I've asked three questions here on relatively advanced but esoteric math topics, yet the questions themselves aren't really research level. In each of them I'm seeking some sort of explanation for why things are the way they are; I'm far from professional mathematician status (MathOverflow's stated audience).

But all three of my questions have gotten very little traction here: I've even earnt a Tumbleweed badge for one of them, while the others had <50 views total and none in the past few days until I brought them up in chat earlier today.

I thought that perhaps there's something wrong with how I ask my questions (too lengthy?) but when I asked on Chat, it was suggested to me that perhaps there simply isn't anyone on this site interested in those particular topics.

I've thus started thinking that it would perhaps be better for my questions to be migrated to MathOverflow, where perhaps they'll be more at home. But I can neither flag nor close them myself for migration with that goal in mind (apparently they are too old to be migrated). So could the following list of all my questions be migrated over to MO?

  1. BPSW Primality Test - Selection of D & Q parameters

    A question about why a particular primality test uses a particular procedure for selecting certain internal parameters, when said strategy has obvious flaws.

  2. Computing the distance between two Linear Congruential Generator states

    A question about whether it is possible, in general, to compute the distance between two LCG states "efficiently" (sublinearly), purely from their current value, and why my procedure appears to successfully do so for my particular application.

    • I ended up self-answering this one, because I happened to encounter, in-person, a world expert on this topic.
  3. Numerical stability of Winograd short convolution algorithm

    A question about why an algorithm for fast convolution nowadays called "Winograd convolution" empirically exhibits such poor numerical stability, when 1) A comparable algorithm relying on the FFT does not and 2) Both algorithms are "correct" in the sense of giving the exactly correct answer when they are executed symbolically.


1 Answer 1


I would recommend cross posting over migration, both because then you don't have to worry about time limits and because you get more exposure that way. I have done this once (I've only ever posted 11 questions on MSE). You can see the original post and then the Overflow cross post. My cross-post was well received, with 18 upvotes and 5 stars, and asked 12 days later after it had received many upvotes but no answers. I think waiting a week or so is a good amount of time to wait (see related questions linked in the comments). I eventually received an answer on Math Overflow, and self-answered my question with a link to the Math overflow.

The MathOverflow Help Center says (emphasis mine):

What kind of questions can I ask here?

MathOverflow's primary goal is for users to ask and answer research level math questions, the sorts of questions you come across when you're writing or reading articles or graduate level books. Of course, individual questions don't have to be worthy of an article, and they don't have to be about new mathematics. A typical example is, "Can this hypothesis in that theorem be relaxed in this way?"

The site works best for well-defined questions: math questions that actually have a specific answer. You'll notice that there is the occasional question making a list of something, asking about the workings of the mathematical community, or something else which isn't really a math question. Such questions can be helpful to the community, but it is extremely tricky to ask them in a way that produces a useful response. So if you're new to the site, we suggest you stick to asking precise math questions until you learn about the quirks of the community and the strengths of the medium. If you have a very broad question (like "Please explain topic X"), try searching Google, Wikipedia, nLab, or looking for survey articles on the arXiv.

Please look around to see if your question has already been asked (and maybe even answered!). If you do post a question that was asked here before, don't worry; somebody will give you a link and close your question as duplicate.

The best way to get great answers to your question is to write a great question. To help you do that, we've written down some guidelines on How to ask a good MathOverflow question.

It's a common misconception that Math Overflow is for professional mathematicians only. It's for questions that are difficult enough to be of interest to researcher and graduate students, even if they don't arise out of research or aren't being asked by professional mathematicians. I think that all three of your linked posts meet that bar


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .