Right - one of the implications is that the parts need an _absolute_ precision limit, e.g. the author should limit 'meaningful' use of triangles to a spacing between vertices of > X where X is a fraction of an LDU, after considering all transforms. Then X needs to be big enough that the accumulated error of some number of transforms doesn't exceed X.
T-Junctions re-visited
« Next Oldest | Next Newest »
Users browsing this thread: 5 Guest(s)