Re: [Gnu-arch-users] Re: practical questions of archive ownership

[Stephen J. Turnbull]

>>>>> "Miles" == Miles Bader <address@hidden> writes:

    Miles> Still, I'm curious what will happen in the future with
    Miles> emacs and other projects where there _isn't_ a strong
    Miles> central maintainer, and where the notion of a gate-keeper
    Miles> for the canonical sources is unlikely to be acceptable.

    Miles> It may be that, as now, people do initial
    Miles> development/testing in their own archive/branch, but merge
    Miles> into a shared central archive/branch.

Look at cscvs.  That's my guess: whoever is most active at the moment
will tend to pull patches toward themselves.  There are three
important archives there, and it looked to me like the center of
development started with Mark, shifted to Charles, gravitated toward
Robert Collins for a bit, then back to Charles.  But I don't even know
what the default archive for my cscvs workspace is.  And pragmatically
I don't care[1] if my impressions are right; from the point of view of
a down stream user, I just want fq patch names (and they're all in my
bash history, so I type C-r cduf C-e and edit the end of my last
replay command).  This wouldn't change if in fact John Goerzen is
doing all the work but somehow the patches are in Charles's archive.

And it seems likely that in the not too distant future, bk2arch and
svn2arch branches may fork off.  No problem (managerially).  The
various players can concentrate on what's important: refactoring the
core functionality to handle different constraints of different source
and target systems.

As things scale up, there are two important tools; one is Andrew's
patch queue manager, otherwise instead of gravitating to someone with
architectural/design/programming skills, the "center" will gravitate
to a manager/maintainer type.  Usually that person will not be Linus,
so that probably is a loss in terms of design.  At least to the extent
that the patch queue can be automated.

The other is going to be an index of all patches.  This isn't
important at cscvs scales.  It seems like Mark retired, but Robert and
Charles basically have all of each other's patches.  But I think it
makes sense to separate the indexing function (basically managerial)
from the design aspects of patch-flow control.  But with that, I don't
really see the need for having the whole archive back to Adam and Eve
that Andrea keeps pushing.  Yes, he probably _will_ keep everything
that Linus ever commits handy, but there are other important players
out there.  With a global index he can keep track of all them,
including people who are important to him, but maybe not to Linus.

Ie, different people will have overlapping but quite different ideas
of what "the whole ball of wax" means for them.


Footnotes: 
[1]  Of course my apologies if I'm not giving credit where it's due.

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.