Some discussion with Oyku suggest that using the canonical loop(n) to look for cycles might not be the best thing to do. loop(n-1) might be preferable. The reason is that with loop(n) we miss the finite-order braids.
The argument against using loop(n-1) was that the generator n-1 acts trivially. But maybe this doesn't matter?
This is fixed: loop(n) no longer has a basepoint. Can't really tell which commit since the conversion to Git seems to have lost some per-file history. Grrr.
Some discussion with Oyku suggest that using the canonical
loop(n)
to look for cycles might not be the best thing to do.loop(n-1)
might be preferable. The reason is that withloop(n)
we miss the finite-order braids.The argument against using
loop(n-1)
was that the generatorn-1
acts trivially. But maybe this doesn't matter?