[GRLUG] WMNTUG Windows 7 Meeting

[john-thomas richards]

On Thu, Sep 17, 2009 at 03:12:07PM -0400, Bob Kline wrote:
> On Thu, Sep 17, 2009 at 12:49 PM, john-thomas richards <jtr at jrichards.org>wrote:
[snip]
> > > I'd thought the releases were going
> > > to come a little more often now.
> > >
> > >    -- Bob
> >
> > Well....sort of.  There is general agreement among the developers to
> > release more frequently, but you will never see anything approaching the
> > frequency of a Ubuntu (who still continue to benefit from Debian's
> > approach).  If Debian could release every 18 months I'd be happy.  The
> > two- or three-year release (such as with Sarge) are really painful since
> > change happens so quickly in the FOSS world.  As I mentioned, I'm
> > running Debian Lenny (the current Stable) and I'm rock-solid.  For
> > example, I haven't experienced any instability with xorg that comes from
> > the new graphics execution manager (GEM) and kernels 2.6.28 and up.
> > (See http://distrowatch.com/weekly.php?issue=20090817#feature for more
> > details.)  That sort of instability cannot* happen with Debian since
> > Debian is so far behind the curve.
> >
> > *Okay; "cannot" is a strong word—perhaps too strong.
> > --
> > john-thomas
> >
> 
> There are issues with this.  In one sense
> the Debian effort simply gets to ride the
> backs of more current distributions, where
> bug fixes are an ongoing process.  In enough
> time  things are pretty well debugged.  The
> most used things anyway.

This assumes that Debian only works on stable releases.  Actually,
Debian contributes a ton of stuff upstream via the Testing and Unstable
branches.  Many, many Debian users run Testing or Unstable and
contribute bug reports (I usually run Testing and have since I started
using Debian back in 2002) though I am currently running Stable (I plan
to upgrade as soon as the GEM / Intel video chipset issue is worked
out).  Debian is not riding the backs of anyone, anymore than other
distros / developers are riding the backs of Debian.

> And unless it does all its own testing, it
> also gets the benefit of legions of testers
> for the current releases of other distributions.
> In many ways the game is to get as many
> knowledgeable users to use something, and
> report any bugs, as possible.

Naturally there is a lot of testing (most of it?) that is done upstream
by the application's developers.  Clearly Debian benefits from the very
wide userbase for, say, Firefox (though Debian rebrands it).  Just like
other distros, Debian also submits patches and bugreports upstream.

> Nothing unfair about all this, but simply to
> say that in part Debian's distribution is
> solid simply because it's old.   Old in the
> software world anyway.

What I was trying to say is Debian Stable is far less likely to be hit
by stability bugs because of its longer release cycle.  Ubuntu 9.04 (I
think) got hit pretty hard with users having Intel video chipsets
because of the instability that the new GEM architecture brought.
Debian would never have released such new software and so the bugs (that
Debian is helping to fix) will likely be worked out long before the
relevant software / kernel combination hits Debian Stable.  It's not so
much that Debian software is old.  It's that the release schedule makes
finding and avoiding such bugs more likely.

> Is there anything to suggest Debian's
> distribution is measurably more reliable
> than, say, an older version of Ubuntu?

I only run Debian so even my anecdotal evidence is unreliable.  The
above mentioned problem is Ubuntu is an extreme example, but is one that
is unlikely to ever hit Debian Stable.

> I'm sure there are fight'n words here
> somewhere, but I'm just looking at the
> situation.

Fightin' words?  It's not like we're talking about vim versus emacs...
-- 
john-thomas
------
I want everybody to tell me the truth, even if it costs them their jobs.
Samuel Goldwyn