🚀 go-pugleaf

Disclaimer:
Maybe this falls more in the arena of comp.graphics, but everyone doing
animation work there has huge requirements compared to my own.

I'm doing a camera-oriented (like a driving or flying simulator) animation
program, and I'm trying to build it as bare-bones as possible.  That way,
when graphics cards get faster, there won't be a ton of routines to rewrite.

But what sorts of things are necessary?  I'm not talking about fancy primitives
like ellipses, circles, etc.  I'm not talking about fancy operations like
pattern fills, hidden surface removal, etc.  I'm just talking about a few
routines, as quick as possible, for doing development work.

The list I've got so far is:

	determine graphics environment (used to allow/disallow various modes)
	set graphics mode
	begin double buffering
	swap buffers
	clear buffer
	end double buffering
	return to text mode
	set drawing color
	point, clipped point
	line, clipped line
	polygon, clipped polygon

I've written a set of routines (only supporting the SpeedStar regular card,
in all 256-color modes) and I'm getting ready to start writing C++ code to
draw 3-d objects using transformations and the above primitives.

Are there any other primitives that I should be handling?  Sure, circle and
ellipse would be nice, but (a) those aren't 3-d objects and (b) I'm using
this just for development until something fast comes along, and I don't
want to spend time implementing things that reduce the frames/sec too much.

Is anyone else even interested in doing 3-d animations, or is that
limited to people who own IRIS machines and don't have to worry about
the lower level details?

Scott Mark
sjm@miles.com