re: libc->glibc: One good point here is that it was a move, advertised
as such; I remember talking to people from RedHat during the libc->glibc
move, they took lots of heat for it but they stood by their decision as
the only viable one, and the world has proven them right.
I've had almost zero experience with Win2K, and even less trying to
figure out what's compatible from NT4 and what breaks; but apparently
they're different enough for the industry at large to care, as there are
grass roots efforts to keep NT-based shops running. At least one
organization has announced a non-Microsoft-authorized NT certification
program to address the fact that M$ no longer certifies MCSE's for NT
(or will discontinue certification soon, I'm not precisely sure of the
time line).
Contrast to Red Hat during the glibc move; they included libraries to
provide as much back-compatibility/interoperability as possible, stood
by their decision as the right one, and let the reality of the situation
bring everybody else into line. Microsoft has to resort to bullying,
which they get away with by being the only show in town.
Some ancient DOS programs will still run on Win2K; but I understand that
there is a significant subset of software for WinNT that will not.
There is certainly a wide range (the majority?) of Win9x software that
won't run on NT, or won't run on Win2K, or both. I can't help but
believe that some of that is deliberate, a way to force customers to
purchase upgrades they would not otherwise need.
Every time I compile a new kernel, I have the option of enabling support
for a.out format binaries, in spite of the fact that the Linux kernel
and most of userland have been ported to ELF for longer than I've been
using Linux. I'm sure that many a.out apps would crash on one of my
boxes, but not because That's What Boss Man Stallman Wants. My point is
that, wherever possible, the poeple who develop for GNU systems have
functionality as the foremost concern; the people who write Windows
platforms are concerned with making money. Cross-platform
interoperability and back-compatibility are advantageous in the former
case, less so in the latter. And look, it shows up in the way the
systems behave.
Warren Melnick wrote:
>
> Rob,
>
> The major "upgrade" in the linux distros was when everything went from libc5
> to glibc. I believe that you will find that almost ALL of the Linux server
> shipments in 2000 were with glibc-based linux. The move from libc5 to glibc
> was quite readical, moreso than the move from NT4 to NT5 (or Win2K as
> marketting calls it). It was far more similar to the move from NT3 to NT4
> which was far more intrusive than the 4->2000 move, similar to the
> Win3.1->Win9x move of 1995/6, which in the linux world was similar in time
> and scope to the move from a.out to ELF binaries.
>
> Any questions? :)
>
> ----------------
> Warren Melnick
> Director of Research and Development
> Astata Corporation
>
> -----Original Message-----
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, March 22, 2001 12:10 PM
> To: [EMAIL PROTECTED]
> Subject: RE: OT: The Truth: Server Shipments in Y2K
>
> > -----Original Message-----
> > From: Michael R. Jinks [mailto:[EMAIL PROTECTED]]
> > Sent: Thursday, March 22, 2001 11:02 AM
> > To: [EMAIL PROTECTED]
> > Subject: Re: OT: The Truth: Server Shipments in Y2K
> >
> >
> > [EMAIL PROTECTED] wrote:
> > >
> > > Also, since we're talking aggregates and versions, should
> > we not be breaking
> > > down the Linux numbers by distribution in the same way the
> > windows stat was
> > > broken out by version?
> >
> > I don't think so. There are cross-distro compatibility issues, sure,
> > but in my experience they haven't mattered very much.
> >
> > If I did want to break out the stats into distros, I'm not
> > sure how I'd
> > do it; many of today's well-known distributions are Red Hat-derived;
> > should they be "less separate" than Red Hat vs. Debian vs. Slackware?
> > Does KRUD count as a distro?
>
> Granted. But could you not make the same case for W2K? That it is
> NT-derived? That's Microsoft's claim, anyway.
> Just so you know, I'm not trying to give you a hard time. I am playing a
> little devil's advocate to refine pro-Linux arguments such as yours so that
> I can use them too :).
>
> Regards,
> Rob
>
> _______________________________________________
> Redhat-list mailing list
> [EMAIL PROTECTED]
> https://listman.redhat.com/mailman/listinfo/redhat-list
>
> _______________________________________________
> Redhat-list mailing list
> [EMAIL PROTECTED]
> https://listman.redhat.com/mailman/listinfo/redhat-list
--
~~~Michael Jinks, IB // Technical Entity // Saecos Corporation~~~
With all due respect to Kenny Rogers,
the best that you can hope for is a return code of zero.
Opinions expressed above are my own, and not those of my computer.
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
