Life after Opie 1.2.0

Hi fellow developers,

now that through Familiar Linux 0.8.2 and OpenZaurus 3.5.3,
Opie 1.2.0 is finally released to the general audience, we
can start the great Palaver about what we want to do next.

>From discussions on IRC in the past weeks, I have got the
impression that basically there are three kind of interests
out there:

a) fixing bugs and polishing what we currently have

b) adding new applications and playing with new concepts
   based on the current set of libraries

c) starting from scratch with something revolutionary rather
   than cloning existing environments

Abstained from the road I personally would like to pursuit, now
a few statements about how we should organize either way:

a) Bug fixes and polishing should be done in CVS HEAD -
when there is a critical mass of fixes reached, the tree
could be tagged as 1.2.x and be prepared for another release.

b) New applications and playing with new concepts can also be done
in CVS HEAD, although to prevent destabilizing the codebase,
for larger and substantial changes I recommend adding new directories
instead of using a CVS branch. For instance, I've heard _alwin_ wants
to look into a whole new Launcher based on Qt/E 2.x - this should be
done e.g. in core/launcher2 together with e.g. core/libqpe2. There
are major advantages in using new directories as opposed to doing a
branch: You can diff more easily, you can build stuff in parallel
do to regression testing more easily, you can see how far things are
without having to work with two trees and tags etc. Having participated
in the last Opie branch hell, I can assure you that merging branches in
a system like cvs is harder than just doing it in two different
directories
of the same branch.

c) Starting from scratch needs to be done in a separate repository. We
could
use this change to make some important things happen, i.e. abandon cvs,
use a more sane directory structure, start with individual versioning of
apps,
start with more self-contained apps, make tarball releases happen, etc.
The major question for this road though is what toolkit should we base
something
like Opie-2 on. There are several alternatives and this should be
discussed in a
separate thread.

What I like to do in this thread is to discuss whether we go for one of
this roads
or whether we will try to pursuit all in parallel. My opinion on that is
ambivalent.
On the one hand I think it is an important feature of free software
projects that
there is room for everyone's contributions. Opie should be no different.
However
one the other hand I am not sure though if the general interest (read:
developer resources) in Opie justifies going all three roads in
parallel. What do you think?

Ah - one last word about the structure of this discussion first: It
could be
possible that this thread may become pretty longish, so whenever
you start to spinoff a discussion (i.e. for each of the three
alternatives a, b, c), then _please_ adjust the subject line and don't
mix it with this thread.

Cheers!

--

-- 
Regards,

Mickey.
------------------------------------------------------------------
Dipl.-Inf. Michael 'Mickey' Lauer <mickey <at> tm.cs.uni-frankfurt.de>
------------------------------------------------------------------

_______________________________________________

http://opie.handhelds.org/cgi-bin/moin.cgi/DeveloperWikiIndex

Opie-devel mailing list
Opie-devel <at> handhelds.org
https://handhelds.org/mailman/listinfo/opie-devel