On May 2, Keith Packard wrote: > So, we have the following constraints: > > 1) Sum of areas from arbitrary tesselation is unity > 2) Invarient under translation > 3) Should look "reasonable" > 4) 64 bit arithmetic > 5) Within a factor of 2 of the fastest software algorithm that > meets requirement 1) and 2)
Thinking about implementations, I can think of several things that I believe meet these constraints. Some are faster than others, and I'm guessing that any of them would "look reasonable". The difficulty though, is not coming up with the implementation. The hard part is coming up with a specification that is precise, (assuming that by precise you mean that every pixel value can be explicitly determined from the spec.), that allows flexibility in the implementation. Most of the implementation ideas I have could not be precisely specified other than giving more or less the exact implementation. It seems that to come up with a precise specification and still allow flexibility in implementation we must nail down criterion #2 more precisely. For example, what if the spec said that the alpha value computed must be within some margin of error of the pixel area covered. Obviously, pixel area coverage is not the ideal basis for specification, but it would be a reasonable standard against which to verify and implementation. And, any better solution based on anti-aliasing would likely be within the margin of error from pixel area coverage. So, that would be moving away from what I think you had in mind for the definition of "precise", but it would prevent the specification from mandating a non-useful implementation. Would that be a reasonable tradeoff? -Carl -- Carl Worth USC Information Sciences Institute [EMAIL PROTECTED] 3811 N. Fairfax Dr. #200, Arlington VA 22203 703-812-3725 _______________________________________________ Render mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/render
