Tom Pledger wrote:
> Andy Gill writes:
> > Galois Connections submitted an entry to the ICFP programming
competition.
> > Here are some notes about our entry.
> > [...]
>
> Impressive! My 815-1401 line(*) weakling certainly doesn't match that
> speed.
That's about how we feel when we look at the O'Caml entry ;-)
>
>
> Did you think any of the sample output files (final versions) were
> wrong? I'm doubtful of the far edge of the right die's shadow in
> dice.ppm.
Hmm... it looks OK to me, and we get the same edge. What were you expecting?
>
>
> What did you do about common surfaces between the elements of a CSG
> intersection or difference? For example, what does the intersection
> of a blue unit cube and a red unit cube look like? I made it blue
> with lots of little holes through it. :-(
>
Depends on the cube's mood ;-) As I recall, they said it was undefined which
surface you get when they coincide. In ours, the choice depended on how the
object was constructed, so it consistently picked whichever cube was
`leftmost'
in the data structure.
--Jeff
