🚀 go-pugleaf

rjwalsh@maths.tcd.ie (Robert Walsh) writes:

>       ----
>      /    \
>      /    \
>      ------

>then I should get a bitmap appearing to hover in three space. The
>bitmap should be scaled squashed at the top and fatter down towards
>the bottom. If I map it onto a square, it should be an exact scale
>replica of the original bitmap.

This is asking for something completely different.

Assuming you meant this sort of thing:

       ____ 
      /    \
     /      \
    /________\

This can be thought of as either a square with the top squashed in
(so you just do a normal resize, but a different size for each line),
OR it could be as you suggest, a 3D inclined plane - the same as
the above, but with the vertical density of lines decreasing towards
the "front", or "bottom" of the trapezium.

>Extending the above further, I should be able to map the bitmap onto
>something like this:

>      /\
>     /  |_
>    /     |
>    \______\

>and have it scaled properly all across it.

"properly"?

>Am I asking to do something that only a Cray could manage in realtime? Is
>this feasible on a 33Mhz 386?

The first to can often be done in real time, since they are just minor
variations on normal rescale algorithms, but a general techniue, using
some set of constraints about how to "fit" an image, would be more
difficult.  I doubt you'd need a Cray for that sort of thing, unless
you want a fair few of these fitted shapes, all large.


Warwick
--
  _-_|\      warwick@cs.uq.oz.au            / eastings genistas /
 /     * <-- Computer Science Department,  / giantess seatings /
 \_.-._/     University of Queensland,    / teasings tsiganes /
      v      Brisbane, Australia.        / -------- -------- /