[Fluxus] more collada commits

[Artem Baguinski]

On Nov 29, 2007 10:54 AM, Dave Griffiths <dave at pawfal.org> wrote:
> Not yet, but the idea is to support collada as fully as possible.
> It's a matter of time, next up is materials, followed by animation - I
> guess physics will be after that.
>
> It's slow at the moment though, the moon buggy example takes around 10
> seconds to load - the reason is that the geometry loading is doing a _lot_
> of processing to convert the data from multiple indices per vertex data
> type to one uniform index for all vertex data types (which is optimal for
> rendering speed).
>
> I need to profile it again, but options are to live with it, rewrite bits
> of it in C++ or use Collada as an intermediate format and have a native
> fluxus scene file which is laid out favorably to how fluxus works -
> neither of which I particually like the sound of.

I though the idea of collada WAS to be an intermediate format to carry
as much as possible from package to package.

In VGE we use scheme-based native format for scenes (and ogre formats
for meshes, materials etc - because we use ogre for rendering). May be
fluxus could use scheme for everything , may be with optional text
friendly encoding for large chunks of binary data, and an external
converter tool to convert from collada to that.

> > does it support physical properties too?
> >
> > On Nov 28, 2007 11:47 PM, Dave Griffiths <dave at pawfal.org> wrote:
> >> Lots of collada scene import fixes, fluxus now loads the complicated
> >> sample scene: http://www.flickr.com/photos/7905273@N07/2072475786/
> >>
> >> I also fixed some unitialised variables in the text editor, nothing
> >> which I could see which would cause problems, but maybe worth checking
> >> for those with problems (I'll look into this a bit more next)
> >
> >
> > --
> > cheers,
> > artm
> >
> > http://lab.v2.nl/
> >
>
>
> http://www.pawfal.org/dave/
>
>



-- 
cheers,
artm

http://lab.v2.nl/