> Ted Husted wrote:
>
> > "Geir Magnusson Jr." wrote:
> > > However, if they are 'independent entities', it is missing one thing -
> > > the grouping of committers. I still chafe at the one large committer
> > > group that spans every package idea. See 'Subproject Guidelines' #15
in
> >
> > I think it's mainly a security/human labor issue, Geir. There's a
> > certain amount of work involved in setting up a new CVS repository, and
> > adding committers to it. Until we can show cause, I think the best thing
> > to do is use the status file to track the committers for each package,
> > which will give us a leg-up in case we want to create more repositories
> > later.
> >
> > I think the mailing lists, on the other hand, are easier to create, and
> > are also self maintaining -- people can sub and unsub themselves. So,
> > its easier to keep those separate from the beginning.
> >
>
> Speaking as the person that would be responsible for both of these things,
> you've got the effort required totally backwards :-)
>
> Setting up a CVS repository is very straightforward (< 3 minutes).
Setting up
> a mailing list is much more involved, and causes incremental overhead not
only
> on the Apache mail server (which serves all *.apache.org mailing lists)
but
> also on the nice mailing list repository sites that are responsible for
> archiving our lists for us.
I like Costin's argument about having only one CVS repository. Given the
apparent overhead of having separate mailing lists for each component, I
would say we should go with only one set of lists : projectname-announce /
projectname-user / projectname-dev.
> (Isn't it scary when Costin and I agree about anything ... that's *twice*
> now! :-)
Yes, something's wrong here :-)
Remy
