In reading the endless discussions of the acceptability of postings by particular users (e.g., are six postings per day too much for one person), or by users in particular situations (e.g., should people be "allowed" to post questions only if they have not been assigned those questions as schoolwork), I find the following slogan helpful:
-- "Math.SE is not a social networking site."
That is, the focused Q & A, low-discussion framework of SE means that although people are users of the system, structurally the site is not set up to directly create relationships between users, but only between questions and answers. For example, the software does not provide a facility for users to send each other messages.
Corollaries of this principle include:
imagining the questioner as an AI is a good habit for reducing the incentive to flamewars.
whether a posting is homework (the answer recipient will be paid in grades), job-work (the recipient will be paid in money), or research work (the recipient will be paid in status), or even math.SE-work (the recipient is trolling for reputation points) doesn't matter. Certainly a questioner can try to make it more attractive to others to spend time answering, by explaining why he wants an answer, what he has done, relevant references and sources, etc. But if the asker does not provide this, it seems to me to be out of bounds to interrogate aggressively for such information or to close a question. The correct solution to the "problem" of homework or other seemingly exploitative postings is to improve the rating and filtering mechanism for questions in ways that make it irrelevant to determine whether anything is homework or not.
Though the above is less a question than an unsolicited answer, and math.SE is not a discussion forum, the beta-test version of meta.math.SE certainly is a discussion forum as it concerns questions of policy and FAQ content, so I assume it is OK to post it here.