HN Explorer

How would you responsibly disclose proof that p=np?

15 points by MarkPNeyer ↗ HN
suppose you find a proof that p = np, and with it, a simple process for solving np-complete problems. given that such a process would allow you to break all kinds of cryptographic protocols, how would you responsibly make this proof public? or would you?

if you announce that you have a proof, and provide evidence that you do, you're saying something like "in one week i will provide all computer users with a skeleton key that can open any lock" - what manner of chaos would ensue?

if you go to a trusted authority, which authority do you go to?

this sort of thing keeps me up at night.

8 comments

[ 4.0 ms ] story [ 11.0 ms ] thread
chewxy
Write a paper, get it peer reviewed in a reputable math/compsci journal (preferably math).

I'm not kidding

wslh
This is the kind of thing that doesn't need peer review in a reputable journal. You can probe you point attacking NP algorithms in polynomial time, this is your proof, the reality.

Since there are security implications, giving an alert in advance is a good way to prevent side effects of your research.

tjgq
Technically, you could prove P=NP without actually discovering a polynomial-time algorithm for an NP-hard problem. Not all proofs are constructive.
icedchai
you should watch the movie "Sneakers".
charlesjshort
Don't tell anyone but very carefully selected disciples.
dsjoerg
the right answer depends on whether you suspect others have already figured it out as well. many discoveries are nearly simultaneously made by others because the components to the solution become available.

remember that powerful countries have secret teams working on this stuff.

so you should remember that your technique may already be in the hands of various good guys and bad guys.

thus, immediate disclosure may be for the best. anything else puts you in serious physical danger. because if you haven't released the secret yet, then kidnapping or killing you may keep the genie in the bottle.

kineticfocus
Unless you actually have it working... just relax... there's a new 'proof' written by an expert every day, that ends up being dis-proven. Also, be mindful of the addictiveness of trying to solve these types of problems. Even the thought of the hypothetical validation that could come with an answer seems to be a very powerful thing.