[vos-d] selecting development priorities

[Peter Amstutz]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I'd like to propose a criteria to help decide development tasks from 
version to version.  One thing I'd like to do (and I say this every 
time...) is to have a more nimble development process, so that releases 
happen every 3-4 months instead of every 6-8 months.  We can do this by 
reducing the number of tasks required for each version.  This shouldn't 
slow down development much (although I should note that releases do 
involve some overhead) and will improve our visibility and "liveness" to 
people who are interested in VOS but not following it closely.  It's also 
easier to get people to try out stable releases rather than always 
checking out from CVS.

So what I propose is that for each version, we pick exactly one task from 
one of each of three categories, and make those the goals for that 
version.  These categories are VOS core (improvement to libvos itself), 
developer-visible features (new library/plugin features), and user-visible 
features (noticable changes to ter'angreal itself).

For the core changes, I'm currently working on the way messages are 
dispatched to methods on the target object.  This should produce immediate 
savings by significantly reducing the memory overhead on Vobjects and 
making vobject creation faster.  I may look into other ways we can reduce 
the memory overhead.  I want VOS to be able to scale up to hold 1000000 
vobjects or more, but currently it chokes at as little as 10000 due to how 
to the fact that every vobject has several kilobytes of overhead 
associated with it (and it really doesn't need to have that much...)

For developer-facing changes, I intend work on the SWIG Python bindings 
and related support (omnivos plugins and such).

For user-facing changes, I want to see loading at least static geometry 
from VRML/X3D working, so that we have access to some of that content.  I 
believe that there are at leaste a few VRML people out there (like 
Hermetic?) who would be interested in using VOS if they could design their 
worlds in VRML.

Maybe we can sneak OS X support under "user-facing changes" but that may 
yet be a large amount of work, so while we'll continue to work on it, I 
don't want to make it a blocking task for 0.24.

How does that sound for 0.24?  Everything else we've been talking about 
will get done, but I just want to lay out the next 3 months or so of 
development.  Also, Jason Moyers is working on a design for skeletal 
models (and animation) so we may have that as well, depending on the 
timing.  Everything else is pushed to 0.25 and beyond.  I will update the 
VOS roadmap to reflect this when I get a chance.

[   Peter Amstutz   ][ tetron at interreality.org ][ piamstutz at anteon.com  ]
[Lead Programmer][Interreality Project][Virtual Reality for the Internet]
[ VOS: Next Generation Internet Communication][ http://interreality.org ]
[ http://interreality.org/~tetron ][ pgpkey:  pgpkeys.mit.edu  18C21DF7 ]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)

iD8DBQFEUNByaeHUyhjCHfcRAtshAJwK2F5zv3KhB8fGe0Rz61tnKWQPUACfeovV
B26WYw3tidQIPN1LRFE+bYs=
=gDFV
-----END PGP SIGNATURE-----