This question is similar to other questions asked on Math SE Meta, however it is distinct:

So, I've done my research but I'm here to as more specific questions unanswered in the links mentioned above.

There are quite a few books that I am going by, from authors such as I. M. Gelfand and A. P. Kiselev, that do not contain solutions, and many (especially in Kiselev's books) are not found even on the web.

They are tough problems, and I hope to put their solutions up on Math SE as reference.

I intend to go through each problem systematically, but some problem have already been put up on Math SE. It would look awfully untidy if some problems are not uniform in:

  • title and reference
  • formatting
  • style of explanation

However, a significant problem will arise if I do so.

If I were to make it systematic, some problems would be considered "duplicate", down voted, and potentially deleted.

How would I accomplish the described task, if at all possible?

  • 4
    $\begingroup$ I do not view the compilation of solutions manuals as a good use of the MSE site. If you want to do that, make your own blog or contribute somewhere like exwiki. As I remarked here, there is a very high threshold for getting a good response to self-answered questions, and attempts to do similar things to this in the past have been met with quite poor responses. $\endgroup$
    – user296602
    May 30, 2016 at 21:02
  • $\begingroup$ @T.Bongers - Yes, that's my biggest concern. However if I blog it, chances are it'll never see the light of day. :-). I suppose I'll refrain from my plan. Thanks. $\endgroup$
    – Fine Man
    May 30, 2016 at 21:38
  • 1
    $\begingroup$ I don't think "without formally published solutions" in the title makes a difference. When posting an exercise, one can give a summary of the published solutions as part of the question, and ask about additional methods, generalizations, references or other aspects that are relevant in the case where a solution is known. Also, some exercises from a book without solutions might be duplicates of exercises in other books with formally published solutions. $\endgroup$
    – zyx
    May 31, 2016 at 4:05
  • $\begingroup$ @zyx - OK, I'll correct the title. Thanks. $\endgroup$
    – Fine Man
    May 31, 2016 at 4:09
  • $\begingroup$ Exwiki has not been updated in the last 4 years. @T.Bongers $\endgroup$
    – zyx
    May 31, 2016 at 4:13

2 Answers 2


It would great to have solutions to as many book problems as possible on MSE, indexed as searchably as possible ---- but self-answered questions by learners are not the best way to upload that content. People solving exercises as a way of learning the material will (on average) produce lower-quality solutions than people with more knowledge.

For learners or posters who consider themselves non-expert on the subject of the exercise, I suggest:

  • post the problem with a sketch of a solution (if you know one) in the question instead of a self-answer. Give the essential idea, but light enough on details to not make the question too long or hard to read
  • do ask about alternative approaches, better solutions, generalizations, applications, references, or connections to theory
  • do not ask as a [proof-verification] question. Verification tags tend to prevent the posting of alternative methods since answerers will focus on your solution rather than the problem it attempts to solve.
  • give precise source information such as the book, edition, chapter, page and exercise number. This will help index it for web searches and the MSE site search.
  • quote the exercise exactly (repairing or commenting on any obvious mistakes).

This is also a good routine for more experienced posters, except that the sketch of solution could be replaced by a self-answer if they think there is sufficient reason to do that.

If a long time passes with no answer better than the one sketched, that could also be a reason to post a self-answer.


One possible way would be to work like this:

  • Start writing list of posts corresponding to the exercises in some editor of your choice.
  • For each exercise either find already existing post on this problem or create a new question if needed.
  • You can share the link to this list in your profile (or elsewhere).

For example, you could create a text containing name of chapters and numbers of exercises. If you prefer, you can also type texts of the exercises. To each exercise you could add links to posts on this site where the exercise is solved. (Of course, you could add some other resources where the problems are solved. It is not necessary to restrict yourself to this site.)

The choice of editor depends on your preference. But if you want to use MarkDown+MathJax (i.e., similar format as you would use in the posts here), one possibility would be StackEdit. You can find some other suggestions if you have a look at some of the posts here on meta tagged .

Once you have created (some version) of the text and published it somewhere online, it is good idea to promote it in some way so that other users (and search engines) become aware of it. One good possibility is to put the link into your profile. You can also mention the text in chat. (Especially if there is a chatroom devoted specially to the subject of this particular book.)

Additional note: Compiling a list with exercises is a lot of work. (In addition to finding links, posting questions, etc.) But if you decide not to make such a list, you can at least post a comment or, if it is your question, mention in the body something like: "This is Exercise 1.2.3 from the book B by the author A." This might help other people having problem with this exercise to find it using Google. For example, by searching for something like Rudin "chapter 3" "exercise 4" or Rudin "exercise 3.4".

I think that in general mentioning source of the problem in the question is a good thing to do.


You must log in to answer this question.

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