Folding in R · Chapter 1

The pattern is the answer. The view is only what a method gets to see.

Panel A is a flat sheet of paper with its creases marked. Panel B is the same sheet, folded. Turn the folded sheet and watch A: the facets you can actually see light up, the ones hidden behind a fold go grey. The pattern in A never changes — only your view of it does. That gap, between the structure that is really there and the part any one view exposes, is what this book measures.

A The flat sheet —
B The same sheet, folded drag to turn
0.65
visible in B hidden behind a fold mountain crease valley crease

The fold is a Miura-ori, and it is rigid: every facet stays flat and all the bending happens at the creases. So the folded sheet in B is isometric to the flat sheet in A — distances measured along the surface are identical in both panels. Worst departure from that, across every facet and every fold angle shown here: —, which is floating-point noise.

That is the property the book is built on. Because A is exactly the unfolding of B, A is not an approximation of the right two-dimensional answer — it is the right answer, known in closed form. Which means you can hand it to a dimension-reduction method as a test with a marking scheme, instead of an opinion.

Every coordinate on this page was computed in R by the book's own code and checked before it was drawn; the browser only renders. Slide θ to 0 and the sheet flattens back into panel A exactly.