The questions about recommendation of books/texts/lecture notes for some topic are to some extent subjective and they will often be left without accepting an answer (since among several good recommendations it is difficult to single out one of them as "the correct one").

So it seems that this argument from Qiaochu's answer applies here:

The question is asking for a list of examples. In this case I think it makes sense to use the voting system as a way to rank the answers, and it's easier to do this if the question is CW because then votes do not affect reputation, so they are more honestly judgments of the answer than of the answerer.

On the other hand this point from an answer to a question at meta.SO seems reasonable to me, too.

Even if there is more than one valid answer (in open-ended questions), individual answers may still have value, and so they deserve the rep gain when they're upvoted. Marking such a question CW just discourages people from putting any effort into their answers.

I'd be glad to find out opinion of members of MSE community on this.


2 Answers 2


I was about to ask this exact question, but this question came up as having been asked before. I feel that it makes no sense that someone should be able to gain large sways of reputation for an answer that

  1. Requires minimal effort (in general)
  2. Is purely an opinion
  3. May not even indicate any knowledge whatsoever - I may have heard that X is a great book on Y and recommend X without knowing anything about Y

On the other hand, I feel that book recommendation questions are an important part of math.SE. To me, the simplest solution is as you've suggested - to allow book recommendation questions but disallow reputation gains on them by making them CW

  • $\begingroup$ If an answer is purely an opinion and does not indicate any knowledge, what good is it? It should be deleted. If the question invites such answers (by asking "hey, what are some books y'all like?"), the question should be closed and deleted, not just made CW. If it's an opinion poll, it's not a valid Stack Exchange question. Book recommendations may be valid SE questions, if they give objective criteria on which answers can be evaluated. $\endgroup$
    – user147263
    Aug 3, 2014 at 15:15
  • $\begingroup$ Right, I definitely agree with that. Nonetheless, reasons 1 and 3 apply to all book recommendation questions. $\endgroup$
    – Mathmo123
    Aug 3, 2014 at 15:22
  • 3
    $\begingroup$ Explaining the contents of a book, and its virtues (and faults), can take significantly more effort than answering many of the other questions here. If you don't think a recommendation deserves rewarding, don't vote it up. $\endgroup$
    – user160609
    Aug 3, 2014 at 16:08
  • 3
    $\begingroup$ It can. More often than not it does not. However, in general, people will vote up a recommendation because they happen to like that book. See here: math.stackexchange.com/questions/886317/… and here: math.stackexchange.com/questions/804716/… for examples $\endgroup$
    – Mathmo123
    Aug 3, 2014 at 16:14
  • $\begingroup$ If you value superb book reviews, then it is logical to let them be motivated by the same reward system that SE uses to generally motivate contributions. Forcing them to be CW does not allow that. $\endgroup$ Aug 3, 2014 at 19:15
  • 1
    $\begingroup$ @BillDubuque that's certainly true... however at the moment we are rewarding awful book reviews. Users will still earn badges for their reviews so there is still some reward $\endgroup$
    – Mathmo123
    Aug 3, 2014 at 19:20
  • $\begingroup$ I don't follow your logic. You think that too many users are rewarding poor reviews, so you think the fix is to remove the possibility to reward good reviews? How does that make sense? I also think that, more generally, too many users are rewarding poor FGITW answers too. By your logic we should make all answers CW! $\endgroup$ Aug 3, 2014 at 19:26
  • 1
    $\begingroup$ My logic is that the nature of the question encourages a very simplistic, one line answer of "I suggest this book because...". This answer requires minimal effort and no expertise. There will also be some fantastic answers, but they are a minority and will remain one so long as one can be adequately rewarded for poor answers. Other questions encourage thought out answers, where even if they are simplistic, they at least require some level of expertise. I can provide a book recommendation without ever having read a book. $\endgroup$
    – Mathmo123
    Aug 3, 2014 at 19:29
  • $\begingroup$ So too for "is my proof correct?" questions. But those questions are not hammered with CW. Some users do answer with nothing more than "yes". But others give more illuminating answers (and they probably would not if such questions were CW hammered). $\endgroup$ Aug 3, 2014 at 19:32
  • $\begingroup$ I agree... although in that case, users who just answer "yes" don't get up voted. In fact they often get down voted. In the book review case, that is not the case - people up vote because they agree with the book choice, not because they think effort/expertise has gone into the answer. $\endgroup$
    – Mathmo123
    Aug 3, 2014 at 19:33
  • $\begingroup$ Then maybe you should post a meta question about what the proper criteria should be for voting on book reviews, and then reference that in comments when need be. Probably some users are not even aware that voting in such ways may annoy others. $\endgroup$ Aug 3, 2014 at 19:39
  • $\begingroup$ I'd love to see richer book reviews here, so I am hesitant to support a proposal that discourages that (esp. if only for the reason that some poor answers got too many upvotes - a problem that is ubiquitous here). $\endgroup$ Aug 3, 2014 at 19:46
  • $\begingroup$ I think that's fair enough. I will consider writing a meta post. A posted this for the sake of discussion, and you present a strong case. $\endgroup$
    – Mathmo123
    Aug 3, 2014 at 19:57

My opinion: it should be CW, unless you expect there is only one correct answer. For example: "What was the first calculus textbook to develop the trig functions starting from their differential equations?"


You must log in to answer this question.

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