Why Math Overflow works, and why it might not

I spent a bunch of time yesterday playing with Math Overflow, the new math Q&A website launched last week by Berkeley grad students David Brown and Anton Gerashchenko. The site is built on the popular Stack Exchange platform, giving users the power not only to ask and answer questions but to vote on other people’s answers, giving those users “reputation points” which they can use to unlock more features of the site.

I was chatting with Tim Gowers last month, in the context of PolyMath, about what made a website “sticky,” or, to put it more pungently, “addictive” — what makes users willing to go back to the same site multiple times a day, and keep it up for weeks or months?  Math Overflow seems to have this quality in a particularly pure form.  Unlike PolyMath — where showing up half a day late might well give you no chance of catching up and making a contribution — Math Overflow offers a constantly changing array of new questions.  Questions to which you might know the answer right off the top of your head — or at least if you take ten minutes to think about it, or just a quick half-hour to scan through some references or…

Now at this point you might say “I could answer this, but I don’t really have the time right now.”  But then somebody else would answer it first! And then you wouldn’t get that warm feeling of helping somebody out!

I think this quality of rightnowness is what’s kind of great about Math Overflow, the thing that will get a lot of people to look at it consistently and thus make it a useful place to ask questions.  But there’s also something worrisome about it.  It shouldn’t be important to be the first one to answer.  A much more rational response to that “right now” feeling would be:  “I don’t need the warm feeling.  An earnest, hard-working grad student will come along and give the same answer I would have given; except the E,HWGS will spend more time and give a more thorough answer with more details included.”  And maybe giving a terse, dashed-off answer as soon as you see the question will actually prevent that E,HWGS from ever writing the ideal answer!

But then, a terse, dashed-off answer is a lot better than no answer.  At the moment I’m very high on this site.  I hope a lot of people — even earnest, hard-working senior faculty — will put a shoulder to it, and see what happens.

Tagged , ,

7 thoughts on “Why Math Overflow works, and why it might not

  1. Richard says:

    Oh no … yet another thing to distract me from work on my paper!

  2. Qiaochu Yuan says:

    I don’t think it’s been important to be the first one to answer. As long as the question is still on the first page people do vote up answers that appear later but are more thorough than earlier answers.

  3. […] Quomodocumque I learned of a new website, Math Overflow, where you can ask and answer mathematical questions. It […]

  4. […] I should try to ask the problem also on “math overflow“. See also here, here and here for what math […]

  5. Anonymous says:

    Terry Tao is quoted as describing MathOverflow as venerable newsgroup sci.math, but with more modern, ‘Web 2.0’ features”. I think this is a perfect description, as long as one drops the adjective “venerable”. Underlying the initial appeal of MathOverflow is a hidden promise that cannot be delivered, and sooner or later the emperor will be revealed to have no clothes. The premise of MathOverflow is that, unbeknown to us, there is a large pool of hidden mathematical talent with the historical, technical, and intuitive knowledge waiting to answer our questions. Sadly, this is not the case. Any expert using this site will soon realize that many of the best answers come from “the usual suspects”. Reference requests and questions outside one’s field may still be usefully be posted to MathOverflow, but I don’t think this will always be the case. As the “experts” gradually realize that they may not have so much to learn from the site, they will correspondingly spend less time visiting — thus decreasing further the proportion of intelligent questions and answers. Of course, there will be people to fill the void left by the experts — there are the addicts (to be fair, there is a small but non-zero intersection of this set with experts), the misplaced egos (the continual bane of mathematicians everywhere — people who don’t know what they are talking about), the enthusiasts, the amateurs, and finally, the cranks. If one looks at recent number theory questions, it certainly seems as though the enthusiasts have been in the ascendancy for quite some time. Even if the cranks are held at bay (which they are, mostly), in another six months MathOverflow will be a pale imitation of its initial self (if it hasn’t reached that point already).

    I also wanted to give some of my thoughts about polymath, but those are slightly more nuanced, so I will leave that for another time.

  6. . says:

    I more or less agree with Anonymous, though for people at somewhat isolated places this sort of thing can be quite valuable. Moreover, the patience and kindness of posts by experts such as Emerton are a nice step in the direction of helping to fight against the reputation of math as a club for the few. Buzzard’s willingness to ask a lot of reasonable questions is a good thing too. But the proportion of idiotic questions has certainly gone up too much, and the grad students at Berkeley who ask way too many questions should make more use of their good location and stop cluttering the site with weird/idle questions. (Of course, the same applies to others, but the ones at Berkeley have less of an excuse.) MO is heading in the direction which Anonymous predicts, but is not quite there yet.

  7. Dirty Davey says:

    FYI, there is now a mathematics forum at the main Stack Exchange site:


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: