My point of view diverges from Matt Pressland's around "to improve the education of the questioner - I see this as much more important". Color me skeptical, but I simply do not believe that Math.SE can function well as a place where mathematics is taught. It is not designed for this purpose, because Stack Exchange, Inc is not in the education business. Online academies do exist, but Stack Exchange is not one. It's good at what it's meant to do: provide answers to questions.

@Matt A suggestion towards a more friendly approach was given by Jayesh Badwaik yesterday, see here. The way I think the current "solution" is to be put into practice is quite nicely worded by Jyrki Lahtonen's new answer to the proposal.

The SE format is certainly limited as a teaching aid, although it can be very good at this. I would prefer that it at least avoids harming the education of questioners, and there are certain things that we can do to try and prevent it from doing so. Just because we can't do this perfectly doesn't mean we shouldn't try at all.

I have a conflict of interest here to a degree because I work in higher education - there's a limit to how much I can be involved with the site if it makes it harder for educators to do their jobs.

@Lord_Farin I think something like that suggestion would work well in the meta answer that's supposed to replace the "how to ask homework questions" question in the comment template. I really like Jyrki's answer to that question - flexibility is certainly important, and I think some of the problems in the discussion so far have come from trying to create a policy that's too rigid.

It seems from the meta discussion that StackOverflow also discourages no effort questions (although I'm not an SO user so I don't know how/if this works in practice) so it seems not to be the intention of StackExchange that the site should literally provide answers to all questions regardless of how they are asked.

I'm trying to take more of a "teach a man to fish" approach here - when someone gets stuck on their assignment problems, helping them learn to solve problems is more useful in the long run than just helping them solve that problem. I'm not proposing that the question remain unanswered forever (providing the OP improves their question enough), so the "goal of providing answers to questions" is still met.

@MattPressland Indeed, some forces (both approving and disapproving of QA) seem to think that everything will be set in stone by a "policy". I can of course only speak for myself, but that was never my intent, and would indeed be harmful. I was thinking more along the lines of "When we see a question needs work, let's have a more-or-less uniform way of expressing this, so as to put less burden on answerers."

@Lord_Farin I think we do need to have a policy, but it should be vague enough to allow for some interpretation on the part of users. This is a little dangerous I guess, and could provoke meta arguments, but is likely to be the practical result anyway because users on the main site can ultimately do what they like.

As an example of how I would like things to work - I was very happy with the development of this question: math.stackexchange.com/questions/377676/how-to-prove-this-claim.

A lead in to the standard comment template (partly inspired by Brian!) seemed to be a good idea. Particularly encouraging is that the question never got further than a single close vote before being improved, and I think the extra work from the questioner made it much easier to see which part of the problem they weren't understanding.

(Speaking of which, it seems to be impossible to cancel close votes - is this really the case?)

@MattPressland (I have been encountering this too. It seems strange; perhaps we should create a feature request?)

I should also clarify that when I said "assignment problems" above, I'm including problems being solved for self-study, as they serve essentially the same pedagogical role.

@Lord_Farin (I thought there was one on meta before, but I couldn't find it. Maybe meta.SO is more likely because it would be a site wide feature.)

@MattPressland There have been some nice results with the template so far. I do however think it reasonably likely that the approach (in lieu of "policy") resulting from this discussion will provoke the need for a new comment. Some critiques on the current one are sensible and I am in growing dissatisfaction with it. It will do for now, augmented with explanation, or adapted, or whatnot; it seems not fruitful to embark on a new journey for one, before this "convention" concludes.

@Lord_Farin The problem seems to be similar to the problem discussing this issue on meta - comments just aren't long enough for nuance. Personalising the comments is definitely friendlier. I wouldn't be averse (if space allows) to modifying the template to account for Brian's objection, and mention that not all users feel that this is important, but enough that the question may be closed.

@MattPressland The relevant thread is here.

This should also serve as a reminder that the situation may change if a majority of users (in as much as we can measure that) want the site to return to a more permissive approach. Thanks for finding that thread - I guess it doesn't matter too much, I doubt that the improved questions are likely to attract the remaining close votes.

@MattPressland Shortening the link to the "Good question" thread provided enough space to include more nuance. See here.

@Lord_Farin That looks better to me.