3D graphics programmers using Python?
Brandon Van Every
vanevery at 3DProgrammer.com
Sun Feb 2 22:27:59 CET 2003
> PyGeo was developed for the visualization of geometric concepts of
> projective geometry, and then the geometry of complex numbers. Its
> architecture was evolutionary - as it was built to serve a
> educational need (first off, my own), and there was no way, I don't
> believe, to architecture it all up front. Refractoring on top of
> refractoring was the process. I am not sure there is any way to
> convince you that the process could not have been as pleasureable or
> (I think) successful if I was working purely in C++ or Java.
I think you knew, going into your project, that your treatment of geometry
was inherently explorational. Python is a sensible language choice for
that. It's rather analogous to what I'll do with game AI someday. I don't
know how I'll solve it, there will be many methods tried. I will first
create a pathfinding and analysis library in C++ for efficiency, then I will
probably use Python to glue different approaches together.
It doesn't sound like balls-to-the-wall performance was a requirement of
your project. Nor did you know that you could rely on a few 3D constructs
to do everything. Really, once my spherical icosahedron is actually a
planet, my job is done. I move on to some other aspect of my game. I do
not need flexibility of geometrical definition, I know it has to be a
> Couple of years into playing with this project, and at some point -
> maybe a year or two from now - I will make the decision as to whether
> I am prototyping or developing.
I am developing. Barring fundamental design mistakes, all code that I write
is "keeper" code. I've improved some of my interfaces by refactoring...
nothing about C++ made that particularly hard.
Brandon Van Every Seattle, WA
20% of the world is real.
80% is gobbledygook we make up inside our own heads.
More information about the Python-list