Linux-Advocacy Digest #616, Volume #31 Sat, 20 Jan 01 16:13:04 EST
Contents:
Re: Windows curses fast computers ("Mike")
Re: What really burns the Winvocates here... (J Sloan)
Re: A salutary lesson about open source ("Chad Myers")
Re: A salutary lesson about open source ("Chad Myers")
Re: Linux 2.4 Major Advance ("Chad Myers")
Re: A salutary lesson about open source (J Sloan)
Re: Linux is crude and inconsistant (Lewis Miller)
Re: Multiple standards don't constitute choice (The Ghost In The Machine)
Re: *** THE PROPOSAL *** (Charlie Ebert)
Re: Windows curses fast computers (Bob Hauck)
Re: Linux is crude and inconsistant ("Ayende Rahien")
Re: New Microsoft Ad :-) (The Ghost In The Machine)
----------------------------------------------------------------------------
From: "Mike" <[EMAIL PROTECTED]>
Subject: Re: Windows curses fast computers
Date: Sat, 20 Jan 2001 20:11:16 GMT
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
news:je2a6.916$[EMAIL PROTECTED]...
...
> The problem was not "screwing with real/protected mode". The problem was
> the computer didn't give the drives (with large caches) enough time to
> completely write out their data before shutting down.
>
> Actually, I think this *IS* a fault of the drive. The drive should hold
> enough capacitance to finish writing out it's cache and then park, but
> aparently the drive doesn't do this.
Yikes! Capacitance to act as a mini-UPS for a disk drive?
A modern desktop drive consumes something in excess of 5W when reading and
writing, and takes somwhere in excess of 1A of current. Most of the chips
can tolerate somewhere between 5% and 10% power supply drop before they fall
out of spec. But that's how the power supply is spec'd _before_ the power is
cut off. The power supplied to the drive may already be at the low end of
the spec when the drive is powered down. There are lots of additional
complicating factors here, including the reduced supply voltages (typically
3.3V or 1.8V) for many of the chips. These chips may have additional margin
as the 5V supply drops.
When the supply drops, the capacitance needed to keep the drive running in
spec for the time it takes to shut down would be larger than the drive
itself, since the voltage across the capacitor would have to remain
virtually unchanged until the cache was flushed. That translates into a BIG
capacitor. How big depends on just how much risk the drive vendor wants to
take. Will things fail if the supply drops another 5%? Or does it have to be
held to 1%? Or less?
A better approach would be to try to use a much smaller capacitor and feed
it into a switching regulator, which would be capable of holding a constant
output voltage as the capacitor voltage driving it drops. There's probably
already a switcher on the drive to supply power to the controller and
channel chips, but the preamp/write driver generally requires a 5V supply,
as does the motor and motor speed control, and the head positioning servo
system. You'd have to add another switcher to supply 5V to those components
when the power fails. The additional switcher adds more cost. In high
performance drives, the motor requires 12V. You'd need to add another
switcher to handle that. High performance drives also take alot more than
the 5W of power that a typical desktop drive might consume.
You may also have another problem: the supply to your disk drive is shared
with other components. If you put a supply in your drive to keep it running
when the power goes down, then it may also have to supply power back to
everything else that's sharing the same supply. So, you can add a relay, and
try to open it when the power goes down. That's a problem because you have
to sense when the power has turned off, and you can't sense a voltage that's
above the spec minimum, and there's no extra margin in the spec. In other
words, by the time you know the power is too low, it's already too low. And,
relays aren't exactly the most reliable things when you're trying to protect
something against multiple g-force shocks. You could try a MOS switch, but
then you get voltage drop across the switch, leading to supply problems even
in normal operation.
Let's say, though, that you could solve all those problems, and all you
needed was a capacitor to hold the energy. Let's also say that your switcher
could operate reliably with 1.8V coming in, and that you charge your
capacitor to 5V off the main PC supply, and that it takes 5 seconds to flush
the cache. If the drive takes 5W off the 5V supply, you'll need to supply 1A
for 5 seconds. Your capacitor is allowed to drop 3.2V, from 5V to 1.8V,
during that 5 seconds, before the drive fails. Assuming that the current
drained off the capacitor stayed linear (it would actually be increasing as
the voltage dropped so that constant power could be supplied at 5V), you'd
need 1A*5sec/3.2V = 1.56Farads of capacitance. A quick search turned up some
100,000uF electrolytic jobs for around $5 each in large quantities. You need
16 of them, and they'll fit very nicely in a space that's roughly 6.25" x
6.25" x 3". I suspect that's going to play hell with the 3.5" form factor,
as well as add lots of money to the price of a drive.
That's pretty extreme, and one problem with my analysis is that I really
don't know how long it takes to flush the cache. But if the cache is fairly
large, and fairly full, and contains chunks of lots of different files, then
there's going to be lots of seeking going on as each chunk is written. Even
if the worst case was only 0.5 second, you'd still need lots of capacitance.
A single 0.1F capacitor is around 1.5" in diameter, and 3" tall. It would
still add significantly to the cost of the drive, and would still change the
form factor. It would be a mess.
You could try some other things. One common trick is to use the stored
energy in the motor to park the head after the power is removed. You can't
use this approach to run the drive, though, because writing data depends on
having the motor speed precisely controlled. If you're using the stored
energy in the motor to do anything, it's no longer running at a precisely
controlled speed. That's fine for parking the head, but really bad if you're
writing data.
So, I don't think this is a disk drive problem in the same sense that you
do. I think the main problem is the ambiguity of the specification. Varying
interpretations of the specification mean that there is no single effective
way to flush the cache and power down a drive reliably. As convenient as it
may be for some to blame Microsoft, all it may take for Linux to die is a
drive vendor coming out with a new drive with a different interpretation of
the specs, and different requirements for flushing the cache safely.
-- Mike --
------------------------------
From: J Sloan <[EMAIL PROTECTED]>
Subject: Re: What really burns the Winvocates here...
Date: Sat, 20 Jan 2001 20:13:57 GMT
Pete Goodwin wrote:
> J Sloan wrote:
>
> > > Yes, and Linuxconf is a Gtk tool, not a Qt tool.
> >
> > Wrong - linuxconf is a system tool.
>
> It uses the Gtk library. How am I wrong? Does it use the Qt library? No it
> doesn't.
Nope - it uses neither.
Only "gnome-linuxconf" uses any X libraries
>
>
> > You can access it through a web interface.
>
> I tried that, it died for some reason.
Well "died" is unlikely - more likely you didn't enable web access.
> > You can access it through a color terminal interface.
>
> Yes I tried that too. Unfortunately, it obscures any error message it
> reports making it a less than useful system tool.
I don't see that problem here.
Perhaps you are telneting in from a windows pc?
That would explain broken terminal emulation.
> Besides, using a CLI tool on a GUI desktop is no better than using multiple
> CLI's. That's not a GUI at all.
AFAIK there is no cli interface for linuxconf.
a curses app is a full screen interface, unlike the
"one shot" cli commands
jjs
------------------------------
From: "Chad Myers" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: A salutary lesson about open source
Date: Sat, 20 Jan 2001 19:59:13 GMT
"Cliff Wagner" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Sat, 20 Jan 2001 15:53:43 GMT, Chad Myers typed something like:
>
> >> >> Please provide proof of this statement.
> >> >> From my experience, most "My Cat Fluffy" sites are hosted
> >> >> on places like geocities and homestead and places
> >> >> like that because people generally don't want to
> >> >> pay money to host something so inane.
> >> >
> >> >If you compare surveys from other parties (besides Netcraft), they
> >> >mostly survey Fortune500, Global500, etc. Those numbers, IIS is
> >> >in the lead or closely follows iPlanet and Apache is far behind.
> >> >
> >> >Netcraft is the only survey where Apache leads.
> >> >
> >> >http://www.biznix.org/surveys/
> >> >
> >> >Netcraft doesn't differeniate between corporate and personal
> >> >sites. It also counts each virtual host on a hosting provider.
> >> >
> >> >The numbers are grossly inflated for Apache. All Netcraft's
> >> >numbers tell us is that Apache is the choice for hosting providers,
> >> >which we already know, so it doesn't really give us anything.
> >> >
> >> >As far as geocities and homestead, there are still many people
> >> >who purchase domain names for personal sites or family web sites.
> >> >Aside from that, many non-profit organizations, clubs, and
> >> >other small organizations have web sites.
> >> >
> >> >The people who have high traffic and who have high demand use
> >> >iPlanet and IIS. The people who show pictures of their family
> >> >or who post meeting calendars for the local VFW use hosted
> >> >Apache virtual hosts.
> >> >
> >> >-Chad
> >> >
> >> By your claim Fortune 500 = top web site.
> >> Someone has already refuted this claim, so I won't
> >> waste my breath.
> >
> >I don't recall saying that Fortune 500 = top web site. Nor do
> >I see that anywhere in my previous post. Please show us where
> >you see this.
> >
> >It's my contention that Fortune 500 sites receive MORE hits
> >and visits than does your average non-profit or personal site.
> >Wouldn't you agree?
>
> Well, duh, that isn't the issue here. You're arguing
> about IIS vs. Apache,
No, I'm not. I'm arguing that the Netcraft numbers are
largely irrelevant. Please participate in the discussion,
or go elsewhere.
> and your only meaningful stat relates to fortune 500,
> making the assertion that
> the netcraft stats are skewed due to "my cat fluffy"
> sites, yet never have backed up that claim with any
> assertion other then "I said so" and "look at the
> Fortune 500"...well, gee, we look at the fortune 500,
> and that covers, well, gee, 500 companies. That
> doesn't mean they are all meaningful players in the
> web business.
I have backed up my claim. I have shown that the web
server stats for Fortune 500 companies varies greatly with
the netcraft numbers. Why is there such a large discrepency?
a.) Netcraft's survey method is grossly inaccurate and/or
unscientific
b.) Netcraft includes every site (including every
invidiual virtual host) which leads to misleading numbers
due to the large amount of low-traffic unimportant sites.
So, if b is the case (which is my contention), and Apache is
the winner, so what? Apache can be the king of low-traffic sites,
I don't care. I want to see what server the people who have large
amounts of money riding on their choice are choosing. And by
and large they're choosing IIS and iPlanet.
It's simple facts, I don't understand why you guys have such a
hard time with it.
-Chad
------------------------------
From: "Chad Myers" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: A salutary lesson about open source
Date: Sat, 20 Jan 2001 20:00:51 GMT
"J Sloan" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
> Chad Myers wrote:
>
> > There are many more Apache's out there than IIS, I agree, but so what?
> >
> > If I set up 15 boxes with Apache and they serve 1 request a day, so what?
>
> That would be silly, you might as well run a pc web server
> like iis or even pws if you get 1 request a day.
Exactly, it doesn't matter one way or the other, right? This is where
Apache is the king, because it's free and doesn't matter. All the numbers
point to this: large numbers of installations of Apache, but little to
show for it.
> > If there's a Fortune 500 company whose business depends on their web site,
> > or a significant part of it, and they choose IIS, this means something.
>
> and if they choose apache it means just as much.
It would, but they aren't choosing apache, they're choosing IIS and iPlanet.
Why? Because they're better. Thank you for making my point for me.
-Chad
------------------------------
From: "Chad Myers" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 Major Advance
Date: Sat, 20 Jan 2001 20:01:30 GMT
<[EMAIL PROTECTED]> wrote in message
news:94cfpp$jo9$[EMAIL PROTECTED]...
> "Chad Myers" <[EMAIL PROTECTED]> writes:
>
> >Not really. The only benchmark I've seen Linux win was with a web server
> >that no one uses. One benchmark. Please show me ones where Linux wins
> >(oh yeah, and the FUD ones from c't don't count, only major reputible
> >companies with standardized benchmarks, not grudges against Microsoft).
>
> So c't, who has a Spec license (can Mindcraft say that?) and comes from
> the same people who, in a magzine called "ix" extensively cover Windows NT,
> now has "grudges against Microsoft"?
>
> Maybe you should, just for a change, *read* the magazine you are criticizing?
Please show me an article in c't that is favorable to Microsoft.
Just one.
Thank you.
-Chad
------------------------------
From: J Sloan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: A salutary lesson about open source
Date: Sat, 20 Jan 2001 20:29:45 GMT
Chad Myers wrote:
> "J Sloan" <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
>
> > That would be silly, you might as well run a pc web server
> > like iis or even pws if you get 1 request a day.
>
> Exactly, it doesn't matter one way or the other, right? This is where
> Apache is the king, because it's free and doesn't matter. All the numbers
> point to this: large numbers of installations of Apache, but little to
> show for it.
And you carefully snipped my references to amazon, google, deja,
which, like yahoo and other large sites, use apache.
Nice try kid, but you're fighting a losing battle and we all know it.
jjs
------------------------------
From: [EMAIL PROTECTED] (Lewis Miller)
Crossposted-To: alt.linux.sux
Subject: Re: Linux is crude and inconsistant
Date: 20 Jan 2001 20:37:52 GMT
Aaron R. Kulkis was heard ranting about <[EMAIL PROTECTED]> in
alt.linux.sux on 19 Jan 2001
>Lewis Miller wrote:
>>
>> Aaron R. Kulkis was heard ranting about <[EMAIL PROTECTED]>
>> in alt.linux.sux on 16 Jan 2001
>>
>> >Kyle Jacobs wrote:
>> >>
>> >Translation...
>> >the $200 product [Lose9x] crashes every couple of hours
>> >the $1000 product [LoseNT] crashes only once/week.
>> >The $25 product [Linux] will stay up for months.
>> ^^^
>> $25?!?! Damn you're gettin ripped off.. last I checked Linux was
>> free. :) I've never paid for it. Even my CD versions.
>
>It's nice having the manuals around, for distribution quirks/features
>so that you can loan it to a friend with some degree of confidence that
>they will return your linux disks within a reasonable amount of time.
Manuals? Last I checked they were on the CD.
>> Besides forget NT get 2000 if you're going to run a Windows box.
>
>Why would i want to do something as stupid as that?
Um, because 2000 supports a hell of a lot more hardware than NT. 2000 is a
lot more stable than NT. 2000 has more tools than NT.
--
l8r
-LJM
a.k.a. Jaster Mereel
a.k.a. MrBobaFett
"Little things used to mean so much to Shelly. I used to think
they were kind of trivial. Believe me, nothing's trivial. "
-- Eric Draven, The Crow
------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Subject: Re: Multiple standards don't constitute choice
Date: Sat, 20 Jan 2001 20:41:36 GMT
In comp.os.linux.advocacy, Pete Goodwin
<[EMAIL PROTECTED]>
wrote
on Sat, 20 Jan 2001 10:46:14 +0000
<LZda6.184771$[EMAIL PROTECTED]>:
>Everyone goes on about how Linux offers me the 'choice' of which desktop I
>can use, unlike Windows. However, choice here does not equate to consistant
>style.
>
>If I want all my file save/open dialogs to all look the same - like the KDE
>style, or MOTIF or Gtk, can I do that with the Linux desktop? No I can't -
>my choice is restricted here to whatever toolktip the application is
>created with.
>
>If I restrict myself to KDE only applications then I lose certain system
>configuration tools as there isn't one written for KDE (that's certainly
>true of the Mandrake distribution). Linuxconf is one example, it can run in
>text mode or GUI - but uses the Gtk toolkit.
>
>It is true that on Windows, application do use different styles of file
>open/save dialogs - however, there is a system wide _standard_ that 99% of
>applications use. Unfortunately, you can't change this standard - like have
>different shapes buttons etc. (and this is what I would call a "choice" -
>not the varying standards Linux offers).
There are also other requesters, such as color pickers and print setup items.
In fact, I've written a test jig -- cf. www.winehq.com, where you'll
see it in $WINEROOT/programs/cmdlgtst -- which exemplifies the
7 major dialogs: Open, Save, Color, Font, Find, Replace, and Print.
(One could quibble about the first two, admittedly.) These are
the set of dialogs implemented by Microsoft's COMMDLG, and
the code is available for browsing at
http://www.winehq.com/source/programs/cmdlgtst/?v=wine20001202
It's not horribly pretty, but test jigs tend not to be. :-)
But this gives people an idea of what it takes to use Microsoft
COMMDLG stuff (it's not bad, actually -- just clunky and overly
complicated). If I get time, I might post a comparison between
using the Microsoft open or save dialog, versus Motif's widget,
which I don't have handy right now because I don't have LessTif
installed.
It should build on NT or Win9x, and it also builds on WinE.
(I used an old Borland compiler, which may create issues.)
I can't contrast it with QT or GTK's file picker right now, or with
Motif's (my system disk died, and I'm building a new one; right now
I'm on an old (5.2) version of RedHat which is what I had handy).
Microsoft probably already has newer stuff sitting on top of this,
anyway. For instance, ADO has some wonderful stuff that can display
as a table the results of an arbitrary SQL query ultimately
fetched from ODBC -- actually, results from any data producer.
Whether it's usable by mere mortals is not clear to me, however.
I can get it to work in VC++; that's all I'll swear to. :-)
>
>--
>Pete, running KDE2 on Linux Mandrake 7.2
>
>PS: Now, before someone starts a discussion on semantics about how Linux is
>not X or KDE, when I say Linux, I mean "Linux + X + KDE" or "Linux + X +
>your favourite Window Manager".
Noted.
--
[EMAIL PROTECTED] -- insert random misquote here
EAC code #191 1d:19h:36m actually running Linux.
I'm here, you're there, and that's pretty much it.
------------------------------
From: [EMAIL PROTECTED] (Charlie Ebert)
Subject: Re: *** THE PROPOSAL ***
Reply-To: Charlie Ebert:<[EMAIL PROTECTED]>
Date: Sat, 20 Jan 2001 20:43:48 GMT
In article <[EMAIL PROTECTED]>, J Sloan wrote:
>Charlie Ebert wrote:
>
>> We have all read and heard the comments about RedHat's
>> beta release compiler which has trouble compiling applications
>> with except for the kernel {of course}
>
>Let's get the facts straight: The debate was over whether
>gcc-2.96 could compile the kernel - there was never any
>question that it compiles everything else with no problem.
>Just to be sure, Red Hat shipped "kgcc", an older gcc which
>has been known to create stable kernels for some time.
>
The problem was compiling X applications I believe.
There was never any doubt it could compile Kernels and I
said so. What 2.96 could not do was compile some X applications.
And your right, they shipped the kgcc also. Why.
>However, it has since become clear that gcc-2.96 compiles
>both 2.2 and 2.4 kernel without problems -
>
There was NEVER any doubt about this.
>> and also this new
>> worm {hole fixer worm} which was unleashed on all the RedHat
>> 6.2 and 7.0 customers just recently.
>
>dunno, I've not seen it on ANY of my servers -
>
I haven't either.
>> Everybody knows I've used Slackware, RedHat, Suse, Mandrake
>> and Debian. I've tried them all and I like Debian's
>> distribution the best for it's stability and quality and
>> it's huge list of supported hardware platforms.
>
>I've tried sls, slack, caldera, turbo, & mandrake. debian may
>be fine, I just haven't had time to look at it. I"ve run freebsd
>and netbsd too. not to mention Solaris, Irix, HP-UX and UW.
>
>The result? RH suits my needs beset.
>
>> We have two key players here.
>>
>> We have RedHat who has 80% of the Linux market or more.
>
>probably more like 50-60...
>
At most there are 4. The rest are just in there.
But the top 2 are Redhat, Debian.
>>
>> We have Debian who has the largest and most stable of
>> the Linux distributions.
>
>I'd say SuSE has the "largest".
>
Nope. It's Debian.
>> RedHat has the BEST marketing.
>
>I dunno about best, caldera's pretty good at that.
>
Okay. Let's say market-share.
>> Debian has the BEST OS.
>
>hmm, isn't it the same kernel?
>
If that's all there was to making a good OS
then let Microsoft install 2.4 NOW!
>> Leave Debian as the all volunteer effort and just contribute
>> to the system. RedHat would assist in the development using
>> the Debian infrastructure. RedHat could then offer two
>> seperate distributions to the public with little effort.
>
>If anyone could pull this off, it might be a good thing - but
>debian is so anti-commercial, and antagonistic to all other
>distros (I think of debian as the freebsd of Linux) that it would
>probably never work.
>
>Some consolidation is probably inevitable, but debian and rh
>are at the 2 ends of the spectrum.
>
>jjs
>
>
Oh, I totally agree with you. They are at extreme opposites of
the spectrum. But can there be nothing of Richard Stallmans
COMMUNITY concept? The GPL ensures you will work together anyway.
All I'm suggesting is let Debian do what it does best, package
management tools , integrating applications and testing software
AND let's RedHat to TIP development to feed into the machine
and market the product.
Both benefit from the relationship.
I am picking the BEST of both organizations.
The most agreeable 2 parts.
Debian has no marketing and want's a better install program.
RedHat wishes they had a bug free solid distribution so they
could persue what they always do, tip development. RedHat's
problem is they are integrating too much not-ready for prime
time stuff which hasn't been tested enough.
And I'm suggesting 2 distributions from this.
RedHat/Debian commercial-stable.
RedHat/Debian home-cutting edge.
Charlie
------------------------------
From: [EMAIL PROTECTED] (Bob Hauck)
Subject: Re: Windows curses fast computers
Reply-To: bobh{at}haucks{dot}org
Date: Sat, 20 Jan 2001 20:43:41 GMT
On Sat, 20 Jan 2001 09:58:14 -0500, mlw <[EMAIL PROTECTED]> wrote:
> What about all this makes it the hard disk's fault?
Nothing, really, but that doesn't matter. Having observed Erik for a
few years now, I can say that I don't think there has ever been a case
where he admitted that MS might have made a mistake. It is always:
Someone else's fault
-or-
Not really a mistake, just a misunderstanding
-or-
They did it for your own good
-or-
The user did something wrong
I agree with that last. The user messed up by buying the product in the
first place.
--
-| Bob Hauck
-| To Whom You Are Speaking
-| http://www.haucks.org/
------------------------------
From: "Ayende Rahien" <Please@don't.spam>
Crossposted-To: alt.linux.sux
Subject: Re: Linux is crude and inconsistant
Date: Sat, 20 Jan 2001 22:48:11 +0200
Reply-To: "Ayende Rahien" <Please@don't.spam>
"Lewis Miller" <[EMAIL PROTECTED]> wrote in message
news:94csv0$gac$[EMAIL PROTECTED]...
> Aaron R. Kulkis was heard ranting about <[EMAIL PROTECTED]> in
> >> Besides forget NT get 2000 if you're going to run a Windows box.
> >
> >Why would i want to do something as stupid as that?
>
> Um, because 2000 supports a hell of a lot more hardware than NT. 2000 is a
> lot more stable than NT. 2000 has more tools than NT.
>
To qoute Linux:
"It's better"
------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.nt.advocacy
Subject: Re: New Microsoft Ad :-)
Date: Sat, 20 Jan 2001 20:59:27 GMT
In comp.os.linux.advocacy, J Sloan
<[EMAIL PROTECTED]>
wrote
on Fri, 19 Jan 2001 07:52:17 GMT
<[EMAIL PROTECTED]>:
>JS PL wrote:
>
>> Easily. I just built a system last week. And it played an mp3 perfectly
>> while simultaneously copying 600mb worth of other mp3's from the cd drive to
>> a folder AND installing office 2000 from the other cd drive. Didn't skip a
>> beat. It was probably "accessing" the internet too, I forget.
>
>Sure, and I'll bet it cured your cancer too...
>
>Meanwhile, back in the real world, my friend just mentioned
>that he clicked on the icq button the other day and windows
>2000 spontaneously rebooted.
>
>Now, that's the windows we all know and love!
That's not a bug, that's, erm, an undocumented feature!
Yeah, that's it! :-)
To be fair, I've hung Netscape more times than I can count with
it chewing CPU (beware Java applets), but that's not quite the same
since a kill -9 killed it, and it may be a config problem since
the Netscape is newer than the Redhat 6.0 system I put it on.
Now that I'm rebuilding, I might be able to address this issue in
some form.
Others have noted that X can crash with Netscape -- but Linux
proper stays up. Give credit where credit's due, though; in
this case, it's X's fault, since X shouldn't really crash either,
even if given a totally stupid request over its socket. (I don't
have specific version numbers and such, though -- sorry.)
I haven't tried Win2k enough to comment much on its stability,
and in any event it's on a machine I only use for games and a
demo piece of software that crashes suspiciously because I
upgraded a video driver -- but the system didn't go down, just
that software demo. The rest of the time, it's powered down
doing nothing. (It's a long story.)
Last comment: a lot of things can be covered up using additional CPU.
I still remember trying and failing to play stereo MP3s on my Pentium 90;
no OS can compensate for lack of "engine power". (Of course,
Windows does a good job of sapping "engine power", not to mention
RAM and interrupt latency, with such ultra-necessary things as sliding
or vanishing menus and "create directory here" save requesters. Sigh.)
[rest snipped]
--
[EMAIL PROTECTED] -- insert random misquote here
EAC code #191 1d:21h:18m actually running Linux.
I'm here, you're there, and that's pretty much it.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
