On Mon, Aug 03, 1998 at 01:53:07AM -0700, George Bonser wrote:
> On Mon, 3 Aug 1998, Joey Hess wrote:
> True, but kernel version is not as important as libc version and
> filesystem layout. Kernel version really means very little. It simply
> provides an easy method of very basic configuration management. Not
> detailed ... I would not expect there to be more than a half-dozen things
> that the LSB should specify ... maybe 10 at most.
Well I'm fascinated watching this debate.
First, I would like to comment that for some weird reason I do not
like "version number maniac" companies, where the version numbers of
programs goes sky high in quite a short term. Major versions every
3-6 months, with little to none core changes. I think even RedHat
did it with a little bad taste, they fired the version number into the
sky, new minor every 3 months, new major every half year? Uhm, not my
style. I won't look forward to RedHat 99. :-o
Other part is the kernel you mention. The kernel version IS important.
Heck, those non-geeks ask first: "Hey, what version of kernel this
distro' have? 2.0.34? Naaahhh, it's old, they told me!" And regarding
this RedHat versions is WAAAAY out of range. I think, for a change,
you should try to convince RedHat to go back to version 2.0.xx because
this matches the kernel. Debian is just okay, v2.0 for the 2.0.xx
kernels. :-)))
So, there are many factors. You like high version numbers, you like
"comparable version numbers" (I think of StarOffice and micro$oft word
here for example), you like a next major version everytime there is
more than 10 files changed in the distribution.
Debian people seem to like the other way. Small version numbers show
that there were few required bugfixes, updated updates, since the
distribution is perfect. ;-> They raise major _only_ when there was
a _major_ change, like libc5->libc6 which could be considered like one.
Neither older ones, nor hamm->slink won't be a big leap (unless maybe
we'll be fortunate enough to have 2.2.xx kernels in our hands then,
but I think not even that would be major enough).
Major versions for ONLY major changes, minors for minor, and
not entering the version-number-hype-marketing-bandwagon is
the hackers' view of version numbers. And non-geeks can realise
it, too. (As for those you mentioned watching only the version numbers,
well, if there wouldn't be stupid people the psychiatrists would
starve to death.)
And, for closing, let me quote a wise man (I think it was Occam but
I let anyone fix my mistakes :)) regarding these numbers, too:
"Small is beautiful!"
:)
bye,
grin
(using an operating system with a
core kernel version LESS THAN the decade.)
--
Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null
