[gst-devel] GStreamer needs a maintainer

[Thomas Vander Stichele]

El mar, 16-12-2003 a las 04:51, Benjamin Otte escribió:
> ... becuse I'm fed up dealing with Thomas and his ideas and arguing every
> change I do. Especially with someone who doesn't have a clue about
> anything but build system related things. [1]

I won't bite if it's that obvious :)
Anyway, you seem to be confusing some bunch of things.
IRC is for CHATTING.  If I disagree there with stuff you do and you
disagree with stuff I do there then that's fine.  It's for CHATTING. 
It's ok to get angry at each other sometimes.  I get angry at friends
too, and they get angry at me too.  It doesn't mean I don't like them.

It also doesn't mean I argue every change you do.  I do argue some of
the methods you do them with.

I particularly find it annoying that you have a hard time collaborating
on stuff people want to collaborate on.  Case in point, metadata stuff. 
The first one to really care about it in GStreamer was me.  At the time,
I said what I wanted to do, asked for how to do it, invited anyone to
help out, and so on.  Sure, it wasn't perfect in GStreamer, but hey, at
least I got some results out of it.

One day you announced you were going to reimplement support for this.  I
repeatedly asked for some clues on what you were planning, if you needed
help, if you wanted to discuss it, and so on.  You either ignored that
or said "I don't want to share what I have until it's finished, I'm not
sure it's good enough yet".  Well, OK, fine.  It left my stuff in limbo
for a while and I could read well enough between the lines that you
weren't really intending for that to matter.  OK, I'll roll with that.

And then one day you commited everything to CVS, remove the old ways,
break media-info/nautilus-media, and leave it at that with a short
explanation of what you did.  When asked at some points, you were
willing to comment that nautilus-media is broken anyway.  I know that; I
don't think I've ever heard you say about any piece software that it was
anything but broken.  When pressed though, no real reason came forward. 
Just that it was broken.

I can understand that for some reason you don't care about stuff that is
already out there using GStreamer.  I can understand that you don't seem
to care about what this does to the perception of GStreamer.  I cannot
really understand why you were the one to push for GStreamer 0.8 in
GNOME when you don't care about whether or not we actually make it easy
for GNOME to use us, but I assumed that your intention on that meant
your committal to making sure things above the core level were working
OK.

All I'm trying to say is - that is NOT collaborative development.  You
may argue that you do not deem me worthy enough because I mostly work on
build stuff lately (the ugly bits that nobody ever cares enough about to
get right, hence I'm stuck with them), and even that is fine by me. 
Honestly, as long as you're writing decent code, which you are doing,
I'll roll with whatever we collectively as a team put out.

If you break the build, I'm fixing it.  If you break my apps that depend
on stuff, I'll try to fix them too.  And so on.  I'm fine with whatever
good code you write.  The things I did change here and there where
things I announced to you beforehand, and were aimed at getting stuff
working.  I told you *last week* that make distcheck didn't pass at all,
you said it did, I asked you for a log so I could check, I never got
one, and two machines doing daily builds showed you it *didn't pass*. 
So I ended up moving *one file* to make it pass.  You can hardly call my
attitude "doing things on my own".

In any case, I think your flamebait was way over the line as long as
spider is the way it is :) (This is flamebait too, to make it clear).

Honestly, I don't see the need for having this discussion.  We have
different views, different ways of doing things, and as long as we share
similar goals there's no reed for petty discussion.  I don't always have
to like what you do, and neither do you have to like what I do, but as
long as we're getting closer to some goals we all share, that's fine by
me.

> I need a definitive instance to decide on things. This "we'll argue for a
> while about it and then everyone does what he wants" doesn't work.

Our problem is that the two authorative figures, Erik and Wim, have
left.  That leaves us with no real leader figure.  Is that a problem ?
Maybe.  It's something we need to work out.  Does that mean we just have
to choose one ? I wouldn't think so, none of us is actually completely
qualified to do so.  You've argued that Erik and Wim maybe weren't
either, and you might be right, I don't know.

Bottom line, there *is no clear leadership* right now.  The best way to
handle that, IMO, is to re-establish some of the policies we had instead
of keep challenging them (for example, "use mailing list for decisions
instead of IRC"), so we can turn our energy elsewhere, divide
responsibilities between team members that have done a decent job at it,
and take it from there.  And discuss stuff if there are real issues.  I
don't believe opensource projects can survive without a lot of
collaboration and a little discussion.

Thomas