Hello, As I've mentioned in previous postings, I've been working on plans for a model of a dodecahedron with sundial faces. I have no posted some files, both source files and the output in PostScript and PDF, as well as compressed versions of the latter.
I must admit that they are as yet incomplete. I was hoping to calculate dials for all of the more-or-less southward-pointing faces (given a particular orientation of the model), before posting anything, but even doing just one face has turned out to be a great deal of work. If anyone would like to take a look at them anyway, they are available here: http://www.gnu.org/software/3dldf/sundials.html#Dodecahedron_Model or here: http://www.gnu.org/software/3dldf/polymodl.html#Dodecahedra They include diagrams illustrating how I went about finding the hour lines, but the explanations are rather scanty. Having looked at Denis Roegel's article, I've begun to wonder if my approach is perhaps naive and my results incorrect. It seems to me that intersection of the plane through two points on the gnomon and a point on an hour line on the equatorial dial and any other plane or surface would be the hour line on that surface. Am I missing something? It does seem a bit boring to have the same sort of dial on each face. Suggestions for making it a more interesting would be welcome. I'm afraid that I haven't implemented some of the fancier features of METAFONT/MetaPost in GNU 3DLDF, such as the `whatever' feature Denis makes use of. 3DLDF, like most programming languages, cannot solve linear equations by itself and one can therefore not write equations using variables with unknown values. I've gotten used to the more conventional approach required when such features are missing. I think Donald Knuth doesn't get enough credit for them; they are both clever and unusual. On the other hand, 3DLDF defines more data types than MF/MP and provides operations to operate on them. For example, it defines a data type `plane' and an operation for finding the intersection of two planes. I think it's interesting to see a different approach, also based on MF/MP. I don't think one is better than the other; each has its advantages and disadvantages. Laurence Finston --------------------------------------------------- https://lists.uni-koeln.de/mailman/listinfo/sundial
