Re: polygons question


RobertSzefler wrote:
> If you are not too precision-paranoid, you should really check out this:
> http://www.cs.unc.edu/~zhangh/hom.html

ha ha, I am supposed to build map for doing something I do just to
avoid building such a map :) really, I can't use it because I don't
know the scale at wich my output will be rendered. ideally, rendering
engine code (which is beyond my control) should do that mapping, but in
this case I *know* that it is dumb and won't do that, so I have to
prevent it drawing invisible stuff.

> Another approximate approach would be to shoot a bunch of rays (while
> paying attention to good distributedness, Halton seq.?)... All the uncountable ray
> tracing acceleration tricks apply here.

"good" distributedness? I suppose this way I would have to distribute
stuff along the edges. trade-of with precise method is not doing union
part, which I'm not sure how worth is that... also, that after-scaling
problem, again. I have no idea of pixel size related to polygon. do I
hear "dumb"? so? I have my constraints to work with :(

> Another question is how much preprocessing time you have and is your
> data static for many queries or the other way round.

it has to fit in a few seconds or two for ~10k polygons. but it is in
flash, which is sloooow times and times compared to, say, c++ :(

.

Relevant Pages