Live data from Hacker News

New proof reveals that graphs with no pentagons are fundamentally different

quantamagazine.org

111–116 of 116 posts

Re: New proof reveals that graphs with no pentagons are fundamentally different

#111

Question for people knowledgable about Ramsey theory - with current computing tech, will we ever be able to find the exact value of R(5,5) within our lifetimes? I've read the famous Erdos quote about R(5,5) vs R(6,6) and he seemed to think that it was at least theoretically possible if the whole human race had to find it quickly or face destruction.

I did some work in Ramsey theory 20 years ago (improved the upper boundaries for R(3,12) and R(3,15) by one digit each [1]) and I agree with Erdös: we could do it, if there was a compelling reason to devote a lot of effort to it.

As others have said, brute force is just plain impossible. But we can limit the search space significantly by proving sub-results about the structures of the possible graphs. Somewhat trivial example to give the flavor: let's state the problem as "find the smallest number of vertices needed such that any graph must have either a 5-connected component (a "pentagram") or 5 independent vertices". We know that the number is at least 43. So if we are looking for a counterexample, a graph on 43 vertices that does not have either of these two subgraphs, what can we say about the possible number of edges that each vertex must have? (the degree, for graph theorists). We can immediately say that if we have 5 or more vertices with no edges (degree 0) then we have our independent set already, so any counterexample can have at most 4 vertices of degree 0.

By the way, my favorite Explain-like-I'm-5 version of Ramsey's theorem is "Total disorder is impossible": If a structure is large enough, there must be some substructure that is ordered.

[1] https://www2.math.su.se/ramseytheory/sciramseyams.pdf

Re: New proof reveals that graphs with no pentagons are fundamentally different

#112

Earlier quoted context omitted.

As a person who briefly pursued a degree in a traditional engineering discipline before finding my way back to the arts (music) and then eventually into a software career, I think the difference - at least for me - is that with the arts I've never gotten the impression that the fundamentals stop being interesting if they're still applied well. To phrase it another way: practicing scales can be boring. But playing a w…

As a mathematician, I would claim that math at its best is exactly about riffing and jamming on shared fundamentals and adding your own insights. But it is often misunderstood what these fundamentals are: not the proofs or the formulas, but the elements of logical reasoning that make it possible to say with absolute certainty that given some conditions, some statement must be absolutely true. A riff on a proof in geo…

Mathematics is indeed not taught in a proof-based manner, but I would also argue that it takes a lot of training to teach proof-based maths to primary grade students. It is far harder than teaching maths to undergrads or high schoolers. Which is the real reason it doesn't happen in schools.

> I didn't enjoy the endless calculations of long division in grade school...

Sure, everyone has their own cup of coffee. But drummers, for instance, practice the same beats over and over again for dozens of hours to perfect them. Not very different from doing long division over and over again.

Re: New proof reveals that graphs with no pentagons are fundamentally different

#113
I bet that Maria Chudnovsky in the article is related to the Chudnovsky brothers (Gregory and David) who were the inspiration of the 1998 mathematical/psychological movie Pi by Darren Aronofsky.

See also: https://www.newyorker.com/magazine/1992/03/02/the-mountains-...

Re: New proof reveals that graphs with no pentagons are fundamentally different

#114

Earlier quoted context omitted.

I wonder who it's written for? I feel like I understand all the articles, which is worrying because how would I know if I'm right.

The usual Gell-Mann way - when they write an article about a topic you know very well, you see whether you find it "mildly annoying" (because there are always small inaccuracies or papering-overs that seem a big deal to us), or "hilariously bad", or "so outrageous the writer ought to be fired". The quality of the other articles you aren't qualified to judge can then be assumed to be in a cloud around where you judged…

I have the opposite problem here. The CS articles seem fine, but I don't know who would be interested in Quanta's articles but only capable of understanding things at their writing level. It doesn't seem like they should have an audience.

Re: New proof reveals that graphs with no pentagons are fundamentally different

#115

Question for people knowledgable about Ramsey theory - with current computing tech, will we ever be able to find the exact value of R(5,5) within our lifetimes? I've read the famous Erdos quote about R(5,5) vs R(6,6) and he seemed to think that it was at least theoretically possible if the whole human race had to find it quickly or face destruction.

That quote is: "Suppose aliens invade the earth and threaten to obliterate it in a year's time unless human beings can find the Ramsey number for red five and blue five. We could marshal the world's best minds and fastest computers, and within a year we could probably calculate the value. If the aliens demanded the Ramsey number for red six and blue six, however, we would have no choice but to launch a preemptive att…

That comment just wants me to fight aliens. What's wrong with me?

Re: New proof reveals that graphs with no pentagons are fundamentally different

#116

Earlier quoted context omitted.

The usual Gell-Mann way - when they write an article about a topic you know very well, you see whether you find it "mildly annoying" (because there are always small inaccuracies or papering-overs that seem a big deal to us), or "hilariously bad", or "so outrageous the writer ought to be fired". The quality of the other articles you aren't qualified to judge can then be assumed to be in a cloud around where you judged…

I have the opposite problem here. The CS articles seem fine, but I don't know who would be interested in Quanta's articles but only capable of understanding things at their writing level. It doesn't seem like they should have an audience.

Their audience is me, a curious not-very-technical layman who is interested in the concepts but bored by the details!

Granted, I'm not sure how common this is.

I do worry about Gell-Mann amnesia with Quanta, but I comfort myself that usually nobody has a stake in distorting this or that number theory thingamajig or physics quandary. However... I'm sure there are academic infights and research tug-of-war to which I'm blissfully blind.

Post reply on HN