This is very cool and I would like to thank you for putting this together. I am a coder and I feel intimidated by POVRay. I know I shouldn't but it seems to me that you need something interactive when making objects. I always thought that you were using something interactive to build it with and then using POVRay as the renderer to describe the scene and put all the objects into it. But if your doing them all by text code then that is quite something.
I dont think POVRay is unique at all being a text language for rendering. Most of the high end stuff does it this way but uses a GUI on top to make the stuff and build it for you. I think Renderman and Gelato are both like this. I also think that even Blender under the hood is kind of like this too. Its certainly trying to build some kind of XML description of the scene. As far as I can tell, the big studios use something like Maya / Alias to build the scene and then renderman as the scene description language then its sent to the render farm to get generated. So learning POVRay is probably a good idea if you want to make a film as it is done in that same kind of manner.
But I would still like to know how to build models in some GUI based thing before sending it to POVRay to render. You ideas about the auto generating towns are a good idea to do in POVRay since you can make a program or script to generate the scene text and then let the engine render it out. You don't have to get involved with the rendering at all. Its like the 3D equivalent of ImageMagick. So what I would like to know is, if you are building something complex, can you get some kind of feedback or preview of what your doing before letting it render the whole thing out ? A lot of what I build is moderately free form and you drag points out and move it here and there by eye. You cant do it with numbers. Just as an instant example, I was making a pair of sunglasses the other day. What process would you go through to make these in POVRay or is it just not geared up for this kind of thing ?