World's Problems?

Speaking of tech advances, i can feel many tech geekers feeling that
the world's problems is all gonna be solved. LOL.

Remember, that as tech grows, the human population also is growing
behind our back, and arguably has past some sustainable size for long.
You know, all the talk about climate change, green energy, often in a
urgent way. Yeah, tech advances has been great, but overall, in terms
of solving the human animal problem, of natural resources, food,
energy, it isn't exactly peaches and creams. (See: Population Under
The Firmament)

Also, tech today can actually provide food for everyone, but that's
not happening, and ain't like gonna be happening. You know, wars,
massacres, famines, happens in parts of the world every day, all over.
And, you seen homeless on the city street asking you for a quarter,
right? That is the human problem, a problem of human nature. This
human nature, human animal problem, is not likely to go away in a
million years.

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On 2010-08-20 18:22:16 +0100, Xah Lee said:

> monopoly is preventing others from competition.
> monopoly is not saturation of a market.

Surely even you could work out that one of these things leads to the other?

 XL> lol. what load of crap.

 XL> many tech geekers don't understand what the word monopoly means.

 XL> monopoly is preventing others from competition.
 XL> monopoly is not saturation of a market.

 XL> when tech geekers speaks of Microsoft monopoly, they mean saturation
 XL> of a market.

I generally agree with you, but when everybody looks for familiar (read: 
like Windows) operating systems and compatible (read: with Windows) 
software, you cannot deny that others are being prevented from competation 
to some extent.
Plus there is something called "vendor lock-in". 

On Aug 18, 6:51=A0am, Tim Bradshaw <t...@tfeb.org> wrote:
> People a generation younger than me are obvously much further
> down the path away from desktops.

I may not be Young People, but I am a Young Person, and I happen to
love my crappy old i686-based desktop.

On Aug 20, 12:41=C2=A0pm, Tim Bradshaw <t...@tfeb.org> wrote:
> On 2010-08-20 18:22:16 +0100, Xah Lee said:
>
> > monopoly is preventing others from competition.
> > monopoly is not saturation of a market.
>
> Surely even you could work out that one of these things leads to the othe=
r?

Dear Tim, get a hold of

Basic Economics, by Thomas Sowell.

http://xahlee.org/Periodic_dosage_dir/jdini/basic_economics.html

there's audio version and paperback. Buy it now!

there's also basic summary around the web and pirated versions, if one
must.

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On Aug 20, 12:42=C2=A0pm, "Captain Obvious"
<udode...@users.sourceforge.net> wrote:
> =C2=A0XL> lol. what load of crap.
>
> =C2=A0XL> many tech geekers don't understand what the word monopoly means=
..
>
> =C2=A0XL> monopoly is preventing others from competition.
> =C2=A0XL> monopoly is not saturation of a market.
>
> =C2=A0XL> when tech geekers speaks of Microsoft monopoly, they mean satur=
ation
> =C2=A0XL> of a market.
>
> I generally agree with you, but when everybody looks for familiar (read:
> like Windows) operating systems and compatible (read: with Windows)
> software, you cannot deny that others are being prevented from competatio=
n
> to some extent.
> Plus there is something called "vendor lock-in".

a bit explanation is in order.

When a company prevents others from entering the market by tricks =3D
evvvill!.  This is what most tech geekers are thinking.

There's not that much example of such in our capitalistic US going on,
to the degree that idioty body or tech geekers need to chat about,
maybe except US government.

When a company saturates a particular market, such as Microsoft's OS
for PC, google for search engine, that's not =3D evviiil. Because, any
joe can come in and sell their shit, and when the public decided joe's
better (e.g. linux), the so called monopoly topples.

quote: =C2=ABJudge Alex Kozinski of the 9th Circuit pointed out that the
key to monopoly is not market share =E2=80=94 even when it is 100 percent =
=E2=80=94
but the ability to keep others out. A company which cannot keep
competitors out is not a monopoly, no matter what percentage of the
market it may have at a given moment.=C2=BB

from =E3=80=88Basic Economics=E3=80=89, First Edition, by Thomas Sowell
ISBN: 0-465-08138-X
http://www.amazon.com/exec/obidos/ASIN/046508138X

Chapter 2: Big business and Government: Anti-Trust Laws, p.102

See also:

=E2=80=A2 Reading Notes on Basic Economics
  http://xahlee.org/Periodic_dosage_dir/jdini/basic_economics.html

=E2=80=A2 On Microsoft Hatred
  http://xahlee.org/UnixResource_dir/writ/mshatred155.html

=E2=80=A2 The Microsoft Hatred FAQ
  http://xahlee.org/UnixResource_dir/writ/mshatredfaq.html

run and buy the book now. Save humanity! as tech geekers might say,
every generation some idiots repeats shooting ourselves.

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On 2010-08-20, Xah Lee <xahlee@gmail.com> wrote:
> On Aug 18, 3:51�am, Tim Bradshaw <t...@tfeb.org> wrote:
>> On 2010-08-18 11:12:53 +0100, Kenneth Tilton said:
>>
>> > They've run up the white flag, Microsoft, haven't they?
>>
>> I think it's what happens to monopolists: they get big by being a
>> monopolist, and because they have no competition they get slow and
>> bureaucratic.
>
> lol. what load of crap.
>
> many tech geekers don't understand what the word monopoly means.
>
> monopoly is preventing others from competition.
> monopoly is not saturation of a market.
>
> when tech geekers speaks of Microsoft monopoly, they mean saturation
> of a market.
> e.g. GNU is a monopoly, linux is a monopoly, google is a monopoly.
>
> btw, what kenny is babbling about? i installed ie8 from ie7 in maybe 5
> min about a year ago?
>
>  Xah ? http://xahlee.org/ ?


I *do* know what a monopoly is, and Microsoft pretty much
definitely engaged in monopolistic practices once upon a time.

For those of you who aren't mathematical economists but still
literate, you may have notice that Microsoft has, in fact, been
found to have engaged in monopolistic behavior by an American
court and by the European Commission.

Cheers, B.

On 2010-08-20, Xah Lee <xahlee@gmail.com> wrote:
> On Aug 20, 12:41�pm, Tim Bradshaw <t...@tfeb.org> wrote:
>> On 2010-08-20 18:22:16 +0100, Xah Lee said:
>>
>> > monopoly is preventing others from competition.
>> > monopoly is not saturation of a market.
>>
>> Surely even you could work out that one of these things leads to the other?
>
> Dear Tim, get a hold of
>
> Basic Economics, by Thomas Sowell.
>
> http://xahlee.org/Periodic_dosage_dir/jdini/basic_economics.html
>
> there's audio version and paperback. Buy it now!
>
> there's also basic summary around the web and pirated versions, if one
> must.
>
>  Xah ? http://xahlee.org/ ?


Thomas Sowell is pretty much an intellectual lightweight
who skates by on the man-bites-dog shtick of being a
Black conservative.

Suggesting anyone learn economics from Sowell would be
pretty much the most stupid thing I've heard today, if
not this week.

In general the whole Chicago School is a crock of shit
based on vastly oversimplified models that have appeal
for the point-and-drool crowd of economic idiots.  If
you're paying attention to them (particularly after the
last couple of years), you are an idiot.

Cheers, B.

On Aug 20, 8:24=C2=A0pm, "Dr. Brian Leverich" <lever...@linkpendium.com>
wrote:
>
> > many tech geekers don't understand what the word monopoly means.
>
> > monopoly is preventing others from competition.
> > monopoly is not saturation of a market.
>
> > when tech geekers speaks of Microsoft monopoly, they mean saturation
> > of a market.
> > e.g. GNU is a monopoly, linux is a monopoly, google is a monopoly.
>
> > btw, what kenny is babbling about? i installed ie8 from ie7 in maybe 5
> > min about a year ago?
>
> > =C2=A0Xah ?http://xahlee.org/?
>
> I *do* know what a monopoly is, and Microsoft pretty much
> definitely engaged in monopolistic practices once upon a time.
>
> For those of you who aren't mathematical economists but still
> literate, you may have notice that Microsoft has, in fact, been
> found to have engaged in monopolistic behavior by an American
> court and by the European Commission.

according to Thomas Sowell, anti-trust began as a measure to protect
the consumers, but in most court cases, it is simply used by
competitors against the most successful.

in his first edition of the book published around late 1999 or so, it
didn't mention the Microsoft case. It 3rd edition or before, it
specifically does. LOL

from my personal understanding, it is true too, in particular with the
Microsoft case. All of you truely incompetent Sun Microsystems, Be,
AOL/Netscape, Linuxes, or fucking expensive Apple, can't beat Windows?
Let's sue Microsoft!

the gist of your argument in the above, is that because big government
said X, thus it must be true. Well, look at the case of American
genocide of the native Americans, or the German goverment's invasion
of europe. So, average people, at those times, would all say, yeah,
jews are idiots and must be exterminated, and Americans Indians are
barbarians and we must elimitate them, all books and newspapers said
so!

you can find other links to intellectuals on this too:

=E2=80=A2 The Microsoft Hatred FAQ
  http://xahlee.org/UnixResource_dir/writ/mshatredfaq.html

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On 2010-08-20, Xah Lee <xahlee@gmail.com> wrote:
> On Aug 20, 12:42�pm, "Captain Obvious"
><udode...@users.sourceforge.net> wrote:
>> �XL> lol. what load of crap.
>>
>> �XL> many tech geekers don't understand what the word monopoly means.
>>
>> �XL> monopoly is preventing others from competition.
>> �XL> monopoly is not saturation of a market.
>>
>> �XL> when tech geekers speaks of Microsoft monopoly, they mean saturation
>> �XL> of a market.
>>
>> I generally agree with you, but when everybody looks for familiar (read:
>> like Windows) operating systems and compatible (read: with Windows)
>> software, you cannot deny that others are being prevented from competation
>> to some extent.
>> Plus there is something called "vendor lock-in".
>
> a bit explanation is in order.
>
> When a company prevents others from entering the market by tricks =
> evvvill!.  This is what most tech geekers are thinking.
>
> There's not that much example of such in our capitalistic US going on,
> to the degree that idioty body or tech geekers need to chat about,
> maybe except US government.

Oh, puhleez.  That's a painfully stupid remark.

You think maybe there's a competitive market in petroleum, crude
or refined?  You think banking or insurance or hospitals or ...
are competitive markets?  Do you think the *jewelry* industry
operates in a competitive market?  ;)

Truly competitive markets are almost nonexistent.  There is a
huge economic incentive for firms to find ways to explicitly or
implicitly achieve monopoly rents.

This is one of the major reasons why the Chicago School is, in
fact, a crock of shit.  Silly Ec 101 markets largely do not
exist in the real world.

Cheers, B.

On Aug 20, 8:30=C2=A0pm, "Dr. Brian Leverich"
<lever...@askin-17.linkpendium.com> wrote:
> Thomas Sowell is pretty much an intellectual lightweight
> who skates by on the man-bites-dog shtick of being a
> Black conservative.
>
> Suggesting anyone learn economics from Sowell would be
> pretty much the most stupid thing I've heard today, if
> not this week.
>
> In general the whole Chicago School is a crock of shit
> based on vastly oversimplified models that have appeal
> for the point-and-drool crowd of economic idiots. =C2=A0If
> you're paying attention to them (particularly after the
> last couple of years), you are an idiot.

LOL. We have one =E2=80=9CDr=E2=80=9D goes in full newsgroup style, bravo.

i dunno what kinda dr u r, but judging from what u just wrote, u
probably r a nobody. i LOL.

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On 2010-08-21, Xah Lee <xahlee@gmail.com> wrote:
> On Aug 20, 8:30�pm, "Dr. Brian Leverich"
><lever...@askin-17.linkpendium.com> wrote:
>> Thomas Sowell is pretty much an intellectual lightweight
>> who skates by on the man-bites-dog shtick of being a
>> Black conservative.
>>
>> Suggesting anyone learn economics from Sowell would be
>> pretty much the most stupid thing I've heard today, if
>> not this week.
>>
>> In general the whole Chicago School is a crock of shit
>> based on vastly oversimplified models that have appeal
>> for the point-and-drool crowd of economic idiots. �If
>> you're paying attention to them (particularly after the
>> last couple of years), you are an idiot.
>
> LOL. We have one ?Dr? goes in full newsgroup style, bravo.
>
> i dunno what kinda dr u r, but judging from what u just wrote, u
> probably r a nobody. i LOL.
>
>  Xah ? http://xahlee.org/ ?


Mathematical economist, Harvard- and Stanford-trained.

Ph.D. in Public Policy Analysis, and a lifetime of work
in the field.

I've forgotten more than you know about economics.  With
probability 1.  In a measure-theoretic sense.  ;)

And I've been hacking computers longer than you have been
alive, though I wouldn't be surprized if you weren't better
at that than I am.

Cheers, B.

On Aug 20, 8:53=C2=A0pm, "Dr. Brian Leverich"
<lever...@askin-17.linkpendium.com> wrote:
> On 2010-08-21, Xah Lee <xah...@gmail.com> wrote:
> > On Aug 20, 8:30=C2=A0pm, "Dr. Brian Leverich"
> ><lever...@askin-17.linkpendium.com> wrote:
> >> Thomas Sowell is pretty much an intellectual lightweight
> >> who skates by on the man-bites-dog shtick of being a
> >> Black conservative.
>
> >> Suggesting anyone learn economics from Sowell would be
> >> pretty much the most stupid thing I've heard today, if
> >> not this week.
>
> >> In general the whole Chicago School is a crock of shit
> >> based on vastly oversimplified models that have appeal
> >> for the point-and-drool crowd of economic idiots. =C2=A0If
> >> you're paying attention to them (particularly after the
> >> last couple of years), you are an idiot.
>
> > LOL. We have one ?Dr? goes in full newsgroup style, bravo.
>
> > i dunno what kinda dr u r, but judging from what u just wrote, u
> > probably r a nobody. i LOL.
>
> > =C2=A0Xah ?http://xahlee.org/?
>
> Mathematical economist, Harvard- and Stanford-trained.
>
> Ph.D. in Public Policy Analysis, and a lifetime of work
> in the field.
>
> I've forgotten more than you know about economics. =C2=A0With
> probability 1. =C2=A0In a measure-theoretic sense. =C2=A0;)
>
> And I've been hacking computers longer than you have been
> alive, though I wouldn't be surprized if you weren't better
> at that than I am.

longer than i'm alive?

can u still get it up?

hi Dr Brian: i lol!

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

much Microsoft hatred has been spreading again, i feel compelled to
fight this war.

so, here is a article on the issue of my take.

=E2=80=A2 The Microsoft Hatred FAQ
  http://xahlee.org/UnixResource_dir/writ/mshatredfaq.html

plain text version follows

--------------------------------------------------------
=E2=88=91 =E2=96=B2 Unix Pestilence: Open Source.

Microsoft Hatred, FAQ

Xah Lee, 2002-05-18

(originally posted to a Mac OS X mailing list =E2=80=9Cmacosx-
talk@omnigroup.com=E2=80=9D)

Q: US Judges are not morons, and quite a few others are not morons.
They find MS guilty, so it must be true.

A: So did the German population thought Jews are morons by heritage,
to the point that Jews should be exterminated from earth. Apparently,
the entire German population cannot be morons, they must be right.

Judge for yourself, is a principle i abide by. And when you judge, it
is better to put some effort into it.

How much you invest in this endearvor depends on how important the
issue is to you. If you are like most people, for which the issue of
Microsoft have remote effect on your personal well-being, then you can
go out and buy a case of beer on one hand and pizza on the other, and
rap with your online confabulation buddies about how evil is MS. If
you are a author writing a book on this, then obviously its different
because your reputation and ultimately daily bread depend on what you
put down. If you are a MS competitor such as Apple or Sun, then
obviously you will see to it with as much money as you can cough out
that MS is guilty by all measures and gets put out of business. If you
are a government employee such as a judge, of course it is your
interest to please your boss, with your best accessment of the air.

When i judge things, i like to imagine things being serious, as if my
wife is a wager, my daughter is at stake, that any small factual error
or mis-judgement or misleading perspective will cause unimaginable
things to happen. Then, my opinions become better ones.

Q: Microsoft's Operating System is used over 90% of PCs. If that's not
monopoly, i don't know what is.

A: Now suppose there is a very ethical company E, whose products have
the best performance/price ratio, and making all the competitors
looking so majorly stupid and ultimately won over 90% of the market as
decided by consumers. Is E now a monopoly? Apparently, beer drinkers
and pizza eaters need to study a bit on the word monopoly, from the
perspectives of language to history to law. If they have some extra
time, they can sharpen views from philosophy & logic contexts as well.

Q: What about all the people in the corporate environments who are
forced to use MS products and aren't allowed the option/choice to use
Mac/Linux/UNIX?

A: Kick your boss's ass, or, choose to work for a company who has
decisions that you liked.

Q: What about MS buying out all competitors?

A: Microsoft offered me $1 grand for saying good things about them.
They didn't put a gunpoint on my head. I CHOOSE to take the bribe.
Likewise, sold companies can and have decided what's best for them.

Q: Microsoft forced computer makers to not install competitor's
applications or OSes.

A: It is free country. Don't like MS this or that? Fuck MS and talk to
the Solaris or BeOS or AIX or HP-UX or Apple or OS/2 or Amiga or NeXT
or the BSDs and Linuxes with their free yet fantastically uneasy-to-
use and network-spamming X-Windows. Bad business prospects? Then grab
the opportunity and become a entrepreneur and market your own kick-ass
OS. Too difficult? Let's sue Microsoft!

Q: Microsoft distributed their Internet Explorer web browser free,
using their =E2=80=9Cmonopoly=E2=80=9D power to put Netscape out of busines=
s.

A: Entirely inane coding monkeys listen: It takes huge investment to
give away a quality software free. Netscape can give away Operating
Systems free to put MS out of business too. Nobody is stopping Sun
Microsystem from giving Java free, or BeOS a browser free, or Apple to
bundle QuickTime deeply with their OS free.

Not to mention that Netscape is worse than IE in just about every
version till they become the OpenSource mozilla shit and eventually
bought out by AOL and still shit.
Netscape struggles, announced open browser source code in 1998-01,
industry shock. http://wp.netscape.com/newsref/pr/newsrelease558.html
(local copy: =E2=84=A4 netscape_os_ann_1998.zip)
Netscape browser code released in 1998-03. Mozilla FAQ.
http://mozilla.org/docs/mozilla-faq.html (=E2=84=A4 local copy as of 1999)
AOL buys Netscape in 1998-11 for 4.2 billion. (http://news.com.com/
2100-1023-218360.html?legacy=3Dcnet, local copy: =E2=84=A4
aol_buys_netscape_1998.zip)
Jamie Zawinski, resignation and postmortem, 1999-04. http://www.jwz.org/gru=
ntle/nomo.html
(=E2=84=A4 local copy)
suck.com, Greg Knauss & Terry Colon, 2000-04, Netscape 6 mockery.
http://www.suck.com/daily/2000/04/10/, (=E2=84=A4 local copy)
Xah Lee, Netscape Crap. http://xahlee.org/UnixResource_dir/writ/netscape_cr=
ap.html.

Q: Microsoft implemented extra things to standard protocols in their
OS so that other OS makers cannot be compatible with their OS while
their OS can be compatible with all. They used this Embrace & Extend
to lock out competitors.

A: My perspective is this: suppose you are now a company who's OS sits
over 90% of computers (regardless how this come to be for the moment).
Now, lots of =E2=80=9Cstandard=E2=80=9D protocols in the industry is a resu=
lt of
popularity (RFC =3D Really Fucking Common), and popularity resulted from
being free, from the RFCs of the fantastically incompetent by the
truely stupid unix tech morons. What can you do if you want to improve
these protocols? If you go with totally different protocols, then the
incompatibility with the rest 10% isn't your best interest. I would
adopt existing protocols, and extend them with improvements. Being a
commercial entity, i'm sorry that it is not my duty to release my
improvments to my competitors. Any of you incompetent IBM/AIX/OS/2 or
SGI/Irix or HP/HP-UX or Sun/Solaris or Apple/AU-X/Mac can do the same,
not that they haven't.

Of course, the universe of unix idiots and Apple fops cannot see that.
The unix idiots cannot see that their fantastically stupid protocols
are fantastically stupid in the first place. The Apple fanatics are
simply chronically fanatical.

Q: Microsoft product is notorious for their lack of security.

A: In my very sound opinion, if Microsoft's OS's security flaws is
measured at one, then the unixes are measured at one myriad. If unixes
suddenly switch popularity with Windows, then the world's computers
will collapse uncontrollably by all sorts of viruses and attacks. This
can be seen for technical person who knows unix history well:

=E2=80=A2 http://xahlee.org/UnixResource_dir/freebooks.html (e.g. ftpd/
proftpd, inetd/xinetd, sendmail/qmail, X-Windows, telnet, passwd,
login, rsh, rlogin.)

=E2=80=A2 on the criminality of buffer overflow, by Henry Baker, 2001.
http://xahlee.org/UnixResource_dir/_/buffer_overflow.html

=E2=80=A2 Fast Food The UNIX Way: http://xahlee.org/UnixResource_dir/_fastf=
ood_dir/fastfood.html

=E2=80=A2 Early editions of the Jargon File.

=E2=80=A2 The Rise of Worse is Better, by Richard P. Gabriel, 1991, at
http://www.jwz.org/doc/worse-is-better.html

and plenty other pre-90s documents to get a sense of just how
fantastically insecure unix was and is. Unix today is not just
technically slacking in the =E2=80=9Csecurity=E2=80=9D department, but the =
unix ways
created far more unmanageable security risks that's another topic to
discuss.

The unix crime, is not just being utmost technically sloppy. Its
entire system and =E2=80=9Cphilosophy=E2=80=9D created a entire generation =
of
incompetent programers and thinking and programing languages, with
damage that is a few magnitude times beyond all computer viruses and
attacks damages in history combined. See also: Responsible Software
License: http://xahlee.org/UnixResource_dir/writ/responsible_license.html.

Q: Microsoft products are simply poor quality.

A: Perhaps this in general is true pre-1997. I think the vast majority
of MS products today have better performance/price ratio than
competitors. This includes their operating system, their input devices
(mouse & keyboard), their X-Box gaming console, their software game
titles, their software architectures and languages (.NET, C#), their
technologies (few i know: SMB), and many of their software
applications (suite of Office, which consistently ranked top since
early 90s).

e.g. Tom's hardware review on x-box, esp in comparison with Sony
Playstation 2. (2002-02): http://www4.tomshardware.com/consumer/02q1/020204=
/index.html

the leading role of MS Office products can be seen in MacUser &
MacWorld magazine reviews through out early 90s.

Q: BeOS was once to be bundled with PC, but MS meddled with it and
basically at the end fucked Be up.

A: BeOS is a fantastically fucking useless OS. No DVD player, No Java,
No QuickTime, No games, no Mathematica, no nothing. For all practical
purposes, fucking useless in a different way than every donkey unixes.
Not to mention the evil Apple computer, refused to pass the QuickTime
technology, and tried to prevent BeOS from running on Apple hardware
by refusing to release their PPC hardware spec. Be founder Jean-Louis
Gassee wrote a article about it. Who's fucking whom?

Q: X inc tried to do W, but MS threatened to depart.

A: Dear X inc., try to find a bigger cock for your needs. If you
cannot find any, too bad! Suck it up to the big brother and hold on to
what you can get! If you have the smarts, milk him dry! Free country,
free to choose partnership. Ladies, previous night's indiscretion does
not become rape the morning after.

Q: I'm not a beer bucket or pizza hole, but i want to do research over
the web. Is there any free stuff on the web i can grab? I'm a
OpenSource advocate, i demand free things.

A: =E2=80=A2 http://www.moraldefense.com/Campaigns/Microsoft/Antitrust_FAQ/=
default.htm
(The Center for the Moral Defense of Capitalism. URL defunct as of
2006-12. local copy)

=E2=80=A2 http://www.cato.org/pubs/policy_report/v21n2/friedman.html (The
Business Community's Suicidal Impulse by Milton Friedmanw, 1999-03)
local copy

=E2=80=A2 http://reason.com/0111/fe.dk.antitrusts.shtml Antitrust's Greates=
t
Hits: The foolish precedents behind the Microsoft case By David B.
Kopel and Joseph Bast. 2001-11. (local copy)

Q: I'm thinking of putting my wife and daughter on the table. What do
you suggest to begin with?

A: Basic Economics by Thomas Sowell. (See Reading Notes on Basic
Economics) Related:

--
 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

 KT> Yep! I have real work to do,

I have real work to do too, so?
It's not excuse for writing "moronic posts".

 KT>  you know: 
 KT>     http://www.stuckonalgebra.com/index.html

I know. Everybody here knows.
You've mentioned it so many times that it is pretty close to spam.

 KT> A pure Lisp Web application with just enough JS glue to make a JS 
 KT> framework transparent, in this case qooxdoo:
 KT>     http://wiki.github.com/kennytilton/qooxlisp/

I'll take a look to evaluate how much it sucks.

 KT> Die, Microsoft! Die!

The, Microsoft! The!

On 8/20/2010 4:04 AM, Antony wrote:
> On 8/19/2010 3:41 PM, Tim X wrote:
>>
>> The bottom line is that taking 20 minutes to install a web browser in
>> this day an age is just ridiculous. End of story.
> I think the time is a good indication that something more than the
> 'browser' is getting updated ;)

I think we all realize that. That is why the comparison I made was with 
a full Ubuntu install.

kt
-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 8/20/2010 1:26 PM, Xah Lee wrote:
> On Aug 18, 9:20 am, Kenneth Tilton<kentil...@gmail.com>  wrote:
>> On 8/18/2010 9:26 AM, Nick Keighley wrote:
>>
>>> On 18 Aug, 11:12, Kenneth Tilton<kentil...@gmail.com>    wrote:
>>>> I just installed IE8 to check out a user report of a problem with this
>>>> triumph of Lisp, Cells, qooxdoo, and jsMath:http://teamalgebra.com
>>
>>>> Installing IE8 over IE7 took about twenty minutes. I think a complete
>>>> Ubuntu install is about half that....
>>
>>> can my mum use Ubuntu?
>>
>> I think so, and indeed I think she would have a better shot with Ubuntu
>> than Windows. This is coming from a Linux hater who has had to work
>> intensely with Linux lately and am thrilled I did not have to mess with
>> the usual Linux maintenance nonsense.
>>
>> kt
>
> o rly?
>
> where can i buy ubuntu? from slashdot?

buy???? http://www.ubuntu.com/

If you want to buy something, buy VM Workstation and then build as many 
Linux machines as you like. Truly point and click.

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 8/20/2010 1:22 PM, Xah Lee wrote:
> On Aug 18, 3:51 am, Tim Bradshaw<t...@tfeb.org>  wrote:
>> On 2010-08-18 11:12:53 +0100, Kenneth Tilton said:
>>
>>> They've run up the white flag, Microsoft, haven't they?
>>
>> I think it's what happens to monopolists: they get big by being a
>> monopolist, and because they have no competition they get slow and
>> bureaucratic.
>
> lol. what load of crap.
>
> many tech geekers don't understand what the word monopoly means.
>
> monopoly is preventing others from competition.
> monopoly is not saturation of a market.
>
> when tech geekers speaks of Microsoft monopoly, they mean saturation
> of a market.

Speak for yourself. I am talking about anti-competitive practices. 
Really unecessary, too, since they had dominant share anyway. And now 
all Bill does is worry about how to give away all the money he grabbed 
unnecessarily. That's the joke. The unfunny part is the people he hurt 
along the way, and the scorn most folks have for him and his company.

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

Kenneth Tilton <kentilton@gmail.com> writes:

> Speak for yourself. I am talking about anti-competitive
> practices. Really unecessary, too, since they had dominant share
> anyway. And now all Bill does is worry about how to give away all the
> money he grabbed unnecessarily. That's the joke. The unfunny part is
> the people he hurt along the way, and the scorn most folks have for
> him and his company.

This is what power is: the ability to decide what to do with others' money.

And indeed,  Bill Gates is really very powerful, since now he can even
decide what to do with his peers billionaires' money  (at least, he
tries to convince them to give a big share of their money in what *he*
decided was worthwhile).

-- 
__Pascal Bourguignon__                     http://www.informatimago.com/

2010-08-21

On Aug 21, 8:50=C2=A0am, Kenneth Tilton <kentil...@gmail.com> wrote:
> On 8/20/2010 1:22 PM, Xah Lee wrote:

> > On Aug 18, 3:51 am, Tim Bradshaw<t...@tfeb.org> =C2=A0wrote:
> >> On 2010-08-18 11:12:53 +0100, Kenneth Tilton said:
>
> >>> They've run up the white flag, Microsoft, haven't they?
>
> >> I think it's what happens to monopolists: they get big by being a
> >> monopolist, and because they have no competition they get slow and
> >> bureaucratic.
>
> > lol. what load of crap.
>
> > many tech geekers don't understand what the word monopoly means.
>
> > monopoly is preventing others from competition.
> > monopoly is not saturation of a market.
>
> > when tech geekers speaks of Microsoft monopoly, they mean saturation
> > of a market.
>
> Speak for yourself. I am talking about anti-competitive practices.
> Really unecessary, too, since they had dominant share anyway. And now
> all Bill does is worry about how to give away all the money he grabbed
> unnecessarily. That's the joke. The unfunny part is the people he hurt
> along the way, and the scorn most folks have for him and his company.

well i paid a lot attention to this over the past decade about
Microsoft being anti-competitive... but i don't really see any, or any
that's worse than say, Sun Microsystems. And even in my 20 years of
using a computer i used Mac from 1990 to 2009-06 as my main PC and
never have the expertise or worked on any of Microsoft's tech as a
professional programer, and i don't like Microsoft, but my feeling is
that Microsoft Windows is simply vastly better than competitors, ALL
things considered.

what anti-competitive practice are are you thinking?

bundling a browsers? i'm sorry but i think that's rather a MAJOR, non-
trivial, INNOVATION.

threaten to cut off partnership if PC makers sleeps with other OS? Sue
ya wife if she threatens a divorse!

branching off Java? FUCK Sun Microsystems with its motherfucking
purity and run-everywhere marketing fuck Java lies FUCK, and i'm glad
Sun is dead. The scumbag pulled the same Netscape =E2=80=9Copen sourcing it=
=E2=80=9D
trick when death is near. AM VERY GLAD THE SUN MICRO FUCK IS DEAD.
Google also broke off ostensibly with their customized non-standard
Java, and Oracle just sued Google (but for different reason). (LOL!)

(see also:
=E2=80=A2 Jargons of Info Tech Industry
  http://xahlee.org/UnixResource_dir/writ/jargons.html

=E2=80=A2 What are OOP's Jargons and Complexities
  http://xahlee.org/Periodic_dosage_dir/t2/oop.html
)

the above are perhaps the most cited Microsoft faults. What others?
some halooween document? Fuck you linux slashdot groklaw tech geeking
social science illiterate sucmbags lying thru your teeths spreading
fucking illiterate motherfucking mis-info jargons FUD for your own
political gain, and fuck you richard stallman for bending the meaning
of the word FREE for your personal gain, and befuddle the concept of
human's freedom of behavior with your brand of software distribution
philosophy.

(
=E2=80=A2 =E2=80=9CFree=E2=80=9D Software Morality, Richard Stallman, and P=
aperwork
Bureaucracy
  http://xahlee.org/UnixResource_dir/writ2/FSF_philosophy.html
)

but tell me kenny what you think are some of the anti-competitive
practices.

O, btw, i've switched to PC now as my main machine in 2009, from 18
years of Mac, 10 of which being a Mac fanatic so to speak.

My story is here:

=E2=80=A2 Switching from Mac/Unix To PC/Windows
  http://xahlee.org/mswin/switch_to_windows.html

and with the decision to switch, i've thrown away my mental block of
not learning any Microsoft techs, and in the past year have read
extensively on MS techs (e.g. over 100 Wikipedia articles  related to
MS's techs), and started to write a bunch of Microsoft related
tutorials. You can see them collected here:

=E2=80=A2 Microsoft Windows Tech Tips
  http://xahlee.org/mswin/index.html

let me give one impression of their language. The Microsoft
PowerShell.

=E2=80=A2 Xah's PowerShell Tutorial
  http://xahlee.org/powershell/index.html

The PowerShell, in my opinion, is far more powerful and well designed,
than the combination of all of perl, python, php. (O my god, no wonder
Microsoft is so successful.) (for those functional programing nerds,
note that PowerShell was called Monad! Yay monad!)

O, u can see the Wikipedia articles i've read.

=E2=80=A2 Links to Wikipedia from XahLee.org
  http://xahlee.org/wikipedia_links.html

i've made estimates, that for each i actually linked to in my
writings, i've read about 6 more, starting about 2004. (was thinking
10, but last time i did a more serious estimate, i think 6 is proper.)

Since about 2006, i spend 16 hours in front of computer each day.
(actually something like that can be said since 1995) 1/2 of that is
in emacs, 1/2 in Wikipedia, and 1/2 in Firefox. LOL

come, Microsoft haters, let's start a flamewar on this. am willing to
entertain with my smattering of english. (except kenny)

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On 8/21/2010 8:37 AM, Captain Obvious wrote:
> KT> Yep! I have real work to do,
>
> I have real work to do too, so?
> It's not excuse for writing "moronic posts".

Responding to you is my excuse for writing moronic posts. Monkey see, 
monkey do.

>
> KT> you know: KT> http://www.stuckonalgebra.com/index.html
>
> I know. Everybody here knows.
> You've mentioned it so many times that it is pretty close to spam.

The day that a breakthrough Lisp application making it possible to 
deliver Web applications almost without knowing either JS or HTML 
(http://github.com/kennytilton/qooxlisp) is spam is the same day anyone 
will respect your opinion.

Don't wait up.

>
> KT> A pure Lisp Web application with just enough JS glue to make a JS
> KT> framework transparent, in this case qooxdoo:
> KT> http://wiki.github.com/kennytilton/qooxlisp/
>
> I'll take a look to evaluate how much it sucks.

In comparison to your competing ABCL-only effort? Is that what this is 
all about? It is more important to grow lisp than fret over personal 
dominion: Let the best Lisp/Web toolkit win, then all Lispers win.

here is what I am doing with qooxlisp:

    http://teamalgebra.com

A full-blown application on the Web.

Your site?

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On Aug 21, 2:06=A0pm, Kenneth Tilton <kentil...@gmail.com> wrote:
> On 8/21/2010 8:37 AM, Captain Obvious wrote:
>
> > KT> Yep! I have real work to do,
>
> > I have real work to do too, so?
> > It's not excuse for writing "moronic posts".
>
> Responding to you is my excuse for writing moronic posts. Monkey see,
> monkey do.
>
>
>
> > KT> you know: KT>http://www.stuckonalgebra.com/index.html
>
> > I know. Everybody here knows.
> > You've mentioned it so many times that it is pretty close to spam.
>
> The day that a breakthrough Lisp application making it possible to
> deliver Web applications almost without knowing either JS or HTML
> (http://github.com/kennytilton/qooxlisp) is spam is the same day anyone
> will respect your opinion.
>
> Don't wait up.
>
>
>
> > KT> A pure Lisp Web application with just enough JS glue to make a JS
> > KT> framework transparent, in this case qooxdoo:
> > KT>http://wiki.github.com/kennytilton/qooxlisp/
>
> > I'll take a look to evaluate how much it sucks.
>
> In comparison to your competing ABCL-only effort? Is that what this is
> all about? It is more important to grow lisp than fret over personal
> dominion: Let the best Lisp/Web toolkit win, then all Lispers win.
>
> here is what I am doing with qooxlisp:
>
> =A0 =A0http://teamalgebra.com
>
> A full-blown application on the Web.
>
> Your site?
>
> kt

lol kenny, u've been more spammy and cranky in the past months.

i think the 10 or so readers of comp.lang.lisp us already know all
your lisp works. If you are getting into some seo techniques, perhaps
other fora or more math oriented places would help. :)

 Xah

On 8/21/2010 5:51 PM, Xah Lee wrote:
> lol kenny, u've been more spammy and cranky in the past months.

No, seriously: what Lisp web app framework do you use? Ah, hang on, you 
already mentioned elsewhere you don't do Javascript -- all hand-coded 
static HTML? That's fine, but you are not my audience. OTOH, if you want 
to step your game up a level, qooxlisp would be a good way to avoid 
learning more than the absolute minimum. Then you get to learn, Cells, too:

    http://wiki.github.com/kennytilton/cells/

That would make you forget Mathematica.

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 8/21/2010 11:46 AM, Kenneth Tilton wrote:
> On 8/20/2010 1:26 PM, Xah Lee wrote:
>> On Aug 18, 9:20 am, Kenneth Tilton<kentil...@gmail.com> wrote:
>>> On 8/18/2010 9:26 AM, Nick Keighley wrote:
>>>
>>>> On 18 Aug, 11:12, Kenneth Tilton<kentil...@gmail.com> wrote:
>>>>> I just installed IE8 to check out a user report of a problem with this
>>>>> triumph of Lisp, Cells, qooxdoo, and jsMath:http://teamalgebra.com
>>>
>>>>> Installing IE8 over IE7 took about twenty minutes. I think a complete
>>>>> Ubuntu install is about half that....
>>>
>>>> can my mum use Ubuntu?
>>>
>>> I think so, and indeed I think she would have a better shot with Ubuntu
>>> than Windows. This is coming from a Linux hater who has had to work
>>> intensely with Linux lately and am thrilled I did not have to mess with
>>> the usual Linux maintenance nonsense.
>>>
>>> kt
>>
>> o rly?
>>
>> where can i buy ubuntu? from slashdot?
>
> buy???? http://www.ubuntu.com/
>
> If you want to buy something, buy VM Workstation and then build as many
> Linux machines as you like. Truly point and click.

Ah, nice data point. A new version of VM Workstation has been released. 
I have that for Vista, Linux, and the Mac. Not really much of a scoop, 
but it's Windows, not IE. Hell, one of the two (!!!) reboots required 
took longer than the Mac upgrade.

I am giving MS perhaps undeserved credit for ever having tried when I 
say they've stopped.

kt

ps. Someone reported a different/much better time for their IE8 install. 
That's nice. Go get a hundred Windows systems and calculate the standard 
deviation. I'll wait here. k

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

Xah Lee <xahlee@gmail.com> wrote:
> On Aug 21, 2:06?pm, Kenneth Tilton <kentil...@gmail.com> wrote:
>> ? ?http://teamalgebra.com
>>
>> A full-blown application on the Web.
>>
>> Your site?
> 
> lol kenny, u've been more spammy and cranky in the past months.

Whoa!

When Xah Lee is calling someone out as a possible troll, that person
better take a look at themselves....

:)

-pete

P.S. No worries, it's all in good fun!

On Aug 21, 6:39=C2=A0pm, Kenneth Tilton <kentil...@gmail.com> wrote:
> On 8/21/2010 5:51 PM, Xah Lee wrote:
>
> > lol kenny, u've been more spammy and cranky in the past months.
>
> No, seriously: what Lisp web app framework do you use?

emacs lisp?

> Ah, hang on, you already mentioned elsewhere you don't do Javascript
> -- all hand-coded static HTML? That's fine, but you are not my
> audience. OTOH, if you want to step your game up a level, qooxlisp
> would be a good way to avoid learning more than the absolute
> minimum. Then you get to learn, Cells, too: =C2=A0
> =C2=A0http://wiki.github.com/kennytilton/cells/ That would make you
> forget Mathematica.

where did a say i don't do javascript?

=E2=80=A2 Xah's Javascript Tutorial
  http://xahlee.org/js/js.html

oh, and yes my over 4 thousand pages visited by 7 thousand visitors
PER DAY are hand coded, but some are elisp generated. See for
examples:

=E2=80=A2 Links to Wikipedia from XahLee.org
  http://xahlee.org/wikipedia_links.html

=E2=80=A2 Google Earth Files at XahLee.org
  http://xahlee.org/kml_links.html

=E2=80=A2 Creating A Sitemap With Emacs Lisp
  http://xahlee.org/emacs/make_sitemap.html

kenny, i don't think u want to do piss links with me. :)

btw, this is a good opportunity for me to brag. Even though my html is
manually crafted love affair, but, i challenge anyone, with whatever
mark-down or wiki system they use, that i can type, format, to desired
effect, FASTER than them.

let me explain what i mean. For example, here's some common markup
syntax

=3D first header =3D
=3D=3D second header =3D=3D

simple paragraph, no need to mark anything

[ http://en.wikipedia.org/wiki/Lightweight_markup_language link done
with square bracket, sometimes double]

{
; program code here. literal format
(+ 3 4)
}

Now, i claim, i when i write html in emacs, i spontaneously generate
all the tags, faster then they can type those single char markdowns.
e.g. the above html equivalent would be like this:

<h1>first header</h1>
<h2>second header</h2>
<p>simple paragraph</p>

<p><a href=3D"http://en.wikipedia.org/wiki/
Lightweight_markup_language">Lightweight markup language</a></p>

<pre class=3D"elisp">
<span class=3D"comment-delimiter">; </span><span class=3D"comment">program
code here. literal format
</span>(+ 3 4)
</pre>

wooah, can you believe that??

let me repeat to be sure.

i don't care if you are a champion of speed-typing in highschool, and
i don't care what kinda simple, efficient, ass-kicking markdown system
you use, my claim is, that when i work in emacs with my own
elaborately customized system, i create the equivalent rendering with
the much verbose html markup, FASTER than you can with your markdown.
(and with more exacting control and rendering flexibility too, with
PROPER CSS! AND, IT'S ALL HTML4 STRICT VALID!!!)

whoaah, did i blew anyone's mind? i'm willing to subject to tests or
open challenges.

humm... i think some honest fellow here may be wondering just what
system i'm using. Let me be humble and tell you straight forward here
without being mysterious.

Keyboard:

=E2=80=A2 The Microsoft Natural Ergonomic Keyboard 4000
  http://xahlee.org/emacs/ms_keyboard/ms_natural_keyboard_4000.html

Keyboard layout:

=E2=80=A2 The Dvorak Keyboard Layout
  http://xahlee.org/comp/dvorak_keyboard_layout.html

App:

=E2=80=A2 ErgoEmacs Emacs distribution
 http://ergoemacs.org/

Emacs Keybinding system:

=E2=80=A2 ErgoEmacs Keybinding
 http://xahlee.org/emacs/ergonomic_emacs_keybinding.html

 (plus extra customization
( can be seen here http://xahlee.org/emacs/xah_emacs_init.html
(more specifically, these files on that page related to efficiency of
keys
    * xah_emacs_keybinding.el
    * xah_emacs_unicode_input.el
    * xah_emacs_alias.el
    * xah_emacs_abbr.el
)
 ) )

Extensive customized emacs html mode system. Part of it are discussed
here:

=E2=80=A2 =E2=80=9CZen-Coding=E2=80=9D and =E2=80=9CEmacs-Buddha-Coding=E2=
=80=9D
  http://xahlee.org/js/zen-coding.html

=E2=80=A2 Emacs and HTML Tips
  http://xahlee.org/emacs/emacs_html.html

Extensive os-wide key macros:

=E2=80=A2 AutoHotkey Example Scripts
  http://xahlee.org/mswin/autohotkey_examples.html

(part of my custom file can be seen at the bottom of that page, here's
the direct url:
  http://xahlee.org/mswin/xah_autohotkey.ahk
)

OS: Mac or Windows, either way. Currently on Windows. If on Mac, the
os-wide macro replacement of AutoHotkey will be the Mac's
DefaultKeyBinding.dict and QuickSilver.

=E2=80=A2 How To Create Keybinding In Mac OS X
  http://xahlee.org/emacs/osx_keybinding.html

well, i'm not satisfied with the brag above. Y'know? many linux tech
geekers brag about vi, command lind, ratpoison system etc shit, or
emacs old farts brag about emac's default keybinding system, or lisp
coders brag about the paren edit mode. I, Xah Lee, hereby claims that
when it comes to general computer operating tasks (such as switching
windows, launching apps, get a piece of text here past there,
processing file, etc ), or specific lisp coding tasks (e.g. general
editing process when coding lisp), that i work faster, more efficient,
less key strokes, than them.

yes, all this i'm willing to test publicly in some way. There was a
fellow who announced such a open competition a month ago. Let's see
how that goes.

why do i dare to brag about this? because i'm the biggest computer
operation efficiency nerd on earth. My methods or the system i
personally developed to use, are not based on hearsay or habit or
hacker lore as most tech geeker idiots are about these issues. My
methods, is developed with a scientifically verifiable ways as much as
possible.

=E2=80=A2 All About Keyboards, Keyboard Layouts, Shortcuts, Macros
  http://xahlee.org/Periodic_dosage_dir/keyboarding.html

well, sometimes i feel bad because i brag too much. I wonder if i'd
damage my public image... but ya know? i have to say what i think is
true, and my =E2=80=9Cbraggings=E2=80=9D above i think is fairly true. Yes,=
 someone
create a public test please.

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On 8/22/2010 12:03 AM, Xah Lee wrote:
> where did a say i don't do javascript?

On 8/15/2010 1:57 PM, Xah Lee wrote:
 > To make it work well one might have to implement javascript so that
 > mouse hover can popup a balloon with info about the char... my
 > javascript isn't so great...

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

On 8/21/2010 10:34 PM, Peter Keller wrote:
> Xah Lee<xahlee@gmail.com>  wrote:
>> On Aug 21, 2:06?pm, Kenneth Tilton<kentil...@gmail.com>  wrote:
>>> ? ?http://teamalgebra.com
>>>
>>> A full-blown application on the Web.
>>>
>>> Your site?
>>
>> lol kenny, u've been more spammy and cranky in the past months.
>
> Whoa!
>
> When Xah Lee is calling someone out as a possible troll, that person
> better take a look at themselves....

You just discovered that Xah is not a troll.

:)

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

Xah Lee <xahlee@gmail.com> writes:

> what anti-competitive practice are are you thinking?
>
> [...]
>
> threaten to cut off partnership if PC makers sleeps with other OS?

Precisely.  I'm amazed that such exclusive contracts are not forbidden.

Nicolas

 ??>> I have real work to do too, so?
 ??>> It's not excuse for writing "moronic posts".

 KT> Responding to you is my excuse for writing moronic posts. Monkey see,
 KT> monkey do.

But you've started it! You fail at causality.

 ??>> I know. Everybody here knows.
 ??>> You've mentioned it so many times that it is pretty close to spam.

 KT> The day that a breakthrough Lisp application making it possible to
 KT> deliver Web applications almost without knowing either JS or HTML
 KT> (http://github.com/kennytilton/qooxlisp) is spam is the same day anyone
 KT> will respect your opinion.

Um, how many times do you need to "report" it?

By the way, we already had thing like this: it is called SymbolicWeb and, as 
far as I can tell, it has same approach -- serverside fully controls 
presentation and just sends pieces of JavaScript to control browser.
Well, there are no cells there, but it might be a plus.

I don't know if yours is better, but definitely they are competitors and 
SymbolicWeb was here like 2 year earlier.
So I'm afraid breakthrough is a bit exagerated.

 KT>>> A pure Lisp Web application with just enough JS glue to make a JS
 KT>>> framework transparent, in this case qooxdoo:
 KT>>> http://wiki.github.com/kennytilton/qooxlisp/
 ??>> I'll take a look to evaluate how much it sucks.

 KT> In comparison to your competing ABCL-only effort?

Uh, no, I don't have any "competing effort", let alone ABCL-only.

I just want to look at it, check whether it is usable by ordinary humans.
Maybe do some benchmarking...

 KT>  Is that what this is all about?

No, you've got it wrong.

 KT>  It is more important to grow lisp than fret over personal dominion:
 KT> Let the best Lisp/Web toolkit win, then all Lispers win.

I don't mind. Althoug I think there is not single "best toolkit" -- people 
do different thing on the web.

 KT> here is what I am doing with qooxlisp:
 KT>     http://teamalgebra.com
 KT> A full-blown application on the Web.

Ok, let me check it.

1. First thing: I see content fully loaded in 9 seconds. (Google Chrome, 5 
Mbit connection, Europe.)
It's a bit too much. If it was plain HTML generated on server, I don't think 
it would take more than half of second to load a page like this.

Yes, 9 seconds is less than 20 minutes it takes to install IE8. :) But you 
install new version of IE maybe once in few years and you can go drink some 
coffee while it is installing.
While you're browsing hundreds of web pages per day, each fucking day, and 
if each of them takes 9 seconds it is irritating and it is huge waste of 
time.
Maybe you're picking on Microsoft because you're dissatisfied with qooxlisp 
performance? :)

2. Site design looks a bit weird and not modern enough. Particularly, fonts 
are tiny.
    Ok, I don't expect a lot from programmer-designed educational 
application.
    But if you're positioning qooxlisp for mainstream web stuff, you need to 
demonstrate that it can actually render mainstream-looking designs.
    Typically design process looks like this: graphic designers make designs 
in Photoshop, then HTML/CSS specialists make HTML/CSS from graphic design 
(perhaps, fixing some stuff at same time).
    Then programmers cut that HTML into templates to be used in dynamic 
applications, fixing design and cursing designers.

    Does qooxlisp fit into this process? I have an impression that design is 
more programmer-centric.

3. I cannot select text on the page. WTF? It doesn't have familiar web 
look&feel.

4. No feedback when hovering button, see 3.

5. When I click "Login/Register" I see blank page for like 2 seconds and 
then stuff appears.
    It is bad both because of delay (delay on each page is worse than delay 
on application load; why do ajax-y stuff if it takes more time than full 
page reload) and weirdness -- blank page instantly produces reacion "WTF 
went wrong?". I've seen all kind of weirdness on the web so I don't really 
care, but less experienced users might freak out.

6. Switching radio buttons has some sub-second delay. It is not as bad as 
with loading new page, but you can feel that it is not client-side thing.
    So moving logic onto server is not free.

7. Typing examples -- usual two second delay on loading new page, two second 
delay on switching lessons.
     You can see how math symbols are being loaded, WTF? Is it an animation 
of some sort?
     It feels like application is half-backed. You don't have shit like this 
on desktop application -- people render stuff on off-screen surface and then 
show when it is ready, so there is no blinking/flickering and people 
percieve application as being solid.

8. Typing is slow, there is a delay after each symbol is being typed. 
(Luckily it is less than second.)
     People can get used to it, but it is not how applications on desktops 
work.

9. Training center. Switching is slow and you can see how elements are being 
drawn, as usual.
     Clicking the select shows another look&feel problem -- when you click 
an item it first goes white and only then changes to blue while elements are 
drawn.
     You know, this experience reminds me a situation when old, single-core 
computer get one task hanging taking 100% of CPU time so all repaint 
operations are slow and you can seem them in all the ugliness.
     Now you were able to reproduce same effect on fast multi-core 
computers. Bravo!

10. Another problem with selects: I'm on "numeric fractions" in left select 
and click "real numbers".
      But it takes a long time to render contents of right select, so I can 
choose "Reducing" before content is replaced.
      Then I see it showing contents of "Real numbers" section for a moment 
of time, chaning to "Numeric fractions" again.
      But left select says it is "Real numbers" -- left and right selects 
are desynchronized.

     You can say "Don't do it". Ok, I've seen lots of UI disasters and I 
know that things might get dysynchronized if you click fast enough.
     But children using your app might be less experienced with web 
usability disasters, so they will freak out and call teacher or something 
like that.

     I know because I worked on web-based educational software too. Our 
application had very traditional web look&feel and it was hard to break it, 
but still students were finding their ways.

11. Now I'm trying it in Firefox 3.6.8. Takes 8 seconds to load even if 
you're doing it for the second time and it should be cached. WTF? Not the 
case in Chrome, btw.
12. In Firefox sometimes you can see side buttons but they do not work while 
it is loading jsMath (takes something like 5 seconds). I'm not sure how to 
reproduce it exactly but I've seen it many times.
      Might be a problem when someone is trying to go to specific section, 
but buttons do not work -- it is hard to notice "loading jsMath" is corner.

So, I don't know if qooxlisp actually saves development time (at least might 
take some time to learn Cells if you're not Kenny Tilton), but applications 
you make with it are going to be sub-standard in all aspects, as we can 
judge from the only demo application.
And I don't know whether it is possible at all to fix lag issues keeping 
"control on server" paradigm.
Of course it would help to move servers closer to users, but why bother if 
you can implement logic in JS itself and have 0 lag in lot of cases?

 KT> Your site?

It is not my personal site (i.e. I get paid for an employer, sort of) and it 
is in close alpha stage, so there is no public site you can criticize.
And I guess you wouldn't want to be an alpha tester.

And, honestly, we don't do anything partcularly impressive on technological 
side.

Previous version tried to use some advanced things -- ucw_ajax, parenscript 
together with home-brewn OO widget toolkit.
But it didn't work very well -- ucw_ajax is too complex and buggy, and we 
hadn't time to fully develop that parenscript widgets stuff.
I guess it could work if we'd take some JS framework like qooxdoo and use it 
from parenscript. But writing one from scatch is too much work for a small 
team which always has other things to do -- like getting application done.

So we've switched to simplier and more straightforward approach -- on server 
side Lisp just provides RPC functions and client side is almost pure JS, 
using jQuery.
And there is also some web site sections made via html-template and a very 
thin framework on top of hunchentoot.
So no fancy stuff. It just works. I can't say all works perfectly, but there 
is no chance that we'll be stuck somewhere because of framework choice -- if 
it is doable at all then we can do it with stuff we have.

On 2010-08-21 04:24:20 +0100, Dr. Brian Leverich said:

> I *do* know what a monopoly is, and Microsoft pretty much
> definitely engaged in monopolistic practices once upon a time.
> 
> For those of you who aren't mathematical economists but still
> literate, you may have notice that Microsoft has, in fact, been
> found to have engaged in monopolistic behavior by an American
> court and by the European Commission.

I do as well, I think.  I certainly do not subscribe to the 
"monopolists are evil people" school of thought, which is pretty dumb 
in my view.   And it seems to me uncontroversial that a useful 
definition of a monopoly is being able to prevent others from competing 
in the market, and also that being able to do that leads to two things: 
(1) the monopolist has a very large market share, and (2) as they have 
no effective competition they end up producing poor products.

Xah probably thought I was claiming MS were a monopoly purely because 
they had large market share: if he'd bothered to ask he would have 
found out that's not what I think at all.  But of course he didn't 
bother to ask.

 ??>> Your busy server powered by Windows Vista is constantly swapping?
 ??>> You're doing it wrong, then. On so many levels...

It looks like you've tried to say something smart, but it totally misses the 
point.
I repeat: if your machine is constantly swapping (i.e. HDD LED is constantly 
blinking and you hear it doing seek noises all the time; aka thrashing) it 
is NOT ok.
Swapping is for emergencies.

I have seen situations where servers didn't have enough RAM for tasks and I 
can assure you that it is unlikely that machine can do its job while 
thrashing.

Particularly, my master's degree work included some educational software 
working via web: it presented some learning material and then did test. I 
thought it would be cool to do it in Common Lisp, particularly, ABCL was my 
implementation of choice, and I've hacked some continuation-based web 
framework.
The problem is that I was given a computer having 128 MB of RAM running 
Windows 2000 to do demonstration.
So, this computer was running Windows 2000 itself with its GUI (Explorer), 
Apache Tomcat with ABCL and my application, MySQL DBMS and perhaps some 
other stuff, I don't remember, maybe Apache for static files. During setting 
it up I had to run Emacs, WinWord and Explorer on same machine.

Well, it was not constantly swapping. I don't remember, maybe it required 
some I/O on each request, but at least it answered them timely. But from 
time to time it went complete GC, which meant that service was totally 
unavailable for a few minutes while it is swapping in/swapping out.
And use of continuations meant that it actually needed full GC from time to 
time -- continuations need to stay in memory for some amount of time, so 
they go into older generations -- ephemeral GC isn't of help here. Well, 
maybe if it had enough RAM ephemeral GC would work, but having enough RAM is 
a problem in the first place.

So, needless to say, I was scared shitless during demonstration. Our whole 
group used my program for demonstration, so it could affect other people too 
and it had to serve considerable number of requests...
So I was tuning heap sizes and continuation lifetimes for hours.
Luckily, they didn't test application intensively and it didn't go full GC.

But since then I kinda dislike thrashing and I think computers should 
normally work without it.

I often run Linux servers without any swap. I've found that even if working 
set fits into RAM swapping might introduce unpredictable performance 
problems (especially if GC is involved), so my policy is -- either it fits 
in RAM or we need to buy more RAM.

So, I don't buy your "SSD cannot last because there might be swapping". 
Either there is little swapping which is not problem at all, or there is 
heavy swapping and computer is unusable. I don't know any applications where 
swapping is a norm, and even if they exist, these applications themselves 
are not norm.

By the way, we can estimate how many writes swapping generates.
Typical HDDs do up to 100 random I/O operations per second. If system is in 
constant thrashing it is probably all random I/O half writes/half reads, 
with typical page size of 4 KiB it is 200 KiB per second written.
That is 16875 MiB per day =  16.48 GiB.

It is within limits of MLC (if we believe Intel's datasheet) and SLC can run 
circles around that.

So the conclusion is that SSD can do what HDD does even if it is constant 
swapping.
If you would say that SSD must do swapping at its full speed I'd say you're 
unreasonable.

 GN> I'm not going to be drawn into a debate on the merits of Windows as a
 GN> server.

Windows Vista and Windows 7 you've mentioned are not server editions.
There is Windows Server 2008 for this.

 GN>   People do it, so deal with it.

People do all kinds of crazy shit, but it is not valid reason for choosing 
hardware.

 GN> In general there is no way to prevent Windows from swapping -

Disable swap file and it will not swap, by definition.
It will do paging, though -- e.g. freeing and loading back pages of file 
mappings.
But it is another thing... Paging does not generate writes, at least.

Page-ins could kill HDD peformance but are no problem to SSD. Maybe you're 
confusing page-ins with this other stuff you wrote?

 GN>  it doesn't matter how much RAM you have.  Windows uses relocatable
 GN> DLLs as opposed to position independent ones, so a busy system
 GN> frequently has many copies of the same DLLs loaded simultaneously.
 GN> DLLs that are not mapped at their image address use the page file as
 GN> backing store.

If it is the case then it is just like memory allocations in applications --  
they are "backed by page file" too.
So it just means that whole need for memory is larger than it would be 
otherwise.
So?

Either it fits or it doesn't.

 GN> Windows gets upset if you disable swapping entirely

I dunno how upset it is, but it works. Although now I do not disable swap 
completely but keep it very small.
I know other people do same thing and

 GN> and the auto-tuning large cache manager in the server versions
 GN> essentially does what it pleases to adapt to dynamically changing
 GN> loads.  There is a fair amount of manual tuning that can be done, but
 GN> it is very difficult to dial in a server running a mix of applications
 GN> ... so it is almost always left to the cache manager.

So your servers are constanty thrashing, is this what you're trying to say?
Some swapping activity from time to time is not a problem because amount of 
data is limited. 

 TB> Do you think that the view that MS and Sun (under whatever ownership)
 TB> are slowly dying is somehow contentious?

Yes. I do not have a strong opinion on this, though.
My point was, however, that it "does not follow", i.e. it is hard to related 
slowness of IE8 installation to death of MS.

 TB> Patches do this by *reading and writing the whole file for each line
 TB> they change*.

Holy crap...

 TB> So, maybe this bug will take 20 years to "fix" where "fix" means
 TB> "ignore until we replace it with something else".  That's not a good
 TB> indicator, at the very best.

Ok, you've convinced me that Sun is not doing very well.

But Microsoft isn't that bad. They fix things. They are not afraid to change 
things for better breaking compatibility with old cruft. They spend a lot of 
efforts making old cruft co-existing with new solutions.

E.g. NT branch does not allow writting to C:\Program Files (like /usr) for 
non-admininstrator users with default settings -- for obvious settings. But 
both users and applications abused this -- users were usually working as 
administrators on XP and applications were writing to C:\Program Files\xxx 
(while they should write to C:\Profiles\user\Application Data\xxx).

In Vista they have fixed this -- first, now administrator is more like a 
member of wheel group who can do sudo and needs to do sudo for 
administrative tasks (although it is implemented via complex machinery on 
kernel level, as far as I understand).
Then applications which try to write to C:\Program Files while being 
executed in low-privilege mode actually run to some directory in 
C:\Profiles\user\... -- this is implemented transparently on filesystem 
level.

So, you see, Microsoft actually cares about users and their habits and they 
can go as far as introducing new OS components just to make their experience 
better.

Of course, not everything is perfect. But in case of slow installation times 
it's not like they've made totally retarded thing and do not want to fix it. 
More like it is not tuned very well. Maybe it will go 2x or 3x faster if 
ideally tuned and optimized, but it is not absolutely critical so they 
didn't do it.

 TB> That "something else" of course will turn out to be Linux.

Do you think that Linux will kill Windows too? Or that would be Apple's OS? 

On Aug 23, 2:48=C2=A0pm, "Captain Obvious" <udode...@users.sourceforge.net>
wrote:
> So, you see, Microsoft actually cares about users and their habits and th=
ey
> can go as far as introducing new OS components just to make their experie=
nce
> better.
[full article pasted below my sig]

well said Captain Obvious! (btw, you name made me lol. pun not
intended here.)

regarding about the permission to modify files in =E2=80=9CC:\Program Files=
=E2=80=9D,
i've written about it too. Here:

=E2=80=A2 Windows Vista VirtualStore Problem
  http://xahlee.org/UnixResource_dir/writ/windows_virtualstore.html

also, recently i happed to read a few articles about Microsoft written
by the venerable Microsoft developer Raymond Chen

=E3=80=88The Old New Thing: What about BOZOSLIVEHERE and
TABTHETEXTOUTFORWIMPS?=E3=80=89 (2003-10-15), by Raymond Chen, at
http://blogs.msdn.com/b/oldnewthing/archive/2003/10/15/55296.aspx

where you read that lots of shit is blamed to Microsoft, while the
real fault is 3rd party asses, and Microsoft spend tremendous time
just to still make things work.

Also, you know how Microsoft startup is like a snail? See:

=E2=80=A2 Windows Starting Too Slow? Disable Windows Startup Apps with
msconfig
  http://xahlee.org/mswin/disable_Windows_startup_apps.html

actually, just about all 3rd party software adds hidden gook of itself
to Windows startup, so when Windows starts slow, ms got blamed, but in
fact it's those fucking Adobe, iTune, Java, games. (in particular,
fuck Java the motherfucker, and Sun Microsystems fuckface, glad u
dead. (note that in a java update about few months ago, it
surreptitiously installs a 3rd party web based backup service
software. you can read about in in blogs if you care to search.))

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On Aug 23, 2:48=C2=A0pm, "Captain Obvious" <udode...@users.sourceforge.net>
wrote:
> =C2=A0TB> Do you think that the view that MS and Sun (under whatever owne=
rship)
> =C2=A0TB> are slowly dying is somehow contentious?
>
> Yes. I do not have a strong opinion on this, though.
> My point was, however, that it "does not follow", i.e. it is hard to rela=
ted
> slowness of IE8 installation to death of MS.
>
> =C2=A0TB> Patches do this by *reading and writing the whole file for each=
 line
> =C2=A0TB> they change*.
>
> Holy crap...
>
> =C2=A0TB> So, maybe this bug will take 20 years to "fix" where "fix" mean=
s
> =C2=A0TB> "ignore until we replace it with something else". =C2=A0That's =
not a good
> =C2=A0TB> indicator, at the very best.
>
> Ok, you've convinced me that Sun is not doing very well.
>
> But Microsoft isn't that bad. They fix things. They are not afraid to cha=
nge
> things for better breaking compatibility with old cruft. They spend a lot=
 of
> efforts making old cruft co-existing with new solutions.
>
> E.g. NT branch does not allow writting to C:\Program Files (like /usr) fo=
r
> non-admininstrator users with default settings -- for obvious settings. B=
ut
> both users and applications abused this -- users were usually working as
> administrators on XP and applications were writing to C:\Program Files\xx=
x
> (while they should write to C:\Profiles\user\Application Data\xxx).
>
> In Vista they have fixed this -- first, now administrator is more like a
> member of wheel group who can do sudo and needs to do sudo for
> administrative tasks (although it is implemented via complex machinery on
> kernel level, as far as I understand).
> Then applications which try to write to C:\Program Files while being
> executed in low-privilege mode actually run to some directory in
> C:\Profiles\user\... -- this is implemented transparently on filesystem
> level.
>
> So, you see, Microsoft actually cares about users and their habits and th=
ey
> can go as far as introducing new OS components just to make their experie=
nce
> better.
>
> Of course, not everything is perfect. But in case of slow installation ti=
mes
> it's not like they've made totally retarded thing and do not want to fix =
it.
> More like it is not tuned very well. Maybe it will go 2x or 3x faster if
> ideally tuned and optimized, but it is not absolutely critical so they
> didn't do it.
>
> =C2=A0TB> That "something else" of course will turn out to be Linux.
>
> Do you think that Linux will kill Windows too? Or that would be Apple's O=
S?

On 8/23/2010 6:19 PM, Xah Lee wrote:
> On Aug 23, 2:48 pm, "Captain Obvious"<udode...@users.sourceforge.net>
> wrote:
>> So, you see, Microsoft actually cares about...

$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$.

Period. Full stop.

On 2010-08-23 22:48:01 +0100, Captain Obvious said:

> Ok, you've convinced me that Sun is not doing very well.

Good.  We can probably then agree to differ on MS then: time will tell 
which of us is right (and I hope I've been clear enough that I'm not 
grinding some anti-MS axe here: I try to keep my emotional investment 
in anything to do with computers to a minimum.  Now, if we were talking 
about mid-late 20th century cameras...)

> 
> Do you think that Linux will kill Windows too? Or that would be Apple's OS?

I do not think Linux will kill MS on the desktop, or ever stood a 
chance of doing so, possibly outside some very special markets (really: 
markets MS is not interested in being in because there is no profit 
there).  I think two things: (1) Apple will continue to have a 
reasonable share, especially among people with a lot of disposable 
income, but they will not displace MS; (2) although MS will continue to 
dominate the desktop (and laptop), the interesting things will move 
away from that to phones/tablets/something, and MS will not manage to 
dominate there.  I don't know who will win there (if I did I would be 
rich), but it looks right now like it's between Google and Apple.  I 
think the really interesting story of the last few years is where Nokia 
dropped the ball, and whether they can catch up, as they ought to have 
cleaned up in this market.

Obviously all this is just my idiot pontificating and worth what you 
paid for it :-)

On Aug 23, 9:08=A0pm, Kenneth Tilton <kentil...@gmail.com> wrote:
> On 8/23/2010 6:19 PM, Xah Lee wrote:
>
> > On Aug 23, 2:48 pm, "Captain Obvious"<udode...@users.sourceforge.net>
> > wrote:
> >> So, you see, Microsoft actually cares about...
>
> $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$.
>
> Period. Full stop.

1.  MS cares about customers
2.  Profit

--------------------------------------------

1.  Sun cares about blogosphere goodwill
2.  Trot RMS onstage
3.  Not profit
4.  Oracle buys Sun
5.  Oracle sues Google

On Aug 24, 2:21=C2=A0am, Tim Bradshaw <t...@tfeb.org> wrote:
> > Do you think that Linux will kill Windows too? Or that would be Apple's=
 OS?
>
> I do not think Linux will kill MS on the desktop, or ever stood a
> chance of doing so, possibly outside some very special markets (really:
> markets MS is not interested in being in because there is no profit
> there). =C2=A0I think two things: (1) Apple will continue to have a
> reasonable share, especially among people with a lot of disposable
> income, but they will not displace MS; (2) although MS will continue to
> dominate the desktop (and laptop), the interesting things will move
> away from that to phones/tablets/something, and MS will not manage to
> dominate there. =C2=A0I don't know who will win there (if I did I would b=
e
> rich), but it looks right now like it's between Google and Apple. =C2=A0I
> think the really interesting story of the last few years is where Nokia
> dropped the ball, and whether they can catch up, as they ought to have
> cleaned up in this market.
>
> Obviously all this is just my idiot pontificating and worth what you
> paid for it :-)

Hi Tim.

I, too, have a opinion on this, by chance.

=E2=80=A2 Kindle, iPad, Android, and All That Jazz
  http://xahlee.org/comp/Kindle_iPad_and_all_that_jazz.html

=E2=80=A2 Apple iPad, Censorship
  http://xahlee.org/comp/Apple_iPad_censorship.html

actually just happened to wrote these last week.

plain text version of one of them follows.
--------------------------------------------------
Apple iPad, Censorship

Xah Lee, 2010-08-19

While writing reviews on Kindle and iPad, i learned that Apple
practices heavy censorship. Not just about technology issues of what
software technology is allowed to run, such as Adobe Flash, but actual
contents from established newspapers, magazines, cartoons. Here are
some details gathered from Wikipedia.

Censorship on Nudity

In May 2009, Apple rejected the first version of 'Newspapers', an
iPhone app that let users read content from 50+ newspapers around the
world, including the New York Times, France's Le Monde, and the United
Kingdom tabloid The Sun. The app was rejected because the topless
=E2=80=9CPage 3=E2=80=9D girls daily features were described as =E2=80=9Cob=
scene=E2=80=9D. A second
version of the application was submitted, removing access to The Sun,
and adding a price tag of =C2=A30.59. The app was made available in the
summer, after the release of the iPhone 3.0 software.[53][54]

Note that Page Three, is a well-known feature of The Sun, basically
features photo of topless or nude females.

Censorship on Ancient Text

Here's a rejection because it contains the book Kama Sutra. Quote:

Another application, of similar nature to 'Newspapers', called
'Eucalyptus' allowed users to download e-books to their iPhone, though
was censored by Apple because one of the e-books that could have been
downloaded was the Kama Sutra. The ban has since been lifted.[55]

The Hypocrisy

They censor nudity, but don't censor big, well-known, porn money
makers like Playboy.

It should be noted that the App Store has Playboy and Sports
Illustrated adult-rated apps that have yet to be removed, while some
apps by others were removed citing adult content which has resulted in
accusations of hypocrisy.[58][59][60]

In November 2009, the application of Stern (a mainstream German weekly
magazine with a print circulation of about 900,000) was deleted for
several weeks without warning.[57][61] In January 2010, Europe's
largest newspaper, German tabloid Bild, removed content from the
iPhone version of its print edition at the request of Apple, and later
it had to modify one of its applications - like in the Stern case
because of nudity.[62] The Association of German Magazine Publishers
(VDZ) warned that with such interventions Apple might be moving
towards censorship.[62]

.... Workers at the fashion magazine Dazed & Confused have nicknamed
their iPad edition the =E2=80=9CIran edition=E2=80=9D.[56]

Censorship On Satire

In December 2009, Apple banned a cartoon app called NewsToons by
cartoonist Mark Fiore, on the grounds that it =E2=80=9Cridiculed public
figures.=E2=80=9D[63][64]

In April 2010, Fiore won the Pulitzer prize for his political satire
cartoons, making history as the very first internet-only cartoonist to
win the prestigious journalistic prize.[65][63][64]

Following public outcry after the story broke in the wake of the
award, Apple asked Fiore to resubmit his app. Fiore said, =E2=80=9CSure, mi=
ne
might get approved, but what about someone who hasn=E2=80=99t won a Pulitze=
r
and who is maybe making a better political app than mine? Do you need
some media frenzy to get an app approved that has political
material?=E2=80=9D[64]

Bravo to Mark Fiore.

Censorship On Music

Also in May 2009, Trent Reznor of the rock band Nine Inch Nails
announced, via his Twitter account, that Apple had rejected an update
to the Nine Inch Nails application due to =E2=80=9Cobjectionable content=E2=
=80=9D.[67]
The developer posted a message on the Nine Inch Nails discussion
boards explaining the situation further: =E2=80=9Cv1.0 is live. v1.0.3 got
rejected due to content yet the app has no content in it. This was
mainly a stability release to fix the bug that crashes the app for
international users. The bug was fixed 24 hours after 1.0 went live
and we have been waiting for apple to approve it ever since. Meanwhile
the app continues to get a growing number of 1 star ratings from
international users understandably frustrated by the bug. =E2=80=9CBut look=
s
like our hands are tied=E2=80=9D.[68] Apple later permitted the update.[69]

Here's what Steve Jobs has to say about censorship:

=E2=80=9CWe do believe we have a moral responsibility to keep porn off the
iPhone ... Folks who want porn can buy an Android phone=E2=80=9D =E2=80=93 =
Steve
Jobs[56]

With all due respect to Steve's visions and innovation, moral his ass.

Here are some comical parts out of the censorship.

I Am Rich

I Am Rich. Quote:

In August 2008, an application known as I Am Rich was released in the
store, =E2=80=9Ca work of art with no hidden function at all=E2=80=9D, with=
 its only
purpose being to show other people that they were able to afford it,
as it cost US$999.99, =E2=82=AC799.99, and UK=C2=A3599.99.[46] The applicat=
ion was
removed from the App Store the day following its release, on August 6,
2008 (2008-08-06).[47] Eight people had bought it before it was pulled.
[48]

YouPorn Response

Due to the exclusion of porn from the App Store, YouPorn and others
changed their video format from Flash to H.264 and HTML5 specifically
for the iPad.[78][79] In an e-mail exchange[80] with Ryan Tate from
Valleywag, Steve Jobs claimed the iPad to offer =E2=80=9Cfreedom from porn=
=E2=80=9D,
leading to many upset replies including Adbustings in Berlin by artist
Johannes P. Osterhoff[81][82] and in San Francisco during WWDC10.[83]

Lol. The war of porn formats.

 Xah =E2=88=91 http://xahlee.org/ =E2=98=84

On 2010-08-24 12:30:09 +0100, Xah Lee said:

> Hi Tim.
> 
> [...]
> plain text version of one of them follows.
> --------------------------------------------------
> Apple iPad, Censorship

Haven't read the other, but I think you're right about censorship.  
We're fairly clearly watching the coming of age of the first generation 
of slaves of the internet: utterly reliant on it, but they don't 
understand (or care) how it works, and their lives will essentially be 
owned by the people who control it (facebook, apple &c &c).

Tim Bradshaw <tfb@tfeb.org> writes:

> On 2010-08-24 12:30:09 +0100, Xah Lee said:
>
>> Hi Tim.
>>
>> [...]
>> plain text version of one of them follows.
>> --------------------------------------------------
>> Apple iPad, Censorship
>
> Haven't read the other, but I think you're right about censorship.
> We're fairly clearly watching the coming of age of the first
> generation of slaves of the internet: utterly reliant on it, but they
> don't understand (or care) how it works, and their lives will
> essentially be owned by the people who control it (facebook, apple &c
> &c).


Actually, Benjamin Bayart explains it very well (but AFAICS only in
French), saying that what we're seeing is the Minitel 2.0, not the
Internet.

As you may know, Minitel was a little terminal (toy keyboard, small CRT
screen, 1200/75 bps modem) distributed by France Telecom to access to
telematic services.  You could access various kind of services such as
yellow pages, train reservations, etc.  They made a fortune with it on
"Messagerie Rose" services, sex chats.  

The point is that this system worked with central servers, and these dumb
terminals connected to one or another of the central server to get their
information.  All the processing was done on the central server, and all
the data was stored on the central server.

In the Internet, there is no central server, and all computers are
equals, containing their own data, and processing it themselves.

When you use a central server to communicate, such as gmail with
millions of mailboxes, or a central blog server, or a central web host,
or whatever other kind of central server of which you don't control the
censor policy, you're not using the Internet, you're using the
Minitel�2.0.


In the Internet, your blog is served from your own PC, your videos are
stored on your own PC and your email address is
yourself@your-own-pc.org, where there could be at most a few tens of
email addresses (if you have a big familly), but not 100,000,000.


Benjamin Bayart also promotes small (eg. associative) ISP, where the
people control directly their own part of the Internet, and can ensure
that net-neutrality is implemented without restrictions.  

When you watch the (very nice) maps of the Internet, what's strikingly
wrong, is how much star-like a network it is, while the true nature of
the Internet should be the mesh.  It is true that the backbone is
coarsely meshed.  But big ISP implement star-like network which afford
them centralized control of their prisonners^W customers.

What I would like to see is communes implementing local networks for the
citizen, and interconnecting themselves with neighboring communes so
that the mesh of the Internet would map the mesh of local road
connecting villages.


In any case, it's ludicruous, and dangerous, that when I send an email
to my friend living in the neighbor house, it should travel accross the
country to the capital, and back to my friend's.


-- 
__Pascal Bourguignon__
http://www.informatimago.com

On 2010-08-24 08:48:47 -0400, Pascal J. Bourguignon said:

> What I would like to see is communes implementing local networks for the
> citizen, and interconnecting themselves with neighboring communes so
> that the mesh of the Internet would map the mesh of local road
> connecting villages.

As much as I agree with you, as a rhetorical matter, if you keep 
repeating the word "commune" it will practically guarantee it to never 
happen.

Cries of "we want the internet to be a commune," are the shortest path 
to the death of net neutrality (because now killing net neutrality is 
fighting communism). Remember, you're dealing with business people and 
politicians here; if you tag one side of the debate with the label 
"commune" you'll ensure that they reflexively support the other.

"Decentralization," and "local control" would sell better.

warmest regards,

Ralph

-- 
Raffael Cavallaro

Raffael Cavallaro <raffaelcavallaro@pas.despam.s.il.vous.plait.mac.com> writes:

> On 2010-08-24 08:48:47 -0400, Pascal J. Bourguignon said:
>
>> What I would like to see is communes implementing local networks for the
>> citizen, and interconnecting themselves with neighboring communes so
>> that the mesh of the Internet would map the mesh of local road
>> connecting villages.
>
> As much as I agree with you, as a rhetorical matter, if you keep
> repeating the word "commune" it will practically guarantee it to never
> happen.
>
> Cries of "we want the internet to be a commune," are the shortest path
> to the death of net neutrality (because now killing net neutrality is
> fighting communism). Remember, you're dealing with business people and
> politicians here; if you tag one side of the debate with the label
> "commune" you'll ensure that they reflexively support the other.
>
> "Decentralization," and "local control" would sell better.
>
> warmest regards,

Sure.  I'm not a propagandist.  Please take the message and "refine"
it for the target audience...

-- 
__Pascal Bourguignon__                     http://www.informatimago.com/

Pascal J. Bourguignon <pjb@informatimago.com> wrote:
> When you watch the (very nice) maps of the Internet, what's strikingly
> wrong, is how much star-like a network it is, while the true nature of
> the Internet should be the mesh.  It is true that the backbone is
> coarsely meshed.  But big ISP implement star-like network which afford
> them centralized control of their prisonners^W customers.

Could you imagine hundreds of small villages trying to create and
enforce peer agreements with their neighbors--complete with political
feuds and other usual human antics? Or, when you reach the county
limits and the nice fiber goes into a mud-filled box and out comes
1930's phone wire?

I'm being somewhat the devil's advocate here, but there are certain
economies of scale which benefit the current topology.

I do agree though, that once a star-like system is in place, using it 
as leverage against the customers is easy and simply will happen....

-pete

Peter Keller <psilord@cs.wisc.edu> writes:

> Pascal J. Bourguignon <pjb@informatimago.com> wrote:
>> When you watch the (very nice) maps of the Internet, what's strikingly
>> wrong, is how much star-like a network it is, while the true nature of
>> the Internet should be the mesh.  It is true that the backbone is
>> coarsely meshed.  But big ISP implement star-like network which afford
>> them centralized control of their prisonners^W customers.
>
> Could you imagine hundreds of small villages trying to create and
> enforce peer agreements with their neighbors--complete with political
> feuds and other usual human antics? 

That's why I used the more generic term of "commune".  It's not
necessarily the political entity that would set up the network and the
interconnection, it can be a non-profit organization, or a local
company (owned by local share holder who are at the same time the
customers), etc.


In any case, the network effect should act at least as well as for
roads.  It's true that you can see some strange things on communal
roads sometimes (eg. a road well maintained half way, and unmaintained
for the rest between two villages).

Here, my bet is that routing algorithms will be able to flood packets
well enough that in general no link would be too predominant, so you
would have an interest in maintaining links with all your neighbors.



> Or, when you reach the county
> limits and the nice fiber goes into a mud-filled box and out comes
> 1930's phone wire?

Well, I'm not sure country rats are more eager to eat fiber than city
rats.  After all, in the country rats have better things to eat.

That said, hetzian links may be more economical.  



> I'm being somewhat the devil's advocate here, but there are certain
> economies of scale which benefit the current topology.

Indeed, this network would cost more than a big for-profit star network.

But it would have a lot of advantages, and they are worth the cost:

- less censorship,

- more coverage (big company only cover areas that are profitable,
  ie. big cities.  The country side is left with third world level
  technology).

- more resilience (since it would be more meshed, and a more local
  scale).




> I do agree though, that once a star-like system is in place, using it 
> as leverage against the customers is easy and simply will happen....
>
> -pete

-- 
__Pascal Bourguignon__                     http://www.informatimago.com/

On Mon, 23 Aug 2010 23:05:03 +0300, "Captain Obvious"
<udodenko@users.sourceforge.net> wrote:

>I often run Linux servers without any swap. I've found that even if working 
>set fits into RAM swapping might introduce unpredictable performance 
>problems (especially if GC is involved), so my policy is -- either it fits 
>in RAM or we need to buy more RAM.

I understand your point but I don't agree.  It is *still* way too
expensive to buy enough RAM (or CPUs for that matter) for some
problems.  My Masters is in database modeling ... it doesn't matter
how much RAM you have, there will be somebody with a database many
times larger that will want to do some operation that is hard to do in
neat slices (e.g., multi-dimensional full outer joins).

GC doesn't play well with VMM - GC's normal access patterns are pretty
much the worst case for VMM page replacement policies.  There has been
much talk (but little action) about giving the user-space program
direct control over its page buffers - but GC/runtime developers don't
want responsibility for all paging in the program ... they want at
most a couple of page buffers for random walks through the backing
store.  Then there is the camp that wants GC built into the operating
system.  But nobody can agree on anything so nothing happens.


>Either there is little swapping which is not problem at all, or there is 
>heavy swapping and computer is unusable. 

There is also moderate swapping which is not a problem.


>I don't know any applications where swapping is a norm, and even if they
>exist, these applications themselves are not norm.

There are plenty of applications that have super large data sets and
irregular data access patterns.  Tera-scale databases are common in
industry and peta-scale and even exa-scale are not unusual in Fortune
500 companies.  Business intelligence apps that try to extract hidden
patterns from such databases routinely bump against the limits of
quite large machines.

There are shockingly large numbers of corporate IT people who would
tell you that a computer with 128GB of RAM (~$50,000+) is strained by
its typical use.  There are a fair number who can say the same about a
computer with 512GB of RAM ($150,000+).


> GN> it doesn't matter how much RAM you have.  Windows uses relocatable
> GN> DLLs as opposed to position independent ones, so a busy system
> GN> frequently has many copies of the same DLLs loaded simultaneously.
> GN> DLLs that are not mapped at their image address use the page file as
> GN> backing store.
>
>If it is the case then it is just like memory allocations in applications --  
>they are "backed by page file" too.  So it just means that whole need for
>memory is larger than it would be otherwise.
>So?

You seem to live in a world without technological or financial
constraints.  I'm very happy for you, but keep in mind that not
everyone lives there.  Any time there is an application that is bigger
than the computing budget, there will be swapping.

George

George Neuner  <gneuner2@comcast.net> wrote:
+---------------
| There are shockingly large numbers of corporate IT people who would
| tell you that a computer with 128GB of RAM (~$50,000+) is strained by
| its typical use.  There are a fair number who can say the same about a
| computer with 512GB of RAM ($150,000+).
+---------------

Which is why SGI is *still* selling cc-NUMA machines with *terabytes*
of main RAM:

    http://www.sgi.com/products/servers/altix/uv/
    ...
    Altix(r) UV scales to extraordinary levels-up to 2,048 cores
    (256 sockets) with architectural support to 262,144 cores (32,768
    sockets). Support for up to 16TB of global shared memory in a
    single system image, enables Altix UV to remain highly efficient
    at scale for applications ranging from in-memory databases, to a
    diverse set of data and compute-intensive HPC applications.

But your main point is still valid: With surveillance [ELINT],
oil & gas, space imagery, multimedia, and other large datasets
becoming so common [some customers at a PPoE routinely threw
around numbers in the large PB range!], even 16 TiB of RAM isn't
"enough" for some apps.


-Rob

-----
Rob Warnock			<rpw3@rpw3.org>
627 26th Avenue			<URL:http://rpw3.org/>
San Mateo, CA 94403		(650)572-2607

On 8/24/2010 9:45 PM, Rob Warnock wrote:
> George Neuner<gneuner2@comcast.net>  wrote:
> +---------------
> | There are shockingly large numbers of corporate IT people who would
> | tell you that a computer with 128GB of RAM (~$50,000+) is strained by
> | its typical use.  There are a fair number who can say the same about a
> | computer with 512GB of RAM ($150,000+).
> +---------------
>
> Which is why SGI is *still* selling cc-NUMA machines with *terabytes*
> of main RAM:
>
>      http://www.sgi.com/products/servers/altix/uv/
>      ...
>      Altix(r) UV scales to extraordinary levels-up to 2,048 cores
>      (256 sockets) with architectural support to 262,144 cores (32,768
>      sockets). Support for up to 16TB of global shared memory in a
>      single system image, enables Altix UV to remain highly efficient
>      at scale for applications ranging from in-memory databases, to a
>      diverse set of data and compute-intensive HPC applications.

Finally a platform that can support Microsoft Tic-Tac-Toe!

kt

-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

 KT> On 8/24/2010 9:45 PM, Rob Warnock wrote:
 ??>> George Neuner<gneuner2@comcast.net>  wrote:
 ??>> +---------------
 ??>> | There are shockingly large numbers of corporate IT people who would
 ??>> | tell you that a computer with 128GB of RAM (~$50,000+) is strained
 ??>> by | its typical use.  There are a fair number who can say the same
 ??>> about a | computer with 512GB of RAM ($150,000+). +---------------
 ??>>
 ??>> Which is why SGI is *still* selling cc-NUMA machines with *terabytes*
 ??>> of main RAM:
 ??>>
 ??>>      http://www.sgi.com/products/servers/altix/uv/
 ??>>      ...
 ??>>      Altix(r) UV scales to extraordinary levels-up to 2,048 cores
 ??>>      (256 sockets) with architectural support to 262,144 cores (32,768
 ??>>      sockets). Support for up to 16TB of global shared memory in a
 ??>>      single system image, enables Altix UV to remain highly efficient
 ??>>      at scale for applications ranging from in-memory databases, to a
 ??>>      diverse set of data and compute-intensive HPC applications.

 KT> Finally a platform that can support Microsoft Tic-Tac-Toe!

Hmm, I can see neither George's nor Rob's messages. Usenet WTF. 

On 2010-08-25 00:20:44 +0100, George Neuner said:

> I understand your point but I don't agree.  It is *still* way too
> expensive to buy enough RAM (or CPUs for that matter) for some
> problems.  My Masters is in database modeling ... it doesn't matter
> how much RAM you have, there will be somebody with a database many
> times larger that will want to do some operation that is hard to do in
> neat slices (e.g., multi-dimensional full outer joins).

The solution to that is not generic swap but special "swap" managed by 
the database engine.  Which, unsurprisingly, is what commercial 
databases do.

> There are shockingly large numbers of corporate IT people who would
> tell you that a computer with 128GB of RAM (~$50,000+) is strained by
> its typical use.  There are a fair number who can say the same about a
> computer with 512GB of RAM ($150,000+).

I work in those places.  We would worry if systems *ever* touched 
generic swap space because the performance impact of that is generally 
utterly catastophic.  We have things wired to the page scanner which 
set sirens off if the rate ever goes above zero.

(There are design deficiencies in OSs which often require lots of swap 
space to be configured - for instance Solaris has some oddities around 
its shared-memory implementation which can cause it to require a lot of 
swap space - but that space is never touched.  Other than these design 
issues, the reason for swap is to provide space for crash dumps.)

On 2010-08-25 02:45:09 +0100, Rob Warnock said:

> 
> But your main point is still valid: With surveillance [ELINT],
> oil & gas, space imagery, multimedia, and other large datasets
> becoming so common [some customers at a PPoE routinely threw
> around numbers in the large PB range!], even 16 TiB of RAM isn't
> "enough" for some apps.

But the solution to that problem (at least in my experience) is not 
"just let it page by adding a bunch of swap space" it is "make it page 
in a very structured way".

George Neuner <gneuner2@comcast.net> writes:

> You seem to live in a world without technological or financial
> constraints.  I'm very happy for you, but keep in mind that not
> everyone lives there.  Any time there is an application that is bigger
> than the computing budget, there will be swapping.

	Alternatively, in a world without sufficiently large problem
sets... these days, a large number of programmers have no understanding
of the fact that you sometimes have to worry about having insufficient
amounts of memory, CPU power, bandwidth, diskspace etc. These people
often know *at most* one programming language - commonly Java.

On Aug 24, 10:52=A0pm, "Captain Obvious"
<udode...@users.sourceforge.net> wrote:
> =A0KT> On 8/24/2010 9:45 PM, Rob Warnock wrote:
> =A0??>> George Neuner<gneun...@comcast.net> =A0wrote:
> =A0??>> +---------------
> =A0??>> | There are shockingly large numbers of corporate IT people who w=
ould
> =A0??>> | tell you that a computer with 128GB of RAM (~$50,000+) is strai=
ned
> =A0??>> by | its typical use. =A0There are a fair number who can say the =
same
> =A0??>> about a | computer with 512GB of RAM ($150,000+). +--------------=
-
> =A0??>>
> =A0??>> Which is why SGI is *still* selling cc-NUMA machines with *teraby=
tes*
> =A0??>> of main RAM:
> =A0??>>
> =A0??>> =A0 =A0 =A0http://www.sgi.com/products/servers/altix/uv/
> =A0??>> =A0 =A0 =A0...
> =A0??>> =A0 =A0 =A0Altix(r) UV scales to extraordinary levels-up to 2,048=
 cores
> =A0??>> =A0 =A0 =A0(256 sockets) with architectural support to 262,144 co=
res (32,768
> =A0??>> =A0 =A0 =A0sockets). Support for up to 16TB of global shared memo=
ry in a
> =A0??>> =A0 =A0 =A0single system image, enables Altix UV to remain highly=
 efficient
> =A0??>> =A0 =A0 =A0at scale for applications ranging from in-memory datab=
ases, to a
> =A0??>> =A0 =A0 =A0diverse set of data and compute-intensive HPC applicat=
ions.
>
> =A0KT> Finally a platform that can support Microsoft Tic-Tac-Toe!
>
> Hmm, I can see neither George's nor Rob's messages. Usenet WTF.

Microsoft might have a hand in this. Kenny told me they quite
powerful.

 Xah

On 8/25/2010 4:38 PM, Xah Lee wrote:
> On Aug 24, 10:52 pm, "Captain Obvious"
> <udode...@users.sourceforge.net>  wrote:
>>   KT>  On 8/24/2010 9:45 PM, Rob Warnock wrote:
>>   ??>>  George Neuner<gneun...@comcast.net>    wrote:
>>   ??>>  +---------------
>>   ??>>  | There are shockingly large numbers of corporate IT people who would
>>   ??>>  | tell you that a computer with 128GB of RAM (~$50,000+) is strained
>>   ??>>  by | its typical use.  There are a fair number who can say the same
>>   ??>>  about a | computer with 512GB of RAM ($150,000+). +---------------
>>   ??>>
>>   ??>>  Which is why SGI is *still* selling cc-NUMA machines with *terabytes*
>>   ??>>  of main RAM:
>>   ??>>
>>   ??>>        http://www.sgi.com/products/servers/altix/uv/
>>   ??>>        ...
>>   ??>>        Altix(r) UV scales to extraordinary levels-up to 2,048 cores
>>   ??>>        (256 sockets) with architectural support to 262,144 cores (32,768
>>   ??>>        sockets). Support for up to 16TB of global shared memory in a
>>   ??>>        single system image, enables Altix UV to remain highly efficient
>>   ??>>        at scale for applications ranging from in-memory databases, to a
>>   ??>>        diverse set of data and compute-intensive HPC applications.
>>
>>   KT>  Finally a platform that can support Microsoft Tic-Tac-Toe!
>>
>> Hmm, I can see neither George's nor Rob's messages. Usenet WTF.
>
> Microsoft might have a hand in this. Kenny told me they quite
> powerful.

No, they are all busy getting their resumes out. No one there has a clue 
what to do with smartphones, tablets, or even the web--it is only a 
matter of time and they know it. They lost control of the platform stack 
and of course all they had was that monopoly...game over.

kt
-- 
http://www.stuckonalgebra.com
"The best Algebra tutorial program I have seen... in a class by itself." 
Macworld

Tim Bradshaw <tfb@tfeb.org> wrote in news:i4gdvv$13b$1@news.eternal-
september.org:

> I think it's what happens to monopolists: they get big by being a 
> monopolist, and because they have no competition they get slow and 

Big software companies have an advantage which makes them more likely to 
succeed in spite of relative incompetence.  Instead of "monopoly" we 
should think of a better word that means the company depends on the 
government to get its power and money; and from the power given by the 
government, grows powerful enough to thwart the competition.

The power given by software copyright law doesn't just increase with the 
amount sold.  It increases more than linearly.  The more the software is 
sold, the more people need it, to be compatible with it.  It gives the 
company not only the money from selling the software, but also gives them 
a kind of part ownership of the users, as if they were slaves.  Without 
the government, a big and relatively incompetent software company could 
not exist.  They depend on the government to enforce their ownership of 
their users, by enforcing the non-linear power of copyright law.

If the copyright law were changed to make it easier for competitors to 
sell products more compatible with the products of the big software 
companies, it might solve the problem.  If a company could sell an off-
brand Windows 7, with new features to make it better than Windows 7, 
Microsoft would have competition in its own market.  But presently 
Microsoft would sue for copyright violation, for copying the "look and 
feel" of Windows 7.

x <a@b.c> writes:

> Tim Bradshaw <tfb@tfeb.org> wrote in news:i4gdvv$13b$1@news.eternal-
> september.org:
>
>> I think it's what happens to monopolists: they get big by being a 
>> monopolist, and because they have no competition they get slow and 
>
> Big software companies have an advantage which makes them more likely to 
> succeed in spite of relative incompetence.  Instead of "monopoly" we 
> should think of a better word that means the company depends on the 
> government to get its power and money; and from the power given by the 
> government, grows powerful enough to thwart the competition.
>
> The power given by software copyright law doesn't just increase with the 
> amount sold.  It increases more than linearly.  The more the software is 
> sold, the more people need it, to be compatible with it.  It gives the 
> company not only the money from selling the software, but also gives them 
> a kind of part ownership of the users, as if they were slaves.  Without 
> the government, a big and relatively incompetent software company could 
> not exist.  They depend on the government to enforce their ownership of 
> their users, by enforcing the non-linear power of copyright law.
>
> If the copyright law were changed to make it easier for competitors to 
> sell products more compatible with the products of the big software 
> companies, it might solve the problem.  If a company could sell an off-
> brand Windows 7, with new features to make it better than Windows 7, 
> Microsoft would have competition in its own market.  But presently 
> Microsoft would sue for copyright violation, for copying the "look and 
> feel" of Windows 7.

There are countries (eg. France) where the law allows for example
reverse engineering with the purpose of interoperativity.  If the
format of MS-Word documents is not published (or what is published is
incomplete) then any competitor can in France reverse engineer MS-Word
to be able to develop a MS-Word document reader and write to
interoperate with MS-Word.


I don't know if that laws covers user interfaces too, but for
sufficiently complex user interface, I would say that it's a matter of
interoperativity to be able to reproduce the user interface, so that
users are compatible with your new software ;-)


-- 
__Pascal Bourguignon__                     http://www.informatimago.com/

 ??>> I often run Linux servers without any swap. I've found that even if
 ??>> working set fits into RAM swapping might introduce unpredictable
 ??>> performance problems (especially if GC is involved), so my policy is
 ??>> -- either it fits in RAM or we need to buy more RAM.

 GN> I understand your point but I don't agree.  It is *still* way too
 GN> expensive to buy enough RAM (or CPUs for that matter) for some
 GN> problems.

It is expensive, yes, but are HDDs a solution? At 100 iops per spindle they 
are definitely not a replacement for RAM.
And I guess you need to change algorithms to make writes sequential.
If you have sequential writes and reads and you need a lot of them, then 
maybe HDD is a right solution.
But it is definitely not a common case. Right tool for the purpose.

My opinion is that SSD is better for usual, average cases if we ignore the 
fact that they are way too expensive.

 GN>   My Masters is in database modeling ... it doesn't matter how much RAM
 GN> you have, there will be somebody with a database many times larger that
 GN> will want to do some operation that is hard to do in neat slices (e.g.,
 GN> multi-dimensional full outer joins).

But you don't rewrite your database each second, do you? Records are added, 
some are updated, but most likely insert/update rates are limited, and so 
are I/O rates, so it won't be a problem.
Unless it is insanely dynamic database where you update it all each hour.

 GN> GC doesn't play well with VMM - GC's normal access patterns are pretty
 GN> much the worst case for VMM page replacement policies.

Worst for HDD's limited iops too. When I had swap enabled I was seriously 
considering adding code to do madvise(..., MADV_WILLNEED) (or even mlockall) 
before GC because I had abysmal GC times even having enough RAM.

 GN>   There has been much talk (but little action) about giving the
 GN> user-space program direct control over its page buffers - but
 GN> GC/runtime developers don't want responsibility for all paging in the
 GN> program ... they want at most a couple of page buffers for random walks
 GN> through the backing store.  Then there is the camp that wants GC built
 GN> into the operating system.  But nobody can agree on anything so nothing
 GN> happens.

I think madvise() could be a solution if it worked right. E.g. GC code 
should call madvise() on a region currently being scanned and kernel should 
try to load it as sequential as possible.
I don't know whether it works now... Looking for couple of hours into Linux 
kernel I came to conclusion that it might work, but nothing definitive.

 ??>> Either there is little swapping which is not problem at all, or there
 ??>> is heavy swapping and computer is unusable.

 GN> There is also moderate swapping which is not a problem.

I don't know how your moderate swapping is different from my "little 
swapping" quantatively.
But probably SSDs can cope with it.

 ??>> I don't know any applications where swapping is a norm, and even if
 ??>> they exist, these applications themselves are not norm.

 GN> There are plenty of applications that have super large data sets and
 GN> irregular data access patterns.  Tera-scale databases are common in
 GN> industry and peta-scale and even exa-scale are not unusual in Fortune
 GN> 500 companies.  Business intelligence apps that try to extract hidden
 GN> patterns from such databases routinely bump against the limits of
 GN> quite large machines.

Again it is worth to look into patterns. If you load dataset, say, once a 
day, and then only perform reads then SSD would be ideal.

 GN> There are shockingly large numbers of corporate IT people who would
 GN> tell you that a computer with 128GB of RAM (~$50,000+) is strained by
 GN> its typical use.  There are a fair number who can say the same about a
 GN> computer with 512GB of RAM ($150,000+).

If they know what is their typical use then they make a decision what 
storage to buy.

OTOH SOHO users won't have this problem to begin with.

 GN> You seem to live in a world without technological or financial
 GN> constraints.

No, in context of Kenny's questions I was thinking more about desktop users 
rather than OLAP and data mining uses.
I don't think that desktop users are currently pushing boundaries...

On Aug 22, 9:10=A0am, "Captain Obvious" <udode...@users.sourceforge.net>
wrote:

> =A0 =A0 =A0You can see how math symbols are being loaded, WTF? Is it an a=
nimation
> of some sort?
> =A0 =A0 =A0It feels like application is half-backed.

half-baked

 ??>>      You can see how math symbols are being loaded, WTF? Is it an
 ??>> animation of some sort?      It feels like application is half-backed.

 wx> half-baked

I had just one typo in the whole text? Holy crap, that's impressive!