tea.mathoverflow.net - Discussion Feed ("Best of Math Overflow")

François G. Dorais comments on ""Best of Math Overflow"" (3991)

François G. Dorais — Mon, 22 Mar 2010 13:45:09 -0700

I'm neither for nor against, but I find your point "the[re] is no way I know of to search for highest-voted answers" a good motivation to improve MathOverflow's search capabilities. Such improvements would greatly reduce the need for "best of" lists.

grp comments on ""Best of Math Overflow"" (3985)

grp — Mon, 22 Mar 2010 11:45:32 -0700

Ilya: There is nothing stopping you from creating a list of links to such questions on a webpage outside of MathOverflow. You could use this as a reference for yourself or whenever you feel it useful to include in a comment or answer to someone else. If it works well enough for you as an individual, you will maintain it. If it works well for the community, they will (ideally) take it over and give you some credit for the initial list.

Gerhard "Ask Me About System Design" Paseman, 2010.03.22

Ilya Nikokoshev comments on ""Best of Math Overflow"" (1734)

Ilya Nikokoshev — Tue, 12 Jan 2010 12:35:40 -0800

I agree we don't want users to think they necessarily have to spend a lot of time writing lengthy, 5+ paragraph questions. But we know some of them do it anyway — so why not give them examples they can follow once they decide to ask a complicated, non-standard question or produce a comprehensive answer?

David Speyer comments on ""Best of Math Overflow"" (1712)

David Speyer — Mon, 11 Jan 2010 11:24:44 -0800

This list sounds like fun, but I don't think it would be useful in the FAQ, because I don't think the best questions/answers would be very comparable to typical questions/answers. A lot of our interactions are of the form: User A poses a standard question outside his field of expertise; User B posts a link to a standard reference. Examples: 1, 2, 3.

These are all good questions that we would like newcomers to imitate, but they would make for a very boring "best of" list.

Ilya Nikokoshev comments on ""Best of Math Overflow"" (1711)

Ilya Nikokoshev — Mon, 11 Jan 2010 10:37:38 -0800

There was some talk about making a list of questions to use for people as an example. I think it's an excellent idea (if Anton doesn't have enough time to do that, I'm sure others would be happy to put effort into it).

Now, however, I also think we could do a combined Best of Math Overflow list, that would also contain answers. I see two arguments against such a list:

It would arbitrarily promote some good questions and answers over other good questions and answers.
It is unnecessary as votes/favorites/seen count already provides a similar list.

However, I see more arguments for it:

We do want people to pick up good ideas on writing questions, and people learn best from examples.
We can put a disclaimer: "this is in many ways a quite random selection" on the list.
The is no way I know of to search for highest-voted answers.
Soft-questions are higher represented among highest-voted and highest-seen questions than socially optimal.
We can include explanations highlighting what exactly we consider worthy of emulation.
It will increase motivation to provide not only mathematically meaningful, but also highly readable answers.

I had this idea after reading David Speyer's answer to Peter Lee's question What is Grothendieck Riemann Roch for flag variety of Lie algebra; as you might have guessed, if there was such a list, I would nominate David's text to it.

What do you think? I understand that one (e.g. Anton) could be worried about people spending too much energy of "complementary" things, and one could make a case against all "best of" lists, but I think this particular project would be beneficial for the community.