In some comment here on meta (I can't find it anymore...), a user made reference to the fact that Math.SE has a one of (if not the) lowest question-closure rate of any Area 51 site.

I was thinking about this, and have come to the conclusion that this is probably because of the difficulty of closing any question "as duplicate" on this site. Computers just aren't that great at searching math formulas yet. :) It crossed my mind that many questions asked here are canonical homework questions, and chances are the same question (from the same book) has been asked before.

I was trying to think of an easily-searchable way to index these questions. My proposal is that we should request users who are asking homework questions to include:

  1. The title of the book (with edition number or publication year, preferably)
  2. The author of the book
  3. The page and problem number

However, it is true that some homework questions are from handouts in a class, and thus don't have the above. So, I suggest askers include in this case:

  1. The college/university they are at
  2. The Professor's name
  3. Course name/id
  4. (If there's a homework sheet number, include that.)

This has two major benefits:

  1. It cuts down on duplicate questions. Now, we can simply search "Problem 35, Boyce DiPrima, page 205", and find all previously asked questions. This is the major benefit that I am seeking with this suggestion.
  2. It cuts down on cheaters. This may or may not be such a big result (cheaters will find a way to cheat, after all).

Why am I mentioning it on meta? In order for this to have any visible effect, it must be a group effort of many people. If I go around saying "please give this info for your homework problem," no one will really pay attention, or they may just be annoyed. However, if the community is asking for the information, it becomes a norm, rather than the exception, and is more easily followed.


  • 4
    $\begingroup$ We can't even get users to tag homework questions with homework, so why do you think requesting that they be labeled with the source will accomplish anything? $\endgroup$
    – MJD
    Mar 10 '13 at 19:01
  • $\begingroup$ @MJD I highly doubt that 100% of homework-question-askers will label their questions with the source. However, I do think that some will if asked. This would enable at least some form of searchable text to be in some questions, so it is possible to find duplicates. $\endgroup$
    – apnorton
    Mar 10 '13 at 19:16
  • $\begingroup$ Not for the same reasons, but there was a similar proposal: Categorizing homework questions with textbook information in a way that makes the site act like a solutions manual $\endgroup$ Mar 10 '13 at 19:33
  • $\begingroup$ @MartinSleziak Thanks for posting that. I thought that this may have been discussed before, but didn't see that post. $\endgroup$
    – apnorton
    Mar 10 '13 at 19:38
  • $\begingroup$ So, it's not just math questions for which it's hard to search for duplicates.... $\endgroup$ Mar 10 '13 at 23:42
  • $\begingroup$ @GerryMyerson Very funny... :\ $\endgroup$
    – apnorton
    Mar 11 '13 at 2:18
  • $\begingroup$ People can always claim that they are self-studying and refuse to tag their questions as homework. I think what's basically wrong with this homework tag is that (a) people simply has no incentives to tag a question as homework, and (b) worse still, they may feel discriminated if they tag their questions as homework. Maybe one way to improve the situation is to allow users to give two upvotes to a homework question if they think the question is well written and show that the OP has spent enough efforts to solve it. $\endgroup$
    – user1551
    Mar 12 '13 at 13:06

You must log in to answer this question.

Browse other questions tagged .