Re: [TCLUG:22510] Re: [TCLUG] Re: [TCLUG:22499] RH7 is proven unworthy yet again

["Thomas T. Veldhouse" <veldy@veldy.net>]

> First, 6.2 is rock solid. So, unless you want bleeding edge. Stay on 6.2.
>
> Second, that the open source mantra of release early, release often. Learn
it.
> LIVE IT, but respect it. Living by this mantra can have some side effects.
>
> My personal opinion is all the heat RedHat is getting is going to cause
them to
> have very long (ala 2.4 kernel) release cycles.
>
> I'd rather have a rapid release cycle, followed by a flurry of bug fixes
then a
> stagnate distribution.

I totally agree.  However, the file descriptor bug is an example of pushing
it "too" quick.  The proper audits would easily have caught that one.  If I
am allowed to infer, I am betting that have a terrible QA team, and instead
they have relied upon their developers unit testing (which is probably
exceptionally good).

Debian is a great example of the opposite extreme.  They release so
pitifully slow that most people are running "unstable".  I think that Redhat
simply should have used "blessed" software unless they develop it themselves
(I am excluding the fact that they purchased Cygnus).  Using gcc-2.96, which
is never to be released and not binary compatible, is not something they
should have done.  Put an RPM on rawhide for that.  They undermine any
indication of quality and instead give the impression of open arteries.

If I wanted the bleeding edge, I would probably compile it myself on a known
stable system.

What it comes down too is that they wanted to release something, but there
was not really anything new to release.  It is a selling point to be able to
say they have bumped up a version number on this package or that.

And the anti-Redhat ramble goes on ...

Tom Veldhouse
veldy@veldy.net