1. 7
  1. 2

    one of the most interesting ideas I’ve come across in a while! it seems vaguely similar to the use of compression techniques to identify copy/pasted code, but I’d never have thought of extending that to this sort of optimisation.