Saturday, February 06, 2016

[nquplbrb] Many proofs

One proof is good.  Many proofs of the same thing are better, sometimes elucidating the underlying structure (often symmetry) of the thing being proved.

Prestige in mathematics might be too skewed toward being the first to publish a proof.  Create more incentives for alternate proofs.

Under what conditions is one proof monotonically better than another?

Once something is formally proved and checked, there seems to be no incentive for another formal proof, probably due to the labor involved.

Vaguely similar is the desire to have multiple implementations of something, e.g., an algorithm, in software.

No comments :