[philwelch]

Actually I think a leak is more helpful than either the author himself releasing the proof publicly or keeping everything under wraps.

Especially since the cold fusion thing[1], researchers have been loath to publicly announce revolutionary findings lest the findings fail to pan out. Researchers don't want to be known as publicity-seeking cranks, they want to be known as earnest and honest academics, which I suspect entails playing to the in-crowd and letting the system work[2]--the system being to talk to your colleagues before hand, release everything through peer-reviewed journals, and if everything passes muster, you've eliminated any risks to your reputation while achieving renown as the guy who proved P != NP.

On the other hand, having the proof publicly available to anyone who can understand it can massively parallelize the process of having it verified, or having errors found and potentially corrected. Meanwhile, the researcher's reputation is maintained, because he did the right thing, and if any errors are found he'll be judged to have acted prudently and conservatively in advance of having his findings reviewed.

[1] Fleischmann and Pons publicly claimed to have discovered a means to induce nuclear fusion at room temperature in 1989, but their results couldn't be replicated.

[2] And the system does work--this is no criticism.