All seems to be working now. Got point vs circle and point vs polygon collisions working as well as circle-circle and circle-polygon, plus more efficient point-in-circle and point-in-polygon routines that don't bother to compute a push vector (for 2D picking).

It is all starting to make a bit more sense now as well. I'm hoping that all this exposure to 2D vector maths will bear me in good stead if I ever get round to attempting 3D games, although I think that 2D polygon based games is going to keep me occupied for quite a long time yet.

I'm probably going to rewrite my polygon map editor, since I reckon I can now do the maths to constrain user-drawn polygons to convex. I'd also like to find a way to optimally divide a user-drawn convex polygon into minimum triangles so that I can automate the conversion from collision shapes into renderable triangles.

Dunno really. All this recent activity has opened up so many new possiblities in terms of feasible games that I'm not sure at the moment which way to turn.

I am tempted to write an Arknoid clone that has non-axis-aligned and non-rectangular bricks, just to show off. I may even call it Seperating Arknoid Theorum.

I really should get out more.