Would he need to code them himself? Or would he just need to turn on/off some stuff?

This program is incredible!  Keep up the awesome work bwang!

I think the writer of the client program should write their own collision detection routines, the reason being some games might want custom routines (walk through walls, physics, etc.)

Aaah ok, I think it would be cool if some people could write routines for others, though, so people can easily do games without having to code much. Some sort of game maker.

and a portal clone for Nspire would be cool. Wouldn't that be hard with raycasting, tho?

The main issue i thought about is trying to view the floor at parralel angle, which may seem weird in raycasting, especially when close to walls. Try Duke Nukem 3D with no mods and look up near buildings. Idk if DN3D is raycasting, but look at the lack of perspective for walls when viewing up/down

I dont think DN3D was a raycaster.

Knowing how RayCasting looks like and works, I think DN3D was using a raycaster.

If bwang implements walls at variable altitudes, we can start building doors, corridors, bridges, rooms and so make the virtual world looks like in DN3D.

I could always render the floor at a very low resolution, or cheat and use an image.

I could always render the floor at a very low resolution, or cheat and use an image.

Using an image for the sky (background - vertical) can give a realistic effect.

But it's not the same thing for the floor...
Unless your image is very simple, it won't look realistic while moving, especially while rotating.

If bwang implements walls at variable altitudes, we can start building doors, corridors, bridges, rooms and so make the virtual world looks like in DN3D.

I've been puzzling about this and thinking whether I should add it. On one hand, the worlds would look a lot nicer. On the other hand, implementing walls at variable altitudes is non-trivial, and without supporting features (such as the ability to look up at the ceiling), would not add much to gameplay, not to mention adding a bit more overhead (although not much) to the main rendering loop.
What do you guys think? What should I add next? I'm thinking doors and multidirectional sprites.