Very Computer

Hi there,

For those of you that are interested, PC Week had a NOS shootout
between Linux 2.2.7, Solaris 7, Netware 5.0 and NT 4.

Here is the URL:

http://www.zdnet.com/pcweek/stories/news/0,4153,401970,00.html

Cheers !

-Kotze.

http://www.computerworld.com/home/print.nsf/all/990510A402

"Lintel will run most processes three times faster than Windows NT on
the same hardware, and you won't have to reboot twice a day."

Clearly, we are dealing with an advocacy piece and not anything mired
in objective reality.  So please, continue shoveling and continue
calling it "the real world", and Linux will continue to be viewed as a
perennial also ran with an immature (as far as being able to see issues
in wider context than just technology) user base and microsoft will
continue getting richer, thanks to you.

--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---

Quote:> I didn't have the time to read the entire page but I do not think it had
any
> statistic to validate linux capabilities!   I don't know why but pcweek
and
> zdnet have articles that seem quite similar in that they say NT is a
better
> web server then linux:)  Of course one to two was (not sure which one) was
> using an old version of linux.   The only good thing about linux is that
its
> making nt alot better, competition is wonderful!
> tr

FYI, They used RH 5.2 version with the kernel updated to the 2.2.7 version
and went through the hard part of Linux OS, optimizing it to max
perfermance. Sounds pretty fair to me, then again, I don't see everything
with Linux glasses on....

Quote:> And here's some stuff from the real world:

> http://www.computerworld.com/home/print.nsf/all/990510A402

The only bad things they have to say about Linux are:

  o lack of applications
  o their box peaked out at 2200 HTTP connections per *second*

--
The wheel is turning but the hamster is dead.

    I've looked at the PC Week NOS shootout; Linux lost fair and square.
    Part of it may not be the kernel's fault (The Apache development
    team is fairly vigourously denying that Apache is a speed demon),
    but part of it is.   Whining about it will do nothing except give
    the marketing people at Microsoft warm and fuzzy feelings as they
    stock up their "it's just like the Amiga, and do you want Amiga
    Fanatics running YOUR mission critical systems?" quotes file.  If
    you want to do something about it, shut the * up, grab a copy of
    the kernel, and fix up what's not working -- Linux is not freeBSD,
    where you have to be voted into the core team, but will let any old
    joe patch the code and submit changes, and yours might be the change
    that will blow away NT in the next NOS shootout.

                  ____
    david parsons \bi/ ``Open Source(r)(tm)(c)'' means that you get to hack
                   \/                          the stuff that isn't working.

[deletia]

        When does anyone buy 'just a NOS'?

--

    Microsoft subjected the world to DOS until 1995.             |||
         A little spite is more than justified.                 / | \

                        In search of sane PPP Docs? Try http://penguin.lvcm.com

->     the kernel, and fix up what's not working -- Linux is not freeBSD,
->     where you have to be voted into the core team, but will let any old
->     joe patch the code and submit changes, and yours might be the change
->     that will blow away NT in the next NOS shootout.

Linus has to approve the patch before it will get into the main source
tree.

In response to a question I had about Linux source in CVS, Linus
Torvalds wrote (on linux-usb mailing list):

-> The above implies that I will just *-stamp whatever people are
-> working on. That's not how you create a good and stable development
-> environment, and that's not the kind of project leader I am.
->
-> When sending me patches, you should basically expect that there is a 50-50
-> chance that the patch gets rejected for some reason. And the risk of
-> rejection goes up very steeply as the size of the patch goes up. Which
-> means that it's much much more likely to work if you send me ten small and
-> clearly correct patches, than it would be if you sent me one large patch.
->
-> Even if that large patch was just composed of the very same ten small
-> patches.
->
-> This does not mean that CVS doesn't work. It just means that CVS alone
-> does not work as a front-end to Linus. As an example of this you can take
-> the patch-sets maintained by Alan: he does a really good job of it, and
-> what happens is that people see his large patch-sets (linux-2.2.7-ac3 or
-> whatever), but by the time he actually _sends_ me the patches he knows how
-> I work, and he splits them up into many smaller independent patches. For
-> all I know he may be using CVS to maintain that all.
->
-> But back when he was trying to force-feed me huge patches that weren't
-> clearly split that way, both of us just used to get very frustrated.
->
-> Think of me as the worst nightmare that Dilbert ever had. Think of me as
-> the boss that micromanages _everything_, and who has a very limited
-> capactiy for thought. Anything above that limited capacity gets discarded,
-> because if it's above my head it must be wrong, right?
->
-> And if you think of me that way, you can still work it all out. Use CVS
-> all you please, and then you sneak that fact past my back by spoon-feeding
-> me each commit one-by-one..
->
->           Linus

So, not only do you have to speed up tcp/ip or whatever, you have to
do it in a way that Linus will accept.

Linux is not the anarchy you may think it is, or even the oligarchy
that FreeBSD is.  It is a dictator-ship just like most successful free
software.

--
David Steuber   |   s/trashcan/david/ if you wish to reply by mail

Boling's postulate:
        If you're feeling good, don't worry.  You'll get over it.

-> FYI, They used RH 5.2 version with the kernel updated to the 2.2.7 version
-> and went through the hard part of Linux OS, optimizing it to max
-> perfermance. Sounds pretty fair to me, then again, I don't see everything
-> with Linux glasses on....

Not to mention that speed isn't an Apache priority.  It is designed
for flexibility, not speed.  IIS is a fast web server if you don't tax
it too much.  I'm not talking about connections, I'm talking about
back end processing to deliver a page.

If your web server is plugged into a T1, I don't see that it would
make any difference what you are running.  At least not in terms of
performance.  I also think that a properly set up NT server can run
with reasonable reliability.  It may not be as stable as a unix, but
it is stable enough.  Who is going to notice a once a week reboot of a
web server?

Hopefully these tests will cause the Linux kernel hackers to improve
the code.  Then this would be a good thing for Linux.

If the kernel was perfect, development would stop.

--
David Steuber   |   s/trashcan/david/ if you wish to reply by mail

Ask Not for whom the Bell Tolls, and You will Pay only the
Station-to-Station rate.

Quote:>but I do not think it had any
>statistic to validate linux capabilities!   I don't know why but pcweek and
>zdnet have articles that seem quite similar in that they say NT is a better
>web server then linux:)

  http://www.zdnet.com/zdnn/stories/news/0,4586,2242246,00.html

then go read this (NT FUD exposed by statistical analysis):

  http://www.cnn.com/TECH/computing/9904/21/ntpaper.ent.idg/

then for the final touch (mindcraft fully exposed):

  http://www.salonmagazine.com/tech/feature/1999/04/27/mindcraft/index....

Before you post here again, I suggest you have the time to read these pieces
(should take less then 20 minutes), because there will be a quiz if you do.

Quote:>Of course one to two was (not sure which one) was
>using an old version of linux.   The only good thing about linux is that its
>making nt alot better, competition is wonderful!
>tr

Quote:> but I do not think it had any
>statistic to validate linux capabilities!   I don't know why but pcweek and
>zdnet have articles that seem quite similar in that they say NT is a better
>web server then linux:)  Of course one to two was (not sure which one) was
>using an old version of linux.   The only good thing about linux is that its
>making nt alot better, competition is wonderful!
>tr

Anthony

<snip>
--
-----------------------------------------
| And when our worlds                   |
| They fall apart                       |
| When the walls come tumbling in       |
| Though we may deserve it              |
| It will be worth it  - Depeche Mode   |
-----------------------------------------

   I apologize for responding to the wrong article; I don't have
the article I intended to tespond to.

--
Chris Costello
Any given program will expand to fill available memory.

Quote:> >I didn't have the time to read the entire page

> Oh no? Then why is your opinion valid?

Quote:> Maybe you should read this ZDNET piece before your foot gets any deeper:

>   http://www.zdnet.com/zdnn/stories/news/0,4586,2242246,00.html

zdnet contradicts themselve way to often to be of any use!

Quote:>Before you post here again, I suggest you have the time to read these
pieces
>(should take less then 20 minutes), because there will be a quiz if you do.

Quote:> "Before you post here again"