Raphael Collet wrote:
I feel Mozart has the best chance to survive if we welcome everyone
willing to contribute with open arms. In the past, there have been
cases where potential developers/contributers where driven away,
simply because there existed no official way for them to add what
they missed (e.g., I have in mind past proposals concerning Unicode
support). Meanwhile we have an official way (MEPs). However, we
should make sure that this process is actually working. Moreover, in
my personal view it would be best if potential contributers can get
the feeling they are indeed welcome -- if the Mozart Board appears
not to care than this may have the opposite effect.
I am a more pessimistic about the future of Mozart. I think we have a
very serious maintenance issue here. Mozart's implementation is too
complex and too obscure for newcomers. It's too complex even for
experienced developers, in fact. I still have myself some trouble for
understanding certain parts of the code. Moreover it lacks modularity
and documentation.
This is unfortunately all too true. The UCL approach, with limited
resources, is to focus
on interesting topics and factor them. We have made some progress in
factoring out
the interesting parts from the complex core of Mozart: the new
distribution subsystem
and the Gecode port for constraints. Each of these is supported by a
project:
SELFMAN for distribution and MANCOOSI for constraints. The new distribution
subsystem is the heart of a new approach to systems design with
concurrent components
and feedback loops. This is a quite interesting topic: my paper at the
BCS 08 conference
explains the vision. The MANCOOSI direction is interesting too,
especially since we
have new and powerful graph constraints (c.f. dissertations of Luis
Quesada, Grégoire
Dooms, and Stéphane Zampelli) and we are applying them to manage complex
system
distributions.
The Mozart board no longer reflects the structure of the Mozart
community. Some of
the old partners have drifted to other research areas. I think that at
this time, the UCL
group is the main one driving Mozart. Probably the Mozart governance
should be changed
to reflect this: just UCL + some truly interested partners should make
the decisions and do
the development.
I will leave the UCL in September, and start a non-academic career. I
will keep around for some time, but I cannot guarantee anything. I am
not sure I will want to spend my free time on hacking. Boriss Mejias,
Gustavo Gutierrez and Yves Jaradin will continue to maintain and
develop Mozart, but I think they will need some help...
We will wish you well and hope that you can continue to give advice from
your store of
Mozart knowledge!
Peter
_________________________________________________________________________________
mozart-users mailing list
[email protected]
http://www.mozart-oz.org/mailman/listinfo/mozart-users
