Linux-Advocacy Digest #876, Volume #28 Mon, 4 Sep 00 00:13:04 EDT
Contents:
Re: Inferior Engineering of the Win32 Platform - was Re: Linsux as a desktop
platform (D. Spider)
Re: How low can they go...? (T. Max Devlin)
Re: How low can they go...? (T. Max Devlin)
Re: [OT] Public v. Private Schools (Rick)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (D. Spider)
Crossposted-To: comp.sys.mac.advocacy,comp.os.ms-windows.advocacy,comp.unix.advocacy
Subject: Re: Inferior Engineering of the Win32 Platform - was Re: Linsux as a desktop
platform
Date: Mon, 04 Sep 2000 03:56:51 GMT
It appears that on Mon, 4 Sep 2000 13:04:18 +1000, in
comp.os.linux.advocacy "Christopher Smith" <[EMAIL PROTECTED]> wrote:
>
>"D. Spider" <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> It appears that on Sat, 2 Sep 2000 13:32:18 +1000, in
>> comp.os.linux.advocacy "Christopher Smith" <[EMAIL PROTECTED]> wrote:
>>
>> >
>> >"D. Spider" <[EMAIL PROTECTED]> wrote in message
>> >news:[EMAIL PROTECTED]...
>> >> It appears that on Sat, 2 Sep 2000 10:58:52 +1000, in
>> >> comp.os.linux.advocacy "Christopher Smith" <[EMAIL PROTECTED]> wrote:
>> >>
>> >[chomp]
>> >
>> >> >> I've honestly never seen this happen. I've seen it *claimed* that it
>> >> >> had happened, but in every case where I've had the opportunity to
>> >> >> confirm the case myself, it was just a hung X server, and restarting
>X
>> >> >> solved the problem - i.e. the OS itself continued running without a
>> >> >> glitch the entire time.
>> >> >
>> >> >Not in this case. Although X is pretty as well.
>> >>
>> >> "X is pretty as well?"
>> >
>> >Sorry, "X is pretty unstable as well".
>>
>> Any GUI is relatively unstable - particularly on commodity hardware.
>> That said, in my experience X is far more stable than any variety of
>> MSWin.
>
>On what hardware ?
>
>On most resonably new videocards, X is quite unstable. Given video card
>vendors reluctance to hand out specifications for their cards, this is not
>entirely surprising. However, that doesn't change the facts.
I refuse to buy videocards that don't have good open source drivers
available, and I've never seen what you are describing. If you are
running proprietary binary-only drivers, then you have no grounds to
expect stability. Those are not supported or recommended.
Does this mean I may have to settle for a card that's 5% slower than
the one I would otherwise buy at times? Sure. Doesn't bother me one
bit.
>I've found NT to be much more stable than X. Win9x isn't, of course, but
>no-one seriously interested in stability is going to be running Win9x.
No one seriously interested in stability is going to be running
Windows anything :>
>
>> >> >To the end user the effect is the same. They lose all the data in
>> >whatever
>> >> >they were working in. The fact it takes 30 seconds to restart the X
>> >server
>> >> >instead of a minute to reboot is irrelevant - the end result is the
>same.
>> >>
>> >> Typically one loses a lot less even in this rare event.
>> >
>> >I wouldn't call it rare. I have the X server crap out at least once a
>> >month. That isn't rare, it's positively common.
>>
>> It is too common. I would take a long hard look at your setup if I
>> were you.
>
>The same hardware runs NT and FreeBSD fine. NT is notoriously fickle about
>the hardware it runs on, so I sincerely doubt the problem lies there.
It might not be the hardware. What distribution are you using? Did you
set up the X server yourself, or did you use some sort of auto-magical
setup? X is a mighty complicated system, with plenty of room for human
error. I know I had a lot of minor errors on a redhat box that was
auto-configured, but I've never seen anything like that on my
slackware box.
>> >On a desktop unix box, you'll most likely be using X, even if just to
>have a
>> >dozen XTerms open. Kill X, and everything goes with it. The difference
>> >between this and having The whole OS crash is largely semantic.
>>
>> I have to disagree. On a Workstation it is an inconvenience, but not
>> nearly so much as having the OS go down.
>
>Please explain why on a workstation, where almost all users typically run
>everything from inside X, why it is merely an inconvenience of a lesser
>scale than an OS crash.
I don't agree that "almost all users typically run everything from
inside X" - that is not my experience. Beyond that, restarting X is a
lot quicker than rebooting.
>
>> And as I keep repeating, I
>> was talking about Servers, not Workstations. A *nix Server doesn't
>> have to face any of this. An NT server does.
>
>An NT Server sitting at the login prompt doesn't have to, either,
>realistically speaking. The GUI isn't going to crash the machine if it
>isn't doing anything.
If you understood how computers run you would realise that is not
true. A loop looking for input and doing screen draws is NOT doing
nothing and it can interfere with other running programs and it can
crash.
>
>> >> *nix programs
>> >> generally autosave pretty frequently.
>> >
>> >So do Windows programs. Doesn't make much difference if you've had a fit
>of
>> >inspiration and written 400 words since the last autosave.
>>
>> By default, emacs autosaves every 300 keystrokes (not words.) That
>> setting is, of course, configurable.
>>
>> I can't remember ever needing that backup, however. The instability
>> you are describing is not characteristic at all. Again, I would urge
>> you to think about your setup. It's possible for an administrator to
>> make ANY system unstable.
>
>How long have you been using Linux ? It has been getting much more stable
>as of late (last year or so). I've only had two Linux crashes in the last
>twelve months, due to buggy drivers.
About a year, on my own hardware, about 5 years, on others peoples
hardware. And you are right, it has improved tremendously over that
time, if you were saying this 5 years ago I wouldn't have nearly as
much trouble believing it. 5 years ago I considered linux a neat toy,
but doubted that it would ever be ready to take over from my trusty
Novell server. It is ready for that now. NT still isn't.
>
>[chomp]
>> >It's happened to me about as often as NT crashing. That is, about 4
>times
>> >in the last 3-odd years. Linux crashes (as in the kernel) are somewhat
>more
>> >common - probably 3 or 4 times a year. FreeBSD is the only OS I've used
>> >that's never crashed on me, but I"ve only been using a machine with it
>> >running for a year or so.
>>
>> What on earth are you doing to your linux systems? I've seen a kernel
>> crash only once, and that was a result of damaged hardware. Now if you
>> are running dev kernels or something...
>
>Using them. Mainly in X, which is probably why it's "so unstable". This is
>only workstation and very basic user-mode development, I might add.
Hrmm well if you are running in X, with KDE or GNOME on a RedHat box
or something similar, that might happen. I couldn't say, I never felt
the need for all that fluff.
Running in console mode or with a stable X setup (I use slack 7,
upgraded to XFree 4.01, with WindowMaker for a window manager) I see
no instability at all.
>
>I'm willing to believe (in fact, I'm certain) the problems originate from
>buggy/beta drivers, but that's not the issue.
If you are using unsupported binaries of course that's an issue.
>
>Only ever 2.[0,2].x kernels.
>
>> BTW, avoiding crashes by rebooting your NT machines once a day is
>> cheating ;^)
>
>My NT machine used to get rebooted every fortnight or so (to Win95) so I
>could play games (much less frequent since Win2k), but the longest uptime I
>ever had was around the 90-day mark. I had to shutdown to move house. It
>was an emotional moment :).
>
>> >I might add that NT has IME recovered a lot more gracefully from sudden
>> >power outages (power is flaky around here) than Linux. I've never lost
>an
>> >NTFS filesystem, I've lost several ext2 ones.
>>
>> My experience has been exactly the opposite.
>
>Funny, that :).
>
>> >> >Mozilla is even worse, it has huge memory leaks just downloading
>things
>> >(M8,
>> >> >at least).
>> >>
>> >> Well, although I am no Mozilla fan, I must point out that M8 is
>> >> positively *ancient* and a LOT of problems have been fixed.
>> >
>> >Indeed, it's been a while. I see no reason to bother with Netscape or
>> >Mozilla when IE is so good.
>>
>> IE GOOD? *ROFL* IE is the only thing capable of making Netscape look
>> good ;^)
>
>IE is better than Netscape in almost any way you can measure. It's faster,
>stabler and more standards compliant. It uses less memory and is more
>configurable. Most importantly, it doesn't come with AOL Instant Messenger.
It's just *barely* faster, it's not more standards compliant in my
view (which could easily spark a side-thread on what it means to be
standards compliant if you wish), it uses less memory mostly because
it "integrates" into the OS and hides its memory usage that way, which
I do not want need or appreciate, and lastly Netscape doesn't come
with IM either, unless you choose to install it. Wasn't even an option
on my linux install, was an option I chose to uncheck when I last
installed it in windows and was never even downloaded.
The worst thing about it is it's a script kiddies wet dream. I tried
to use it for awhile, I upgraded it, tediously went through the
security settings and tried to fix that, then it wouldn't do anything
without squawking all the time and demanding I turn the security
settings down. So I uninstalled it. Now Media Player, without so much
as a by your leave, takes it on itself to download new codecs every so
often, waits till it downloads them completely, then complains it
can't autoinstall and demands I change IE security settings and dumps
the download to the bitstream, without ever once even letting me know
what codec it thought it needed. Very poorly behaved program in my
view. Completely insecure by default, and allowing the user to change
that is obviously an afterthought without any real support. Much like
a host of other trash that I uninstalled about the same time.
If you want to see what I consider a good browser, check out
http://www.opera.com - they are porting to linux and that's just one
more nail in the old MS coffin so far as my continuing to use them
(willingly) goes...
>
>As I said, the only thing Netscape has going for it is the fact it's
>available on other platforms. I wouldn't use Netscape on Windows with
>someone else's computer - it's just _crap_. The last decent version of
>Netscape was 3.0.
>
>You are rapidly coming across as an anti-Microsoft bigot.
Heheh I'm an anti-Microsoft bigot eh? Perhaps. If so, it was Microsoft
that made me that way though. I've been using their software since Dos
3, I was once one of their biggest fans. If they would pull their
heads out, I could be again. But they've shown time and again they
don't care about people like me, they don't care about the companies
that made the programs that made their OS usable, they don't care
about anything but extirminating those companies as soon as they have
a competing program and inflating their bottom line. And that's why I
have decided to quit contributing to it.
>
>> >> Ok, you do understand the difference between running a program and not
>> >> running a program, no? You can run a program and just let it sit
>> >> there, and it still takes resources (memory and cpu) while it runs,
>> >> even if from the user point of view it's doing nothing, it IS doing
>> >> something - if only running an idle loop waiting for input.
>> >
>> >If the login box is using enough resources on your machien to be even
>> >noticable, they it's simply underpowered for anything you're going to be
>> >using it for. I doubt you could even measure the amount of CPU time it
>> >would be taking and the amount of meory would not only be small, but
>almost
>> >all paged out in any event.
>>
>> It's not the resources that are worrisome generally, but the
>> instability. It's using the CPU, it's running a loop, accessing
>> hardware, and it is doing so as a kernel process. When you hit the
>> buggy bit of code it can take the whole box down without warning.
>
>I'd feel more than confident in saying the piece of code that loops the
>login box is bug free, after about 10 years of QA.
I'd feel more confident saying Unix is stable, after over 30 years of
QA with a lot more programmers looking at the code. Anyway, it's not a
theoretical question - MS Tech Support very often blames crashes on
video drivers.
>
>> As to the resources, that may not be a big concern in one case, but it
>> may be in another. It all depends on what your system load is. If that
>> extra memory requirement causes disk access that could have otherwise
>> been avoided, it's a problem. And no, you can't swap the routine out
>> either - remember it's drawing the screen and looking for input,
>> whether you have a monitor, keyboard, and mouse hooked up or not.
>
>And as I said, if the few hundred kilobytes of memory that it uses is of
>concern to you, then your machine is too underpowered to be doing anything
>with NT in the first place.
Heheh like I said, if I don't like bloat I should just buy more
hardware - that's been the NT line for a long time. I don't buy it.
That's just a sign of sloppy design.
>
>> I know people that are constantly working to get more out of their
>> companies web server, reprogramming, optimising, figuring out ways to
>> shift some of the processing to the client and so on. And you airily
>> dismiss any concern over wasting resources on some silly thing that
>> shouldn't even be installed to begin?
>
>No, I dismiss concern over wasting such a small amount of system resources.
>We are long past the days of "every kilobyte counts". Now it's more like
>"every 10 MB counts".
>
>> It's very easy to say "if
>> wasting that small amount of resources bothers you, you need a bigger
>> machine" (this seems to be the MS mantra - any complaint about the
>> software simply means you need to buy more hardware) but in the real
>> world the advantage goes to the one that can do more with less.
>
>"Less" in the case of removing the GUI from NT would be, at most, a couple
>of MB of RAM "less".
>
>If a couple of MB of RAM is going to have any worthwhile benefit/impact on
>performance then your machine _is_ underpowered for anything remotely
>"mission critical".
"Sloppy design" I repeat.
>
>> Linux clearly delivers that in server space. Just try to do anything
>> with NT on a 386 with no hard drive! Or a 386 period for that matter.
>
>Why are you mentioning "mission critical" in one post and 386s in another ?
Why does NASA still use 386s?
>
>I'm more than willing to admit Linux (and FreeBSD, and others) do a fine job
>of resurrecting old machines to sue as things like firewalls, and to keep
>some old machines running as file/print servers. However, this does *not*
>automatically mean they will be able to use higher end machines more
>efficiently.
And firewalls are not mission critical?
And I agree, working well on less hardware does NOT mean that it will
automatically run better on larger newer machines. In fact the linux
kernel didn't work very well at all on multiprocessor boxes until
recently. But the same good design qualities that lets it scale well
downward also lets it scale well upward, now that the necessary coding
has been done.
Different jobs require different tools. I am not stuck on any one OS,
I'll use whatever is best for the job. But honestly, the only way I
can see NT filling that bill would be as a vanity server so MS will
buy the company a lot of hardware.
>
>Btw, NT 3.51 would probably work fairly well on a 386 with fast SCSI disks
>and enough RAM, if you really wanted to.
NT 3.51 didn't work well on pentiums. I know, I had to babysit several
of those boxes. I ran Netware and NT side by side on the same hardware
for quite awhile, NT 3.51 couldn't hold a candle.
>
>> >> When the program in question is a GUI, and the video drivers are
>> >> running in kernel space, that's just more potential trouble. Taking a
>> >> risk like that may make sense in some cases, but not having the choice
>> >> of whether to do it or not is just poor design.
>> >
>> >Not if that design lowers development and production costs, especially
>when
>> >the risk is almost entirely theoretical.
>>
>> Well, the risk isn't theoretical, it's been acknowledged publically
>> that a large portion of NT crashes/locks originate from the video
>> driver.
>
>On Workstations, maybe. If this is true for servers then those servers are
>being poorly administrated, in which case a counter-argument is as simple as
>poorly administrated Linux machines.
And how does an NT admin unload the GUI again? *peer*
>
>An NT *Server* should have only the vanilla VGA driver loaded, and sit at
>the login screen all the time. At least, if you hold stability paramount.
Yep, and they still crash.
#####################################################
My email address is posted for purposes of private
correspondence only. Consent is expressly NOT given
to receive advertisements, or bulk mailings of any
kind.
Since Deja.com will not archive my messages without
altering them for purposes of advertisement, deja.com
is barred from archiving my messages.
#####################################################
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: How low can they go...?
Date: Mon, 04 Sep 2000 00:02:51 -0400
Reply-To: [EMAIL PROTECTED]
Said Simon Cooke in comp.os.linux.advocacy;
>"T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
>news:[EMAIL PROTECTED]...
>> Oh, c'mon. We know you'd be biased. I hate card studio software,
>> anyway. I'll bet its the genealogy stuff, right?
>
>Sorry, can't comment either way as to whether you're right or not.
Before you said you simply didn't want to sound biased. Now you're
saying you 'can't comment'. What's up with that? Secret project, eh?
Then why'd you bring it up?
--
T. Max Devlin
-- Such is my recollection of my reconstruction
of events at the time, as I recall. Consider it.
Research assistance gladly accepted. --
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: How low can they go...?
Date: Mon, 04 Sep 2000 00:05:28 -0400
Reply-To: [EMAIL PROTECTED]
Said JS/PL in comp.os.linux.advocacy;
[...]
>> My company pays for it; I certainly wouldn't. The *only* reason I use
>> Windows, or any Microsoft software, at this point (years ago I would
>> have voluntarily used Word and Excel, but they've gone seriously
>> down-hill, and weren't really all that good to begin with) is because of
>> the monopoly.
>
>How come the monopoly forces you to use it but not the other millions of
>users who get by without it?
Because my company uses it and my customers use it, because its a
monopoly. Get it?
> Is there a guy from MS standing next to you
>with a gun to your head? Or are you lying again.
Let me ask you something; did you ever upgrade your IE and have to
replace your video card?
>No one is FORCING you to use any operating system. Shit or get off the pot,
>either stop supporting Microsoft by choosing their product with each tap of
>the keyboard or stop complaining about it.
I don't support Microsoft by using their products; I support me by using
their products. If my customers use something different, and my company
uses something different, I'll use something different. You see how
that works? Microsoft has a monopoly because they make it hard for
everyone to use a different OS. Now stop boring me with this childish
bullshit.
--
T. Max Devlin
-- Such is my recollection of my reconstruction
of events at the time, as I recall. Consider it.
Research assistance gladly accepted. --
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: Rick <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.os2.advocacy,comp.sys.mac.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: [OT] Public v. Private Schools
Date: Mon, 04 Sep 2000 00:10:50 -0400
"Aaron R. Kulkis" wrote:
>
> [EMAIL PROTECTED] wrote:
> >
> > Aaron R. Kulkis" <[EMAIL PROTECTED]> said:
> >
> > >Rick wrote:
> > >>
> > >> "Aaron R. Kulkis" wrote:
> > >> >
> > >> > Rick wrote:
> > >> > >
> > >> > > "Aaron R. Kulkis" wrote:
> > >> > > >
> > >> > >
> > >> > > > > Finally, vouchers. As many of you may have seen, there are now studies
> > >> > > > > from three states indicating that minority students, particularly
> > >> > > > > Hispanic and African-American, do better in private schools. I think
> > >> > > > > vouchers are a great idea, *provided* that *extra* taxes are implemented
> > >> > > > > to pay for them. The bad thing about vouchers is that the money for
> > >> > > >
> > >> > > > Why do you need *extra* taxes for vouchers, when you have just
> > >> > > > stated above that the per-pupil costs of private schools are LOWER
> > >> > > > than that of public schools (Primarily to lower salaries, reduced
> > >> > > > red-tape and significantly smaller bureacracy).
> > >> > > >
> > >> > >
> > >> > > First, private schools charge "more" per student.
> > >> >
> > >> > Yeah, there are a couple of HIGHLY SELECT blue-blood boarding
> > >> > schools like Andover that cost more than the tuition at most
> > >> > colleges...
> > >> >
> > >> > But, by and large, the OVERWHELMING majority of private schools
> > >> > have lower per-pupil costs.
> > >> >
> > >> > Reduced management overhead is the primary reason.
> > >> >
> > >>
> > >> Really? Do you know the typical management structure of a typical public
> > >> school? Its very light. Very.
> >
> > >In most school districts, 30-50% of the per-pupil cost is MANAGMENT OVERHEAD.
> >
> > This would mean that at best, for every three teachers there is one
> > non-teaching administrator. The idea is complete nonsense. I've never seen a
>
> There you go again. Are you alleging that all administrators get
> paid the same as teachers?
>
> Are you forgetting that MOST administrators tend to get paid MORE than
> the teachers AND that they tend to run up a lot of highly dubious
> expenses like limos, air fare and "workshop" costs.
>
Public schools administrators in limos??? And you want us to take you
seriously?
> > public school with that kind of overhead -- and you haven't either. Now, how
> > would like to refine your point.
>
> Yeah, you made the absurd assumption that teachers salaries are
> anywhere close to administrators' salaries.
>
--
Rick
* To email me remove theobvious from my address *
------------------------------
** 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 (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
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
******************************
