From: Jon M. Taylor <taylorj@ecs.csus.edu>
  To  : ggi-develop@eskimo.com
  Date: Tue, 18 Aug 1998 12:45:14 -0700 (PDT)

Re: 3d

On Tue, 18 Aug 1998, teunis wrote:

> On Mon, 17 Aug 1998, Jon M. Taylor wrote:
> 
> > On Mon, 17 Aug 1998, teunis wrote:
> 
> [clip]
> > > 
> > > And on the -really- high-end cards (ie. 3DFX-2 and the like :), the entire
> > > card is programmable ergo next response:
> > 
> > 	How's this?  I still am stuck with a Voodoo I.  Does Voodoo II
> > have a coprocessor of some kind. 
> 
> Hmm - could be wrong.  Last I heard the 3DFX series ARE programmable.  At
> least in the 3D-engine component...
> (there are chipsets that are programmable - I was just fishing for
> examples)

	Yeah.  There a high-end workstation vide cards that contain a CPU
and some RAM on the card, and you download the whole X server to the card. 

> > > > > There allready exist a minor API in Mesa for rasterization, which means
> > > > > a driver can accellerate these.
> > > >  
> > > > I like to see library independent accelleration. Threw ioctl not special
> > > > library stubs. 
> > > 
> > > Not likely.  There's really no compatibility below the OpenGL/whatever
> > > layer.  None whatsoever.  Even OpenGL doesn't really cut it with some
> > > hardware (OpenGL is too bare-metal for them or they're planar hardware)
> > > but it'll do for now.
> > 
> > 	See my z-buffer/polygon/pluggable shader idea elsewhere in this
> > thread.  It'll work closely enough with most hardware.
> 
> It's a good idea - now to implement and see how it'd run :)
> [the big advantage to OpenGL is an existing API / commandset to start
> from]

	I had to implement z-buffering, polygon rendering and pluggable
shaders in a class last spring.  It is quite clear and simple. 

> (aside: My vote goes towards a GGI3D list...  Too many messages annoying
> the non-3D crowd)

	I'm game.  Anyone else?

Jon

---
'Cloning and the reprogramming of DNA is the first serious step in 
becoming one with God.'
	- Scientist G. Richard Seed