Comment by jmatt | Hacker News Reader

Comment by jmatt | Hacker News Reader

Parent Full thread

All these academics that lay low then pop up with problems and call Deolalikar's contribution unfixable infuriate me.

They bring out the worst in the same way that bad QA employees do. They create no wealth and add little value. They slow down and confuse the process. But in the end they are still necessary. There is something to be said about how one reports such problems. Most here know a great QA person that reports relevant problems in a intelligent manner and is there to help find solutions. Versus the QA person that just cranks out an endless stream of wonky insane problems. All make you look bad and all are difficult to reproduce, understand or address.

It's my understanding that his approach is a radical and new way to attempt to solve the P!=NP problem. So it's no surprise that at first glance the vast majority of average math and computer scientists would think it was not a valid approach. And, again no surprise to me, that has ended up to be the case. I don't pretend to understand this flaw or Deolalikar's paper. But I do understand the type of person that posts comments like that...

It looks like the math community has settle on a wiki to unofficially collect news and information on the paper:

http://michaelnielsen.org/polymath1/index.php?title=Deolalik...

The claim is that the proof is unfixable. Proofs are either right or wrong. Good academics are supposed to checks proofs of their colleagues. There is no claim as to the merit of contributions made by Deolalikar. So, why exactly do these academics infuriate you?

So, why exactly do these academics infuriate you?

I'm quite clear if you read more than one sentence.

It's the way this guy responded to a comment of a blog about the paper. It just seems like a bad place to post a real unfixable flaw to the paper. Assuming the commenter is right (also possibly a bad assumption) why not put some effort into your response that will be read by every computer science and math person you'll ever work with in the future. For instance how does one respond, follow up, contact this guy? I have to go google him wtf. It's the equivalent to nitpicking. In the end someone else will have to pick up the pieces of what he said and analyze it and rewrite it and repost it. So to me he seems like the guy from QA that no one wants to work with - a bad team player.

I'm sure that QA guy loves working with you, too...

Most people's work, when submitted, gets reviewed by four people, and if published, read by a few people their sub-area of their area.

All of Computer Science is paying attention to this paper, and it wasn't even published anywhere. People popping up out of nowhere is a good thing. It's even a compliment to the work.

Nobody is saying it's a worthless approach. On the contrary-- the reason there's so much hubbub over the paper is that the approach is quite novel. Even if the end result turns out utterly wrong, much of the lead-up in getting there can be salvaged and used for entirely new directions in TCS.

In TCS, new strategies/approaches/techniques are often just as valuable as the results (and in top conferences, they are sometimes considered even more important).

We need to subject every paper to serious criticism in order to establish which morsels have merit. Hand-waving and weak logic seriously impediment our progress, so many researchers get very frustrated by it.

here is the deal, it is extremely difficult to prove P ?= NP. However it is comparatively easier to have a wrong proof.

The only thing that you can say is that this approach was one of the trickier ones to do it in a wrong way.

Yet it does not "necessarily" means that this approach is actually a starting point.

Just it seems to be more complex than the other approaches.

> The only thing that you can say is that this approach was one of the trickier ones to do it in a wrong way.

Or, rather, that someone (well qualified) thinks it was. I am very sceptical of a refutation based on the assumption that the proof founders on a beginner's mistake; it's not impossible, but, if Deolalikar is a beginner in the relevant field (I have no idea!), then it seems very likely that he would have consulted with someone who wasn't to avoid such trip-ups.

> Yet it does not "necessarily" means that this approach is actually a starting point.

I am a mathematician, not a computer scientist; but I think it's safe to say that any approach to a well known problem that is not (1) immediately dismissible (by experts) as crackpottery or (2) a laborious advance an inch down a familiar road of reasoning is likely to represent a promising starting point for something (future research in general, if not the answer to the specific question). If it is (3) described by experts as a novel and/or unexpected approach, as I think that this paper was, then that likelihood becomes a near-certainty.

Reply on news.ycombinator.com