My friend Andrew Helwer argues that the best way to communicate an algorithm is with PlusCal or Python. He has “come to believe that both are superior to an ad-hoc unspecified pseudo-math language,” of the sort that researchers often invent for their papers. He recommends a subset of Python that avoids “fancy features” like list comprehensions, and avoids the standard library. He shows an apt example: a 43-year-old algorithm that was specified in pseudocode and has therefore been plag...
