Re: Some perspective from the cheap seats...

Around 15 o'clock on Mar 25, Allen Akin wrote:

> You were able to reconcile a lot of those conflicts, and explain where
> the OpenGL semantics fell short in some others.  But I still would have
> preferred adding a few small OpenGL extensions and re-using the rest of
> the 3D infrastructure rather than creating a new parallel
> infrastructure.

We may yet have a chance to give it another try -- the Xc library sits 
atop several different rendering systems to provide a unified API for the 
Xr library (or perhaps other libraries).  It would be interesting to see 
if Xc could be targeted to OpenGL, or what extensions would be required in 
OpenGL to make that possible.

> Every project is short-staffed, so reducing duplication of effort would
> have been good.  And OpenGL as a platform has decent availability; shipping
> on cell phones next year, for example.

OpenGL still has the perception of being a large API with a lot of 
required functionality that isn't needed by 2D applications.  As I 
understand it, the cell-phone version is a subset of the standard OpenGL 
API.

One other issue with OpenGL is that DRI-based implementations limit
accelerated graphics to local applications using direct rendering.  I hope
that will be resolved in the future, otherwise using OpenGL for all 2D 
operations isn't feasible.

> (I still owe you a report on using OpenGL to render antialiased text --
> something I've had working since November, but need to design an experiment
> to quantify the differences between several approaches.)

I'd be very interested to see this in any state; that's obviously one of 
the key functions in Render which is "hard" to do in OpenGL.

-keith