[Bf-committers] Nodify logic , a GSOC project of idea list , need info

Dalai Felinto dfelinto at gmail.com
Fri Mar 19 18:00:33 CET 2010

Hi there,
In case this is not a task for a GSOC (what I think may be or may not,
depending on the specific goal and current state of Benoit's work) those are
more interesting projects someone could pick up for GSOC BGE projects:

1 - Hardware Skinning - GLSL vertex shader built-in/out of the box system to
speed up bone animations. I have some work on that (VBO working but in the
need for optimization and a working example with a real .glsl shader). There
are still plenty of work to be done in this project. Highly needed for BGE
in my opinion.

2 - real DDS support - BGE currently supports DDS but doesn't make use of
its advantages. Also different textures (alpha/normal/color) may need
different compression methods. On that scope I would also add
creation/convertion of DDS textures from Blender UV/Image editor. * note: I
don't use dds myself but I alway heard about how great they are.

Is the wiki open for adding those ideas?


2010/3/19 Charles Wardlaw <cwardlaw at marchentertainment.com>

> > A general purpose nodal mechanism will be great for a number of things
> > and may even become a project on its own rights. But in order to do
> > that it should be:
> >  1. BSD licensed et al.
> >  2. Lightweight - using lua instead of python
> >  3. designed from the ground up as an external module, not using any
> > of the current blender infrastructure but highly similar in design and
> > concept (in the context of the game engine)
> I think that might be a bit above and beyond the GSOC boundaries, though.
> :)
> The fact of the matter is, a lot of the node work is already done and
> in Blender already.  It's more a matter of generalizing what's there
> and making connections work based on the bpy.data pathing.  Going for
> a separate library so nodal logic can be integrated into external game
> engines sounds like a wonderful feature, but as there are already
> open-source packages that accomplish such things and since the number
> of Blender users who would benefit from that feature is a small
> percentage (IE, people writing their own engines or using Open Source
> engines, or possibly Torque), it might be better to focus on generic
> nodal connections for within Blender only and add externalizing the
> functionality as a separate project for down the road.  The generic
> system project would have immediate impact on workflows for all
> Blender users across the board.
> ~ C

More information about the Bf-committers mailing list