long term support release

[Horst H. von Brand]

Ralf Corsepius <rc040203 at freenet.de> wrote:
> On Wed, 2008-01-23 at 14:55 -0500, seth vidal wrote:
> > On Wed, 2008-01-23 at 19:35 +0100, Ralf Corsepius wrote:
> > > Further: IMO, fedora legacy did not fail. It was discontinued by
> > > management, because it collided the certain business interests.

> > Ralf,
> >  I know there's no way to convince you of this,

> No there isn't.

Sad.

> This thread added further to this. It once again made it clear that many
> parties being involved into Fedora aren't willing in implementing a
> Fedora LTS or simply extending the lifetime of Fedora for various
> reasons - We will see if Ubuntu LTS will be a success. I would assume it
> to further contribute to Fedora loosing ground.

The people who were supposed do do the work, didn't. That's how OSS
projects die,

> >  however, this isn't at
> > all what happened. Legacy just died.

> As I see it, it died, because people wanted to let it to die for various
> different individual motivations.

... and not enough people stepped up to keep it alive. This was /not/ the
result of some "Red Hat conspiracy", Fedora Legacy could very well have
gone its own way, /if there had been enough interest/. Much more irritating
as competition than a "Fedora LTS" are projects like CentOS, and those are
doing well.

To give my own perspective: Yes, we do use Fedora day-to-day here, and
sometimes the new version/EOL came at /very/ incovenient moments, so we
used Fedora Legacy, but only for a couple of months. Then we moved our
servers over to White Box and now CentOS (long enough lifetime to allow
planing upgrades), the workstations at the computer labs (via kickstart)
can usually be migrated over a weekend (after some testing on guinea
pigs). So our (in any case very limited) need for Fedora Legacy went away.

> > It wasn't decided or dictated by
> > anyone other than the reality that nothing had happened for months.

> Yes, but not because the "product sucked", the project was poorly
> implemented and poorly lead.

Maybe. Do it better this time around.

> Anyway, meanwhile, things in Fedora-land have changed. 
> 
> What prevents Fedora from launching a "Fedora LTS" as part of Fedora,
> using the Fedora infrastructure, e.g. by simply imply declaring "Fedora
> 7" life time's extended for, say 2 years, when FC9 comes out?
> 
> The price would be fairly small.

s/fairly small/extremely large/

>                                  In particular, the infrastructure
> already is in place. All what would have to be implemented would be some
> regulations/conventions concerning "ABI/APIs" and ACLs.

And having /three/ branches of Fedora instead of /two/ means roughly 50%
more manpower, a resource that is sorely lacking as things stand. Moreover,
you need more of the (very scarce!) people willing to do the thankless job
of backporting bugfixes to ancient codebases. This requires /much/ more
expertise on the code than just taking the upstream package, following the
discussions of the developers and applying urgent proposed patches to the
current codebase. In essence, when you follow upstream, there are many eyes
and hands upstream that do most of the hard work for the packager; when you
maintain a legacy package, you are almost on your own.

No, "then just get the newest package then and integrate it with the rest"
doesn't cut it. If I want LTS it is exactly because I don't want ABI/API or
configuration changes /at all/.

And, again, if you really want LTS, RHEL or CentOS fill that niche nicely.
Sure, it is not exactly Fedora, it doesn't have all the glitter, but it
works.

> After these 2 years, you'd have the results of a fair comparison to
> "Fedora". If it goes down the drain, so be it ...

> > No one wanted to take on the rather large effort to keep it going. It
> > stopped. There was no management who discontinued it. 

> May-be there had not been an explicit RH management decision on letting
> it die. But there also had been no will to support it.

Because of the large effort required, that no one stepped up to commit?

> However, I recall FESCO (or had it been FAB?) having decided on FC's
> short life-time

Very good idea, helps keeping the distribution focused on bleeding edge and
minimizes spreadout of the (very limited!) developers.

>                 and to support EPEL.

Helps bridge the RHEL/CentOS <--> Fedora difference, nice idea.

>                                      Both decisions have been severe
> mistakes, IMO.

Disagree strongly.
-- 
Dr. Horst H. von Brand                   User #22616 counter.li.org
Departamento de Informatica                    Fono: +56 32 2654431
Universidad Tecnica Federico Santa Maria             +56 32 2654239
Casilla 110-V, Valparaiso, Chile               Fax:  +56 32 2797513