Hi,
I personally like the idea of POV4 moving quite a mile. And it should,
considering the change that came along the path from 3.0 to 3.1 :-). And
for the unix-folks even a through e is very significant.
But reading about extensions of POVRay I usually find a mixture of three
very different problems:

a) Writing a good raytracer
b) Writing a good scene description language
c) Writing a good programming language and interpreter

Each and every one of them is well suited to fill up some books, several
lectures or some non-profit organisation. And in general moving from
'good' to 'best' is impossible.

The main problem in my eyes (and the best point in Eugenes proposal) is
the closed architecture of POVRay. I know, that it's religion, but that
would be a major step.

I happen to switch tools faster than anything else, and I will stick to
create scene files with Perl and matlab, just because I don't want
POVRay to be good in solving eigenvalue problems :-)
Exhibiting a well documented interface to what happen to grow an
object-tree sometime, will open up grand new possibilities and at the
same time bring down the load on the POVTeam.

They then may proceed in writing THE BEST raytracer around.

Axel

Post a reply to this message