As a primarily Python programmer I would have to disagree with this vehemently. In my experience, the Python bindings on existing engines tend to be a second rate affair in terms of what is exposed and (....)
True, this is what I experienced as well... With the exception of soya3d, I really like the framework, although the function/object naming is not very consistent - but it's VERY easy to load a heightmap and get a charactor blender-model to walk your earth (10-20 lines including ground-collision) http://home.gna.org/oomadness/en/soya/index.html When I last looked for something more than PyOpenGL or a 3D
engine to use, I ended up using Nedu as a base. http://trac.zeitherrschaft.org/neuro/wiki
I personally would not recommend neuro, since it seems to be tailored for the "I need the lastest pixel shader magic" - demo-hackers. Doesn't have very game-specific stuff by itself - but I didnt browse the docu that extensivly.
> My reasoning is: > > -That bad 3D is worse than good 2D;
For all my projects, be it for professional or personal use, I start with a simple (!) 2d engine to test out the basics (gameplay mechanics you would call it for a game)... keep it simple. check out this guy, who did prototypes for Spore - he has a blog online somewhere with a lot of screenshots of Spore prototypes... they are composed of 2d lines, colors and some numbers. thats it - doesnt even look like a game. http://www.gamasutra.com/php-bin/news_index.php?story=11628 take a look at http://www.kamilche.com for
> some screenshots of a 2D game I'm currently developing.
mh.. couldnt find screenshots, but the NPCs look nice indead :-) greetings simon
