a575@mindlink.UUCP (Michael G. Henders) (10/02/90)
Sorry, I've lost the original postings here, but... Someone was looking for hidden line routines suitable for use with a plotter, and someone else suggested a couple of possibilities, and referred in passing to a book by `someone with a Dutch name', for which he couldn't remember an exact reference. (Did you keep all those `someones' straight? :-) ) I have the book; it's `Programming Principles in Computer Graphics', by Leendert Ammeraal, published by Wiley. Copyright date is 1986, and the ISBN is 0 471 90989 0. It does have C code for a hidden-line routine, which is why I bought it, but I haven't had time yet to try the code. Just for your reference... -- | Michael G. (Mike) Henders | "The quizzical expression on the face of | | Vancouver, B.C., Canada | the monkey at the zoo is due to his | | a575@mindlink.UUCP -or- | wondering if he is his brother's keeper, | | uunet!van-bc!rsoft!mindlink!a575 | or his keeper's brother."
jroth@allvax.dec.com (Jim Roth) (10/04/90)
In article <3397@mindlink.UUCP>, a575@mindlink.UUCP (Michael G. Henders) writes... >Sorry, I've lost the original postings here, but... > > [ ... followup giving Leendert Ammeralls book ... ] Thanks, I'd forgotten to follow up on that... I bought the book for exactly the same reason Michael did - its hidden line code! Here is the citation for the COSMIC software I mentioned: M87-10038 - NASA Dryden Flight Research Center Hidden Line Computer Code D. R. Hedgley I don't know if the source is online anywhere, though I have a copy of the indigestable FORTRAN code itself backed up somewhere. To decipher it you'd have to get a copy of the technical report - the code has *no* comments! We're talking dusty deckware. I also wanted to mention that plane sweep algorithms while rather neat in theory can have problems dealing with degenerate cases (multiple edges at a vertex, edges with an endpoint epsilon from another edge, etc.) which are a real headache in practice. Though there has been some work on resolving this (so-called "simulation of simplicity" by Guibas et al) I'm unsure how practical such solutions really are. In 2-D it is possible to use rational arithmetic and get a reasonable and reliable implementation, but in 3D it looks bad. I haven't gotten very far with such problems yet. Making plane sweep algorithms reliable in the face of such problems is an active area of research. - Jim