c10_h006@jhunix.HCF.JHU.EDU (Stdnt 06) (03/15/89)
Does anyone know of an algorithm to project an arbitrarily large number of dimensions onto a plane in perspective? It seems logical that one could (with the exception of n=2) project any n dimensions to n-1, and therefore project any n onto any other n. Is this true? If anyone has an algorithm to do this (even without perspective, although perspective wouldn't be too much trouble, mathematically) I would appreciate it greatly. Any pointers to literature on the subject of n-dimensional geometry would also be helpful. Gunther Anderson ins_agwa@jhunix.bitnet or g49i3381@jhuvm.bitnet "I will not be pushed, filed, stamped, indexed, briefed, debriefed, or numbered. My life is my own."
jbm@eos.UUCP (Jeffrey Mulligan) (03/16/89)
I'm going to commit the cardinal comp.graphics sin of replying to a posting off the top of my head. From article <1133@jhunix.HCF.JHU.EDU>, by c10_h006@jhunix.HCF.JHU.EDU (Stdnt 06): > Does anyone know of an algorithm to project an arbitrarily > large number of dimensions onto a plane in perspective? It > seems logical that one could (with the exception of n=2) > project any n dimensions to n-1, What is the problem with n=2? You can project the plane into a line; for example you can project the plane through the origin onto the line y=(-1) by: x'= -x/y. Since in 3-D perspective you divide by z, in n dimensions it would seem that you could just divide by the coordinate along which you want to project. As far as doing this iteratively to go from n dimensions to 2, I would say you could do it, but I won't guess what it would "mean." -- Jeff Mulligan (jbm@aurora.arc.nasa.gov) NASA/Ames Research Ctr., Mail Stop 239-3, Moffet Field CA, 94035 (415) 694-6290