Comments on: Selecting another polymath project

By: John Sidles

John Sidles — Tue, 28 Jul 2009 18:49:03 +0000

To extend a comment on the deterministic prime-finding thread, it would be good to distinguish between NP-Poly Projects (whose objectives are in NP, but whose success in uncertain) as contrasted with P-Poly Projects (which are pretty guaranteed to succeed, with sufficient resources, diligence, and skill).

It is easy to think of historical examples of P-Poly Projects … they include some of humanity’s most lustrous scientific enterprises … the Genome Project … the Apollo Project … the Manhattan Project … the Large Hadron Collider … and hopefully even large P-Poly Projects will be launched this century.

It seems to me that it is the P-Poly Projects that have the most scope for creating jobs and careers on the scale that our planet (not to mention our respective academic disciplines!) urgently requires.

For me, the very best posts on the blogs of Nielsen, Tao, Kalai, Gowers (and Fortnow/GASARCH too) are those that touch upon this theme.

By: Lior

Lior — Tue, 28 Jul 2009 17:49:49 +0000

I think the “polymath ethic” needs to be adapted for selecting problems. Basically, selecting comments requires both wide expertise, considerable off-line thought (such as done by Tim or Gil with their proposals), as well as consultation with outsiders on behalf of the project. It is a “global optimization” issue, unlike the solution which involves many “local” steps. Thus my guess is thus that it would be better for some kind of “committee of experts” [self-selected by participating in a comment thread] to do this, probably involving the “guardian council” of Gil, Terry, and Tim.

By: Terence Tao

Terence Tao — Tue, 28 Jul 2009 16:36:18 +0000

Dear ramanujantao,

It may be best to post it on a separate blog (as per the second paragraph of this post). You can see examples of existing polymath proposals at

http://en.wordpress.com/tag/polymath-proposals/

If you are using a wordpress blog and add the tag or category “polymath proposals” to your post, then it will be automatically added to that page.

By: gowers

gowers — Tue, 28 Jul 2009 16:25:28 +0000

Another thought about the selection process is that it might be better to select not one but four or five projects that we definitely intend to go ahead with. I don’t mean by that that we would start them all at the same time — far from it — but simply that we wouldn’t have a new vote every time we felt that a new project could be supported. That would have the advantage that preliminary work, such as setting up useful resources on the wiki, could be done more thoroughly before the later projects started.

By: ramanujantao

ramanujantao — Tue, 28 Jul 2009 16:16:52 +0000

For example, I have a proposal involving a quantum entanglement problem. Should I post it here in the comments?

Thanks

By: ramanujantao

ramanujantao — Tue, 28 Jul 2009 16:13:37 +0000

Dear Terry,

Would I be able to propose a problem for consideration of polymath? Or can only the three moderators of this blog do that? Where should I post the problem for consideration?

Thanks

By: John

John — Tue, 28 Jul 2009 16:08:30 +0000

Dear Terry,

The thing is all of these projects depend on internet. What happens if there is a power outage and the internet goes down for days?

By: Noam

Noam — Tue, 28 Jul 2009 15:30:46 +0000

It is not clear to me that the possible audiences of different “polymath” projects are the same. One could think of concurrent ongoing projects aimed at (nearly) disjoint sets of participants. (E.g. a polymath project in Algorithmic Game Theory will probably have a pretty low overlap with any of the projects proposed here)

By: Terence Tao

Terence Tao — Tue, 28 Jul 2009 15:23:22 +0000

I agree that more of a push to widen the accessibility of the DHJ project would be a good idea. It may be difficult to hold back the research side of things, though. For instance, for the finding primes project, I’m now scrambling to put into place the other aspects of the polymath paradigm even though we haven’t officially launched it yet, due to the non-trivial amount of progress made so far; ideally we should have some wiki notes on pseudorandom number generators, Cramer’s conjecture, etc. before we start but I guess this will have to be created on the fly (and we may have to launch this particular project ahead of schedule if the pace picks up any more).

I think one can adjust the positioning of the comment metadata by editing the CSS, but don’t know exactly how; I’ll pose this question in the formatting thread in the hope that a CSS expert comes along to read it.

By: gowers

gowers — Tue, 28 Jul 2009 15:00:49 +0000

I agree. And I have a related proposal, which is that when it is decided what the next project will be, if it is one that involves a certain amount of understanding before one is up to speed, there should be a discussion of this kind. If we were doing DHJ now, I would propose that a lot of the wiki material (things like proofs of Sperner’s theorem, Roth’s theorem and the corners theorem) should be well-developed so that people had the opportunity to familiarize themselves with the kinds of arguments that the main eventual participants knew thoroughly. And in parallel with that, there would be plenty of opportunity for people to ask questions of the main proposer (and for others to answer those questions if they want to). I think that quite a long time — a month even — would be suitable for this.

[Small technical point -- I've just replied to the wrong comment. I find having the information about a comment appearing at the bottom of a comment quite confusing. Is it possible to adjust the theme so that that stuff appears at the top?]