2006/05/12

[Academic] Cryptic proofs

img_1134
I am getting in heated arguments with one of the students I am working with, who refused to rewrite his proof because he claims that cryptic proofs improve the chance of a paper to be accepted. This student is brilliant, but sometimes gets a bit cocky and refuses to listen to reason from his supervisors, less even from me. ;) Lately, after I managed to show him that one of his cryptic proof was hiding better results from himself by writing my own proof of his result, he agreed to rewrite some of his proofs and found several mistakes, but he still claims that "reviewers should not understand the spirit of the proof, they should just agree that it is correct". Those arguments exhaust me and I am giving up on them, but it seems that those beliefs are at least partly well-founded: I read on Lance Fortnow's weblog how, in some communities, papers are judged on the technicality of the proofs rather than on the beauty and novelty of its concepts. I am satisfied with my proofs only when I can explain why you should prove this result this way and not another way, when I can reasonably argue that minor changes to the proof won't improve the result, and basically when I can explain the main idea of the proof on a few slides (being pedagogical): I see this as pruning the tree rather than just finding a branch leading to a good leaf. I am always annoyed at papers which describe just a path, without explaining the WHY of the proof as well as the HOW, but it seems as a good way to look wise as a Dragon, through cryptic quotes...
Post a Comment