I think Kolmogorov complexity is spot on comparison, considering programs as proofs, theorems as types in the Curry–Howard isomorphism.