Re: [Gnu3dkit-discuss] Advanced Rendering Interface

[Gerard Iglesias]

Happy new year and best wishes....

On Thursday, December 26, 2002, at 01:18 PM, Philippe C.D. Robert wrote:

> > OK you want to make a new QuickRenderman implementation, not a bad 
> > idea, in fact maybe it is a real good idea and make sense on the Mac 
> > platform ;)
>
> I am not sure if a pure QRM implementation would be worth it, but I 
> have to look more into this... anyway, the most interesting part would 
> probably be the RenderMan shader language.

Sorry to be more precise I was thinking of implementing the Renderman 
API in OpenGL, that is using real time rendering technique to implement 
shaders, but that is true that it doesn't mean getting something 
running in real time....

QRM was a simplification of the rendering process to get something 
running at a correct speed on a NeX 68040 ;)

And sure the tough part will be to implement renderman shader in 
OpenGL, in fact a very interesting engineering/research problem, good 
point and a subject by itself, good ....

> > Well, ok maybe if people want to make new stuff very powerful, i.e. 
> > access to the low level, they will need to expand the 3DKit by 
> > programming in C/ObjC/OpenGL, maybe ?
>
> This is my idea as well - to make the kit be a foundation for 3D 
> graphics which can easily be customised for personal needs. Thus I 
> also decided to remove some functionality found in the old 3.x 3DKit 
> as it can easily be integrated on top of the core 3DKit (ie. LOD 
> groups, 3D sound rendering, switch groups, ...).

Agreeing completely with this approach, we need the brick and mortal, 
not the building....

But like with the SJ Metaphor we want to begin to make the building at 
the right level and with the good tools...

> > Fine, Ok I will not have the time to write a new Maya app in my spare 
> > time ;)
>
> You have some spare time? Where did you get this from...:-)

I don't know ;)

> > In the same field, I wonder if it would be interesting to put the 
> > math algorithms we can find in '3D Game Engine Design' in the 
> > GeometryKit ?
>
> If it is additional stuff, yes why not - esp. having more physics 
> functionality would be cool! As for the current C implementation, I 
> think it is quite well optimised, but then there are always ways to 
> improve..

I was thinking about new geometrical stuff, there is also physical 
stuff in it, a simple port would be possible, I don't think that the 
given code is very optimised, and I will take a look at the licensing 
agreement.


Regards

        Gerard