If a 3d modeler needs to get rid of polygons in a 3D mesh, but don’t want to add triangles, here’s two good ways. I call them “3-to-1” and “2-to-1 corner”. I worked these out in 1988 with Joe Formicola and Ken Fedeson, building finite element models for CFD and FEA. Many other 3d artists have discovered or learned these topological tricks, since (and probably before then, too). I just noticed how the first one resembles the shadow of a hypercube.
Is it gauche to repost my comment on my own blog?
I wrote this on this blog (which I enjoy), then decided it’s good enough to post here.
My favorite code advice is “if you need something >3 times, make it a function/object. Otherwise, re-code it inline.” Without this rule, I 1) underestimate the cost of developing reliable objects; 2) overestimate the effort required to code the same thing a second time; and 3) overestimate the likelihood that I’ll need to that function again (in that language/context) in the future.
What’s yours?