🚀 go-pugleaf

Normally for 3D transformations, it is convenient to use a 3x3 matrix
with some constant items at the left:

	| a b 0 |
	| c d 0 |
	| e f 1 |

Assuming that your vectors are of the form [x y 1], you can then perform
translations as well as rotations and the operations can easily be combined.
PostScript does all it's scaling and transforming operations using 2x3 matrices
that are filled to the above form.

Doing just rotations and scales doesn't make much sense, since you can
save a lot of multiplies by just adding up the rotation angles and
multiplying the scaling operators instead of using matrices.

I'm not aware of any personal computer games that would use matrix
transforms on bitmaps. I seriously doubt that Wing Commander is using
anything like this. From what I have heard (I have only seen the game
once or twice), all the shapes are stored as pixel maps.

Once you move into polygon or vector based graphics, matrix operations
become more plausible. As you noted, the nice thing is that you can
scale and rotate (and translate) as much as you want. If you have seen
the arcade game "Assault", then you know what kind of effects are possible
(although Assault has some hardware assistance and probably uses a DDA
to scale and rotate the pixel maps).

-- 
  Juri Munkki					       Windsurf: fast sailing
 jmunkki@hut.fi					      Macintosh: fast software