I think that generally speaking asking for papers might be a non-research issue. If it's merely a google issue, then it shouldn't be in MO.

What about Survey/Review papers? Obviously those are hard to find, and moreover it is hard to find a good comprehensive one. What do you think?


3 Answers 3


Asking for survey papers might be a valuable question.

That said, such a question should have some context:

  • Some indication of the background at which you are starting. Survey papers can be for beginners, or for experts, or somewhere in between, and it's important to know what kind you are searching for.
  • Some context. Why do you want a survey? Is there a particular take on the subject that you are particularly interested in, and that other well-known surveys (if you happen to know of some) don't address?
  • Or perhaps a parallel from subject X, asking if there might be something similar for subject Y. For example, you might be particularly interested in the algorithmic point of view for a particular problem, and you are familiar with a survey that focuses on such a point of view for a different problem. Mentioning the latter as an example of what you are looking for will help get answers that are useful and relevant.
  • Indicate what resources you already have on hand (or are easy to find using standard resources, such as google, etc).

I think asking for survey papers is a good idea. Surveys are broader and useful to a wider audience than other papers. Therefore I think suggestions for survey articles would be useful to future readers as well, not just the OP, so they fit the site better. This, of course, assuming that you cannot find a good survey paper by simply googling for one.


MathOverflow is (among other things) intended to be a repository/clearinghouse for certain information not easily found elsewhere. If the area to be surveyed is narrow enough, it may be harder to find such a paper. However, many people have given their opinions online on good resources for various areas, and those online sources should be probed first before attempting such a question here.

I reviewed my answer to "Is this a known/interesting result?"-type questions . I would tweak it to indicate what searches you have done, and why they came up short. Even though the flow there talks about a specific result, I recommend trying to adopt the flow toward your question. When you come to what it is you really want and what you have tried, write that down, then stop. Then wait. Then look at the question and see what web search you could try to answer it.

After a few iterations of the above, you will either have an answer, or you will have a short but reasonable length (greater than 3) list of attempts and reasons why they failed. That should refine your question to exactly the kind of survey you want as well as the kind of survey you will accept. If you ask a question that is looking for a specific enough survey AND contribute at the same time information which might be helpful to others but won't help you (because of your clearly explained and specific requirements), you will have done a service to the MathOverflow community. You may even be rewarded with a link to the survey you want.

Bonus: if you offer to review a survey that is being drafted (or otherwise assist with such a draft), you may help a current author complete the survey you want to read.

Gerhard "It's About Helping The Community" Paseman, 2015.05.11


You must log in to answer this question.

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