Should we allow questions about homework problems?

Opinions about this question on other sites of the SE network range from “no way” to “of course”. The overall consensus seems to be that the answer is “yes”—for pragmatic reasons (it’s impossible to check) if not for philosophical ones—though the guidance given to askers differs a bit from site to site (e.g., Stack Overflow, math.SE, and CrossValidated).

What should our approach be? I’ll post my opinion in an answer below so that others can comment and/or post their own answers.


I believe that we should allow, and even encourage, questions about homework, but that the asker must show that they have tried to solve the problem already, and that they must ask specific questions about what they are stuck on. Furthermore, because we are trying to create a body of OR knowledge that will be of use to many people, not just the individual person who asked the question, we should encourage people to ask questions about the models, algorithms, etc., not about the specifics of the homework problem itself.

A lot of the discussion in the threads linked to above is about how much detail to include in the answers we write in response to homework-related questions. I think that’s important, but I think a lot of the tough decisions might be avoided if the questions are good. If someone just copies and pastes the homework problem, we have to decide whether to provide a hint or a full solution or somewhere in between (assuming we want to answer the question at all). But if someone shows their work and asks a specific question about it, usually I think that this deserves an answer, probably a complete one.

Obviously we’ll still get plenty of “do my homework for me” questions, but we should send a consistent message about what kinds of homework-related questions we encourage and what kinds we don’t.

So, I suggest that we include a post like the one below in the FAQ or Meta, and when someone asks a poor question, we post a comment like “This seems like it might be a homework problem. We are happy to help, but we can best do so if you (a) show your work so far, and (b) ask a specific question about where you got stuck. For more, see [link to post]."

The post:

Questions related to homework are allowed—even encouraged—here on OR.SE. Note the important phrase “related to”: You can ask questions related to your homework. You can not post homework questions and expect them to be answered here.

When asking questions about homework, please:

  1. Make a good-faith effort to solve the problem first, and show us what you have tried so far.
  2. Ask about specific elements of the problem or your partial solution; show us where you are stuck.
  3. Couch your question in the context of the model or algorithm you are asking about, not in the context of your homework problem. Remember that the questions and answers on OR.SE should have the potential to help lots of people, not just the asker.

If you don’t do these things, your question may be voted down and/or closed.

For example, don’t post this:

Consider the following LP: $$\begin{alignat}{3} \max \quad & z = & x_1 & + & 2 x_2 & \\ \mbox{s.t.} \quad & & 3 x_1 & + & x_2 & \geq 5 \\ & & x_1 & & & \leq 6 \\ & & x_1 & + & 5x_2 & \leq 10 \\ & & x_1 & , & x_2 & \ge 0 \end{alignat}$$ What are the corner-point solutions, and what are the constraint boundaries for each?

Or this:

I am supposed to find the corner-point solutions for the following problem: $$\begin{alignat}{3} \max \quad & z = & x_1 & + & 2 x_2 & \\ \mbox{s.t.} \quad & & 3 x_1 & + & x_2 & \geq 5 \\ & & x_1 & & & \leq 6 \\ & & x_1 & + & 5x_2 & \leq 10 \\ & & x_1 & , & x_2 & \ge 0 \end{alignat}$$ Do I need to include the infeasible solutions too, or just the feasible ones?

But it’s fine to post this:

I am working on a homework assignment that asks me to find the corner-point solutions, and the constraint boundaries for each, for the following LP: $$\begin{alignat}{3} \max \quad & z = & x_1 & + & 2 x_2 & \\ \mbox{s.t.} \quad & & 3 x_1 & + & x_2 & \geq 5 \\ & & x_1 & & & \leq 6 \\ & & x_1 & + & 5x_2 & \leq 10 \\ & & x_1 & , & x_2 & \ge 0 \end{alignat}$$ I can find the corner-point solutions, but how does one find the constraint boundaries for solutions such as $(6,-13)$ that are feasible with respect to the functional constraints but not the non-negativity constraints?

It is also important that you follow the policy of your school and instructor regarding whether it is acceptable to seek outside help with your homework assignments.

Finally, don’t be surprised if the answers you receive guide you toward a solution without actually providing all the details. Our goal is to help, and giving you complete solutions to homework problems won’t help you, in the long run.

  • $\begingroup$ In point 4, it really is not "most important" to be following policies that Stack Exchange has no particular clue about nor interest in dealing with. If anything, it is most important to do what is mentioned in the first point: show your prior effort. A question that happens to be against some unknown policy may be totally fine otherwise, and still a good question, while a question with no effort made is still junk regardless of whether the homework is totally open-book. $\endgroup$
    – Nij
    Jun 11 '19 at 19:51
  • $\begingroup$ @Nij OK, that's a fair point. It was the professor in me that put that. :) Do you suggest I eliminate that point, or "demote" it somehow? (I prefer not to eliminate it, because I do think it's important to send the message that we support whatever policy/limitations your professor imposes.) $\endgroup$
    – LarrySnyder610 Mod
    Jun 11 '19 at 19:56
  • $\begingroup$ I think merely mentioning that students should keep their institutional policies in mind before posting would be helpful. Perhaps a reminder that once you post, it becomes SE's question for everybody, though point 3 somewhat does this already. $\endgroup$
    – Nij
    Jun 11 '19 at 20:03
  • $\begingroup$ @Nij I edited the post to change the position and language of the statement about institutional policies. I didn't mention the part about the post becoming SE's question for everybody, but feel free to suggest a further edit. $\endgroup$
    – LarrySnyder610 Mod
    Jun 12 '19 at 0:12
  • $\begingroup$ Can we clairify: "... if someone shows their work and asks a specific question about it, usually I think that this deserves an answer, probably a complete one."? --- Ask about something complicated, show effort, and ask for help with one specific part. IE: Not "too broad" (do my homework) and the expected answer should be of a reasonable length (people can write lengthy answers, but it shouldn't be a requirement of a 'homework question'). $\endgroup$
    – Rob
    Jun 12 '19 at 0:34
  • $\begingroup$ @Rob I agree, but are you suggesting a change to the post (below the divider line)? I'm not proposing to post the part above the line as the "official" guidance for HW-askers. $\endgroup$
    – LarrySnyder610 Mod
    Jun 12 '19 at 1:05
  • $\begingroup$ I'll be able to discuss further when I have more time available, but I'll leave a few links (in chat I suggested that you do your homework and see what the community consensus is). See: "Improving “demonstrate a minimal understanding” close reason", "Why the false dichotomy in discussing how to respond to suspected “XY problem” questions? (and comments), ... --- just ran out of time. I'll be back - Arnold $\endgroup$
    – Rob
    Jun 12 '19 at 1:38
  • $\begingroup$ @Rob That's fine, I was just asking for clarification on your suggestion. $\endgroup$
    – LarrySnyder610 Mod
    Jun 12 '19 at 1:39
  • $\begingroup$ Math.meta.SE has a How to ask a homework question FAQ-styled Q&A. $\endgroup$
    – Rob
    Jul 30 '19 at 13:51
  • $\begingroup$ @Rob yes, I think I was aiming for something similar. $\endgroup$
    – LarrySnyder610 Mod
    Jul 30 '19 at 15:50
  • 1
    $\begingroup$ I put that here since your other Meta Q links to this, so this seemed like the slightly better place. Ultimately we will have our own policy and standards/expectations which can differ from other sites. It's important to have some idea early as we wouldn't want to later decide we need to dial it back; leaving legacy questions that are no longer wanted. $\endgroup$
    – Rob
    Jul 30 '19 at 16:11

You must log in to answer this question.

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