Re: [grt-talk] Picking targets. Was: Some suggestions.

[Anton N. Mescheryakov]

Nikodemus Siivola wrote:

> Re: hardware acceleration: if someone is willing too look into implementing
> it, that's fantastic. Utterly so.
>
> On my personal list of priorities it's rather low, however: not something
> I'm too interested in pursuing right now.
>  
Personaly, it isn't hard for me, providing that some OpenGL lisp 
bindings are available. CL-SDL will probably do, and it isn't way too 
hard to do on my own, as only a small portion of OpenGL API is 
nescessary. BUT: I completely agree that it isn't something urgent or 
even required for the next release (and porbably a couple of releases in 
the future). Here we come to some interesting point: assigning tasks. Of 
course, grt is a voluntary project, nobody can fire me for... er, not 
much work on my side, but I'm not especialy happy to fly over CVS as a 
ghost over waters or whatever. Currently, I work on somethat more 
clarified version of math.lisp so it'll be more terse and (I hope) 
effective. For example:

(defmacro vector-add* (&rest args)
 `(op3* + ,@args))

and even:

(defun vector-mul (v f)
 (op3* (* f) v))
(defun dot-product (u v) 
 (op* + 3 * u v))

Someone asked me for a patch, if anybody remembers.
But despite probably more compact strucure and ease to maintainance, 
this work is just a reimplementation of already existing code. Roadmap 
to 0.2 and TODO is a bit too terse, feature that grt in general is 
currently lacking, so picking tasks and coordination with existing work 
in progress is somewhat obscure. So let's discuss that to do and in 
which order, or I'm too exacting and lame?

Best regards,
   Anton.