[Asterisk-Dev] benevolent dictatorship, or inclusive developper community?

[Steven Critchfield]

On Thu, 2004-01-08 at 12:23, C. Maj wrote:
> On Thu, 8 Jan 2004, Steven Critchfield waxed:
> 
> > On Thu, 2004-01-08 at 10:32, C. Maj wrote:
> > > On Wed, 7 Jan 2004, Chris Albertson waxed:
> > > 
> > > > (see update command in cvs manpage). So, yes you could have
> > > > multiple lines of developmentand merge them back into a main line. 
> > > 
> > > Yeah and live in a nightmare.  The kernel only uses CVS as a
> > > daily (or whatever) dump of what's in BitKeeper.  People
> > > submit patches against CVS, sure, but the "branching" is
> > > done with BK repositories.
> > > 
> > >     http://www.bitkeeper.com/
> > 
> > Well without dredging up the BK vs. every other revision control
> > software flame war, lets just point out that that wouldn't be a viable
> > option here. 
> 
> Point being the kernel doesn't use CVS, so it's apples and
> oranges.  You seemed to imply previously in this thread that
> the kernel worked like that and this is how branches or
> "mini forks" are created, through CVS.  My apologies if that
> was a glib interpretation of your comments.  I'm just trying
> to determine whether it is your lack of knowledge about BK
> that would lead you to suggest that it's not a viable option
> or something else.  Could you please explain ?
> 
> > I would suggest subversion, but it is easier to stick with what more
> > people know at this moment and not force anyone to deal with the
> > conversion of the tree one more time.
> 
> They don't have to.  Only the developers who want to keep
> their own branches would.  Their bleeding stuff could be
> pushed back into Digium's repository, for example, and
> run BK2CVS on it there for the masses.

Please understand that I don't want to get into the Bitkeeper flame war
here. It is well documented on other lists, and it was argued by people
more appropriate than us. 

What I am trying to suggest is that there should be some way for Digium
to host and manage the repository while they open it up for a few to
further develop portions to be folded back. A central authority is
needed for tracking and managing this. I already documented my concerns
about this control leaving Digium. I feel we could essentially pull off
what the kernel developers have done in such a way that would allow us
to regain some of the speedy development that was happening before the
mailing list exploded. 
 
-- 
Steven Critchfield  <critch at basesys.com>