Third Poll on P vs NP and related Questions is out now! And the winner is Harambe!

I took a poll of the theory community (and others) about P vs NP and related issues in 2002, 2012, and 2019 (sorry its not an arithmetic  sequence --- read the third poll article to see why). The 2002 and 2012 polls are on the web (see here and here ), and now the third poll is out and its here or here . All three appear in Lane's Complexity Column in SIGACT News.

I'll give some highlights and thought about the third poll, but for more read the article. Or read all three and see how opinions have changed over time.

0) How many respondents: 100 the first time, and 152 the second time, 124 the third time.

1) P≠NP is at about 80%. When restricted to people who have thought about the problem a lot (my judgement) then it goes to 99%. The strongest P≠NP is by Lance who says

                                People who think P=NP are like people who believe Elvis is alive.

I disagree. I think Elvis is alive and is trying to prove P=NP.

2) When will it be solved? 66% thought it would be solved before 2100, a bit more than in prior years. But also more thought it would never be solved. Some commented that a computer might solve it.  I doubt that, but I do think this poll will be done by a computer in 10 years.

3) Because I used Survey monkey (1) each question got more answers, and (2) most  questions got a forced YES or NO  so less funny comments or room for creative answers. People could still leave comments, and some did.

4) Related to point (3): The last time I did the poll P=BPP was thought by everyone who answered the question. This year far more people answered it and quite a few thought P≠BPP. This may be because Survey monkey had a psychological affect of making people vote even if they didn't really know (people who have thought a lot about P vs NP  all thought P=BPP). Has  there been evidence that P≠BPP that I am unaware of? Or since there has not been that much progress on it maybe they are unequal. 10 years ago I would have thought we would have L=RL by now.

5) The last time I did the poll 10 people thought factoring IS in P and the NSA or the KGB knows that. This time around nobody thought that. Why? Those 10 people have vanished mysteriously.

6) Five people thought P vs NP will be resolved by Harambe. That is more than any person got.

7)  Is this poll worth doing? I would say yes (gee, I have to having done his poll three times) because it is good to have an objective record of subjective opinion.

8) I'll give some of my answers: P≠NP, Elvis is dead, and both will be proven in the year 2525. For more about the future see this.


Computational Complexity published first on Computational Complexity