On 4/25/07, Tim Jackson <lists(a)timj.co.uk> wrote:
Johnny Hughes wrote:
> We should at least, in my opinion, agree on a consolidated group
of SPEC
> files.
Absolutely.
And on maybe a methodology of how SPEC files are to be
written/documented/etc. This would be the biggest win from the Fedora
items as the problems that we see in the Enterprise environment can be
fed back to make things work better.
> I am not trying to be negative, but "it seems that"
EPEL just assumes
> that they will be the defacto standard in this space ...
I don't think that's the case; I think users will decide that.
I don't think so either. I think that each repository comes across
that way at one point or another in the conversations.. if nothing
more than we end up identifying with what we work with. That being
said, I identify myself with the SmoogeSpace Repo :).
> It just seems backwards to me is all ... maybe I am missing
> something?
You're missing the fact that although EPEL may "on paper" have 0 users
right now, it is not coming from nowhere out of the blue: it is an
extension of Fedora Extras which is a very popular project. The repos
you mention have lots of users, true (I assume), but they have a notably
small group of "owners". So if you add together the *collective* time
that has gone into Fedora Extras from the many contributors, it's at
least possible that it may already exceed that of (Dag + Axel + Matthias
+ Dries + ...) But anyway, arguing over numbers isn't the point. It's
I agree with you on that point. Current numbers do not matter in the
argument. What one brings to both existing and potential customers is
what counts.
probably fair to say that what 3rd party EL repos really have going
for
them is deep technical knowledge and extensive real world deployment
experience. What EPEL may lack in that area, it probably makes up for in
the community/participation aspects *that already exist* and
infrastructure. In that specific sense (and only that sense), some of
the 3rd parties may have had it *RELATIVELY* (I don't mean to belittle
the efforts in the slightest) easy, because those owners have had the
freedom to do what they wanted in their repos without having to decide
by committee. Now, the end results are good (and one can argue that's
all that matters), but to be fair, that acknowledgement should work the
other way too, and the inherited experience & infrastructure of FE in
EPEL ought to be at least acknowledged.
I recognize that Fedora Extras has a lot of people who are interested
in packaging up stuff. However that does not translate into lots of
people wanting to support things for enterprise environment when that
entails supporting very old releases, etc. I also see a problem in the
case where a package needs to differ from Extras and EPEL because of
maintainability or other items.
What are the customers that EPEL is trying to 'service'?
Are they the customer that wants the same version of clamav on EL2,3,4,5?
In some ways I would argue that this is audience that rpmforge/atrpms covers
Are they the customer that wants a stable version of clamav for
EL2,3,4,5 for a short time?
Other repos may cover this???
Are they the customer that wants a stable version of clamav for
EL2,3,4,5 for a long time?
Are the customers willing to replace EL core packages with
'newer/different' items to make other packages to work [example:
update ruby for puppet, update python for plone?]
Do the customers want a wide variety of packages?
Do the customers want a consolidated working set of packages (say
security only, plone website, etc)?
While they probably want all of the above, the customer is limited by
the fact they are getting this for free.. they can find someone who
will pay? EPEL will need to figure out which customers it wants to
help service and why it wants to help service them.
--
Stephen J Smoogen. -- CSIRT/Linux System Administrator
How far that little candle throws his beams! So shines a good deed
in a naughty world. = Shakespeare. "The Merchant of Venice"