[maemo-developers] Diablo extras repository proposal

[From: Tim Teulings]

Hello!

> This is all good, but I don't think it makes sense to do in now.
> The main problem with extras for now are the maintainers' attitude.
> There were a lot of discussions about extras previously on this list and
> no real actions were done by community.

I don't think it is fair to blame the developers without looking for
reasons. I'm sure there are reasons for the current situation (some
already responded to your mail). The situation needs discussion but not
simple blaming. Situation is frustrating but worse could happen.

> After we accomplished Misha's plan[1] several months ago[2] we had no
> feedback and almost nobody used autobuilder and promoter.

I can speak for others but I personally tried to find a solution between
improving existing and developing new software, switching the employer
(plus working overtime), moving to a new home, buying a new car and
other stuff. I'm not payed for developing open source software so other
things have priority.

> With this attitude I don't think it's feasible to go further and develop
> additional features which will not be used again. Instead we should
> probably concentrate on encouraging maintainers to actually use the
> tools and to give feedback. From this point of view Niels proposal is
> more than enough for now.

So back to the possible reasons for the situation (I cannot speak for
others, so other developer please add or correct!):
* Nokia did not give a long term release date but expects developer to
just be prepared. I can understand why Nokia does not give long term
release date (the reasons are obvious) but as a result it must ramp up
such stuff far ealier (should have made advertisment and plans for
strategie for diablo much earlier) to get a result in time. Perhaps
annoucement was made but at least me was still suprized by some aspects.
Assume that every (re)action of an open-source developer may take 2-4
weeks, even if a task of 2 hours is to be completed.

* Nokia must realize that the current autobuilder and extra repository
infrastructure possibly still give no real gain for me and possibly
others. As long as I can add *.install files to my own repository in the
downloads section having an own repository is simpler than using the
autobuilder.
The current infrastructure:
+ Is slow (autobuilder needs *much* longer than my local computer)
and currently even seems to be overloaded.
+ There seem to be smaller bugs, making some builds fail for no reason
+ Needs more initial setup to be usable
+ Cannot handle dependencies, neither does it seem to compile packages
in queueing order (so having one library and multiple depending
applications I cannot "upload and forget").
+ I cannot push packages for multiple OS versions on the same time
via the web interface (will take a look at dput)
+ Does not build packages for older distributions, so I still have to
compile packages for older OS version myself, having twice the effort.
+ Does not support other niftly stuff like lint etc...
+ Integration with downloads could be much better
+ extras is more a mess than an really atractive place to be for my
packages ;-) Quality was requested but control and enforcement seems to
be weak.
+ Suggestions for improvements will not be rejected but the reaction
suggests that things will take a long time (and suggests that I can slow
done pace, too, which is definitely the wrong signal ;-)).

The current situation is the direct result of this and for me personally
no suprise (in fact in the last big discussion I participated I even
hinted that this will happen).

However I don't think that keeping things the way they are or at least
slowing down the development of this the totally wrong way (under given
goals). This is the totally wrong signal and will kill extras
repository. Instead  a way to accelerate the development and invest in
improvement must be found. It looks like Nokia is waiting for the
community to jump in, but possibly critical mass is not yet reached for
this. Or why we discuss 2010 agenda?

How?

If the community cannot help or does not want to help (because hacking
software is cooler than improving infrastructure?) Nokia possibly must
either help,  extpectations must be lowered, or another strategy must be
found. For me it looks like we are still no team.

Another question: How could the proposed community board could help in
such situation?

-- 
Gruß...
       Tim