DebConf10: Final Call for Contributions!

[Reinhard Tartler]

On Fri, Apr 30, 2010 at 23:14:38 (CEST), Reinhard Tartler wrote:

> Anyway, the deadline seems to be tomorrow night. Please review and amend
> my draft by replying to this mail. I'll compile and submit the final abstract
> tomorrow night based on your replies.

On behalf of our team, I've now registered two events, a 'Lecture' and a
'BoF'; both have probably 60 minutes lenght. Here are the abstracts that
I've submitted for them:

== Brave new Multimedia World ==

Many user expect "proper multimedia support" from the operating system
of their choice. Meeting this is challenging because of the different
requirements users have.  For casual and heavy consumers of movies,
trailers and music, a large variety of video and audio codecs and
formats must be supported.  Professional and semi professional users
expect software that supports *reliable* multi-track recording and
playback - some of them even on slow hardware!

Satisfying these requirements turns out to be pretty challenging.
Several individuals have been working very hard on creating and
maintaining the necessary software in Debian.  During the squeeze cycle,
the two existing Debian Multimedia Teams have merged to a common team on
Alioth: pkg-multimedia. Like other packaging teams, pkg-multimedia has
been (and is still!) suffering from the typical symptoms "shortage of
manpower", "shortage of motivation", etc.  Additionally, there are some
more unique constraints that this contribution will highlight.

First, the complexity multimedia stack on Linux is impressive. In the
past we have seen the competing audio drivers OSS vs. ALSA.  In more
recent times and on a higher level, the traditional enlightment sound
daemon seems to have lost in favor of the more advanced sound server
pulseaudio. As if this wasn't complex enough, the pro-audio community
has developed a soft-realtime capable framework, the "jack audio
connection kit" along with several different and competing
implementations. Almost all of these sound daemons impose different API
and guidelines how to use them for multimedia software. pkg-multimedia
has the the task to "make it just work".

The second hurdle the pkg-multimedia team is facing is with supporting
the latest and increasingly popular audio and video codecs and formats.
Luckily, for the most poplar "modern" audio codecs DSFG "free"
implementations do exist. However, because of patents Debian has
problems with including these packages so users have to resort to
3rd-party archives or other vendors if they wish to use such formats.

In order to provide the best multimedia support for Debian users, the
pkg-multimedia team has been collaborating with derivates (studio64,
ubuntu studio) with packaging important pieces of software in the
multimedia stack and working on a high quality supplement repository
(debimedia). While there is always room for improvement, squeeze is in a
decent shape for current multimedia requirements. This talk explains
the challenges and Debian's approaches to them in detail.



== bof on multimedia patents ==

Debian is doing a very suboptimal job at supporting modern multimedia
content since years. Bugs like #522373 show clearly how horribly
confused we act when it comes to the handling of software packages that
implement encoders and decoders. This confusion has led to contradicting
decisions by both package maintainers and ftp-masters.  Not only they
fail how to interact each other, but also how to communicate these issue
to users, downstreams and upstreams.

The 'multimedia patents' discussion is already under way since long ago
in the FOSS world. The most recent zenith is the Mozilla foundation,
which actively tries to push xiph's theora format as standard codec for
the upcoming HTML5 standard. The FOSS community as a whole however does
not agree that theora is the answer to all problems. FFmpeg for
instance, the building block for almost all multimedia applications that
are shipped by Debian has serious concerns about theora and strongly
favors theora's competitor h264 (as described by MPEG4 part 10) for
being technically superior over theora.  Moreover, most recent
developments [1] indicate that h264 has won this race.

[1] http://blogs.msdn.com/ie/archive/2010/04/29/html5-video.aspx

This BoF intends to bring all interested and affected contributors to a
round table to openly discuss past and current issues. The goal of this
discussion is a vision about what role Debian as a project wants to play
in this discussion. Do we want to continue to keep out of this whole
discussion and leave our users in the cold? Or do we want to take a more
serious part here?  How shall we all, package maintainers, ftp-masters
or the DPL engage, reply or otherwise with other FOSS projects like
Mozilla, Xiph, FFmpeg, the FSF or other distributions on this topic?


-- 
Gruesse/greetings,
Reinhard Tartler, KeyID 945348A4