[gst-devel] GStreamer needs a maintainer

[Christian Fredrik Kalager Schaller]

I go to the Cinema to watch a special showing of LOTR 1&2 as a warm up
for the third instalment and a flame war errupts. Can't I be away even
for a day ;)

Anyway I think having a 'maintainer/project lead' would be nice. But for
that person to be able to function he (theorectically she) would need to
command enough respect that everyone here is willing to accept the
decisions made. If I remember correctly even when we did have a
maintainer, in Wim or Erik, and he made decisions they didn't always
become graciously accepted by the person whose idea got turned down. In
other words having the maintainer didn't cure us of malcontent at times.

The modularized design of GStreamer has made it possible for us to keep
clashes low as people have been able to work on their own in smaller
compartments of the code, often with little effect on what happens in
other parts. To me it looks like the contributions done by everyone over
the past leaderless year has moved us fast forward, probably faster than
we progressed the year before that.

During this year we have in my opinion made decisions based on rough
consensus, and some of the points listed here had already been decided
upon in that matter. For instance that we would try to get 0.8 in GNOME
2.6 is something I thought we before and during our Barcelona retreat
agreed upon. Given that everyone concede we did make that decision it
has a direct impact in my opinion on other issues here, like the
declaration of API freeze. While we as a non-dev plattform lib have
freer terms, we do have a commitment to ourselves and our application
developers to stabilize stuff. Which I think means that the time for
changes that break stuff in a major way is over for this round. (A new
opening comes if we decide to do a 0.10/1.0 for GNOME 2.8) It also means
in my opinion that the time for API changes at all is closing in fast. I
think we should use the GNOME feature freeze as our API freeze date. We
need the time remaining in my opinion to stabilize stuff of we are to
make GNOME 2.6.

The paragraph above isn't meant as me trying to make decisions more like
me trying to show that I think we have made some decisions indirectly by
earlier choice.

As for the issue of a separate gstreamer-gnome package, and wether all
headers and plugins are put in gst-libs/gst. I think Thomas do clearly
have a right of decision here as it is he, with some feeble assistance
from me, who will end up carrying the brunt of the work for maintaining
extra modules/inclusions. We might not have a overall project
maintainer, but I think we have a fairly well established tradition that
the person who made/maintains a part of the code gets to decide upon
that code. 

As to get back to the original question of an overall project
maintainer; well the only person I think has a chance of functioning in
such a role is David Schleef. Afaict he is the only one among us who I
think there is a chance everyone will be able to accept the decisions
of. But as stated earlier, our experience when Erik and then Wim was
project leads shows that having a project lead isn't a guarantee against
everyone of us at some point feeling unjustly/wrongly/incompetently
overruled.

Christian