Not signed in (Sign In)

Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

    I've been a contributor to MO for over a year now, and I want to begin by thanking everyone who in an official or unofficial capacity helps to maintain the high quality of the content of the site. I have learned things here that I don't think I could have learned anywhere else.

    In my time here I have asked and answered lots of questions, and I have received mostly positive feedback from the community for my participation. I say this with the hope that I will be able to claim some legitimacy when I raise some concerns about the community.

    My main concern is about perceived hostility by the community toward newcomers. I am a graduate student, and I have had conversations with other graduate students and even some postdocs who follow MO but who are anxious about participating actively. I think that the problem lies in the swift and sometimes harsh way in which MO community standards are enforced. I glanced through meta before initiating this discussion, and it seems that this perception is not limited to my circle of colleagues. I can certainly appreciate the need for the swiftness and even the harshness - without diligent moderation MO would probably lapse into yet another homework help forum - but I think it's worth brainstorming ways to make MO more inviting while preserving our standards. I created this discussion to do exactly that.
    Here are some ideas that I've come up with so far.

    1) Often inappropriate questions are closed with little more feedback to the OP than a link to the "How to Ask" page. This is understandable: it is unreasonable to expect the volunteer "moderators" who do the closing to write out a detailed explanation every time someone posts a question about the quadratic formula (though I have noticed that some users do seem to make an effort to offer a polite explanation, and I think it probably helps in a number of ways). My suggestion is that we improve the list of reasons that one can choose from when closing a question. As of now most inappropriate questions have to be sorted into "off topic", "not a real question", and "too localized". (What does "too localized" even mean?) Options like "too basic for this site" or "too vague" would at least help people understand why their question was closed and improve their question for next time. It occurs to me that there may be technical obstructions to implementing this idea, but it's worth throwing it out there.

    2) Improve the "How to Ask" page. As of now the page offers a wordy and rather vague set of guidelines for asking good questions - I'm not sure if it would be helpful for somebody who doesn't already know how to ask a good question. I think that shortening the explanations and providing good and bad examples in each category would help people ask better questions and understand what was wrong with questions they already asked. By now I bet it would be possible to formulate a near-comprehensive list of examples of "bad question types". It would be even better if we could tie in the first idea and link the "vote to close" options to specific points in the "How to Ask" page.

    3) Improve questions instead of closing them. For every good question X and every epsilon there is a bad question Y within epsilon of X, and I think that it can be hard to find X among the sea of Y's. I don't see many examples of people improving other users' questions, but I think in many cases the OP would rather get answers to a better question than get no answers at all.

    4) Create a mechanism for migrating questions to to math.SE. This might make it seem less like the OP asked a "bad" question and more like they asked the question in the wrong forum. I think this is an important distinction from the point of view of someone who doesn't know much about MO. However I suspect there are technical hurdles associated with this suggestion as well.

    5) Sometimes it might be easier to just answer a question than close it. I have seen examples where somebody asked a fairly easy but borderline question about Hilbert spaces or something, and a debate erupted in the comments about whether or not it was appropriate for MO. Often these debates keep the question on the front page of MO long after it would have remained if the question were just answered, and I worry that these debates drive away people who otherwise might have been good MO contributors.

    6) It might not hurt to relax our standards a little bit. I think one of the most common types of questions to be closed is an overly simple question about an advanced topic. I believe our policy is to get people to ask such questions on math.SE instead of MO, and I don't think this is a terrible policy - many such questions get good answers on math.SE. But people who ask overly easy questions about advanced topics can often also ask good questions about advanced topics, and we might be discouraging those people from participating in MO when we send them to math.SE. I guess this comes down to the question: are we willing to put up with more mediocre questions in order to make MO more inviting to our target audience? Also, more abstractly, I'm not sure I would agree that "research level" questions necessarily have to be "hard" questions. I think it is not uncommon when borrowing tools from another area to get confused by some of the fundamentals in that area. At least, I hope it happens to other people :)

    Those are my thoughts on the matter. Does anyone have other suggestions? Which of these ideas can we actually act on? I look forward to other people's thoughts.
    • CommentAuthorquid
    • CommentTimeOct 1st 2011

    Paul, thank you very much for writing on this subject in such a detailed way. In my opinion while discussion like this have come up before I personally too feel that it could be good to have a (or another) general discussion on this.

    For the moment I just make some very short technical points, but intend to write something longer later:

    ad 1. changing the reasons is technically impossible.

    ad 4. Would be automatically present if the site is migrated to SE-network, which seems not unlikely to happen in the nearer future (without that I guess it is impossible, except in the indirect way done now, close and suggest).

    • CommentAuthorWill Jagy
    • CommentTimeOct 1st 2011
    Paul, on one of your points,4, MSE migration, I now sometimes leave the comment:

    Consider posting at
    but first, please read

    as Gerry Myerson had gotten sick of the stream of crappy questions.
    Any comment may get a hostile reaction, I got tired of it. I also do not put in much effort to improve questions, so often the OP has put in no effort.

    I note you make no mention of answers by newcomers at all. I checked, you have many good questions and many good answers. For me, I posted answers from some point no later than Jan. 9, 2010. (A few days earlier actually, at some point I deleted my answers with nonpositve scores). Then I asked my first question on Jan. 21, 2010. By that time I had a strong sense of what I wanted to see in a question. Sort of a Golden Rule: ask questions of others as you would have them ask questions of you.

    Dear Paul, thank you for compiling this list. Many of these points have been brought up in the past, most have gathered community and/or moderator support. Still, it's nice to see such a compilation.

    As quid pointed out, the only obstructions to 1 and 4 are technical ones that can't be fixed at this time but we are actively seeking better solutions to these (and many other technical issues).

    For 2, we welcome suggestions for improvements to the FAQ or How to Ask pages. We have full control of these pages and so good suggestions can be incorporated very quickly. Since writing such edits takes a fair amount of time and effort, suggestions with proposed text (rather than vague ideas) are more likely to be implemented quickly.

    For 3, there is a general misunderstanding of what closing a question is supposed to do. Closing a question prevents new answers to be added, and has absolutely no other effects. (Closed questions are just as visible as open ones*, they can be voted on just like open ones, they can be edited just like open ones, and comment threads remain open after closing.) So why would you want to close a question? There are three basic reasons (and multiple variations):

    1. Off-topic questions need to be closed because such questions should not be answered here. In this case, it is better to include a link to a more appropriate resource.

    2. Big list questions need to be closed after the big list has been generated and new list items are unlikely to provide any new insights.

    3. Questions that are too vague, lack context, or otherwise need serious editing should be temporarily closed until appropriate edits are done. This is because answers appearing before the edit are very unlikely to correctly answer the final form of the question, so it is best to prevent such answers from appearing altogether.

    Closing of type 1 are permanent; closings of type 2 are permanent but usually occur several months after the question has been posed; closings of type 3 are intended to be temporary, but if the edits are never made they become permanent.

    For 5, I don't really see a problem here. There is usually enough time before closing to jot down an answer to an easy question. These answers are not deleted upon closing. There is not much need for multiple answers to easy questions, so closing these really doesn't have much side-effects. (How do debates keep questions on the front page?)

    For 6, the community standards are constantly fluctuating. There are people on both sides of the fence (even among moderators) and that's healthy for our community. The system is slightly biased toward harsher standards. Mostly because you can't vote to reopen a borderline "off-topic" question until after it has been closed (and forgotten). For technical reasons, there is no way to fix that bias right now. However, I strongly encourage people to use the reopen button for questions that they believe are not off-topic. It's probably a good idea to post a link on meta to raise awareness among similarly minded users.

    *) There is a preference option to "apply ignored-tag styles to closed questions." This option was reluctantly added a long time ago in response to a big problem that has since disappeared. The use of that option was discouraged then and it is also discouraged now.


    Re: 5, I think there's a slight misunderstanding here. Comments do not bounce the post back to the top of the front page; just answers and edits, etc.


    One problem with 1. and 3. is simply that they require more work than many people are willing to put in. As Will Jagy says, in many cases we are talking about questions that received almost no effort from the OP, and it is hard to motivate oneself to put serious effort into their improvement or into detailed feedback.

    To be honest, I don't see why a graduate student should be intimidated or stopped from participating upon seeing how swiftly questions on solving quadratic equations are dealt with. I just don't see a connection there.


    I think that the problem lies in the swift and sometimes harsh way in which MO community standards are enforced.

    That's probably a big part of it. The effort required to stem the black mud tide of shitty questions that come rolling in is at times considerable, and very often this leads to an all-too-quick assessment of MO-worthiness. A recent example concerned a diophantine equation which was immediately deemed to be homework, with votes to close, and it turned out to be a bit non-trivial.

    I wouldn't mind seeing more case studies, more concrete examples, where one could argue the judgment was overly quick and harsh, before deciding that we have a real problem here. Paul, have you or people you've talked to collected examples of these?

    I would like to see less snark from some of the regular closers, admitting at the same time that sometimes it's hard to hold one's tongue in check, say in cases when OP can get an answer in 20 seconds via google and wikipedia. (Thinking here of a recent question, what is a well-quasi-order? Frickin' google it!)


    @Todd: This is not a great example, but the question, was initially judged too simple for MO, before several weighed in to counter the initial reaction. (It still has three votes to close.)

    • CommentAuthorquid
    • CommentTimeOct 1st 2011

    Regarding 6 I tend to agree, though I do not think it is so common. In particular, in my opinion there is a subtle issue. Namely, in my opinion it makes a huge difference what the general background of the questioner is. Say, if somebody is just doing a graduate course, or is a graduate student and does independent study, on some subject and asks a reasonable question on that material that is by present standards just somewhat too easy/mainstream I would definitely be in favor of letting this question 'live'. (Of course, same for people asking outside their field of expertise.) However, if the essentially same question would be asked by sombody with significantly weaker background I would be against it. This is however hard to implement and communicate in practise; and depends on ad-hoc judgement, which makes it problematic.

    Regarding the reopening Fran├žois suggests: here I would really find it nice if those voting to reopen also (like the closers in virtually all cases) gave a reason why they think the question should be reopened. Personally, I find not doing so at least as rude as voting to close without stating a reason.

    Regarding 3. A major problem here in my opinion is that there is a certain type of user that is in my opinion the most problematic kind of all that would most profit from this. These are those users that like to do some chatting: asking about things they do at best vaguely understand, throwing together some buzzwords they picked up somewhere, and perhaps believing what they do is somehow 'advanced mathematics.' On the one hand, I find these questions simply annoying. I could live with that. Yet, on other hand, I believe that these people are actually doing themselves a diservice if they intend to eventually do some actual mathematics. And, thus I consider to support this type of behavior as harmful. An extreme form of this behavior over the last month caused quite some discussions, but in my opinion it was just an extreme form of something fairly common.

    Regarding 5. Discussions can be a problem. I know this well, since I am often enough involved in them :) But, it goes in some sense in both ways. Sometimes it also could be less harmful for everybody, including and perhaps foremost the questioner, if not somebody well-meaning would start making an argument why this question is 'good' after all. Because this essentially forces those that expressed (before) a different opinion, to really spell out in detail how 'bad' this question is.

    Regarding 2 (and cf 6). I think stating the personal motivation and background for asking the question can be really important. For some questions it is not, but it hardly can hurt, and in most cases it will or would avoid problematic situations. If I had one suggestion to make to a new user it would be to be explict regarding background and personal motivation (real name and userprofile do not suffice, people do not look at it and sometimes even looking is not sufficient). In my experience, people do not or or much less likely to say things that can be insulting if the situation is clear. I never saw that somebody S that stated right away to work in field X and asking in Y was told 'you still should really know this basic fact' or something like this. By contrast, it can and does happen that if the situation is unclear discussion arises whether S might be an undergraduate, looking for homework help or something a long these lines, which amounts to saying the same thing.

    In short, coming back to Paul's original motivation: I really doubt that a graduate student who is explict about why and with which background s/he is asking will face a hostile environment. Yet, I can understand why one might think so, which is a problem.

    To make a concrete suggestion for the 'how to ask': I would mention the point 'motivation' earlier, perhaps first or second and emphasize that this is also or mainly about personal motivation/background not only mathematical motivation (perhaps in the listing one could replace just 'motivation' by 'personal and mathmatical motivation'). And where it is written perhaps a slanted or boldface for this sentence. [I believe the default meaning for many is the mathematical one and I know of at least one example where this different understanding of the word motivation caused major discussions.]

    • CommentAuthorgrp
    • CommentTimeOct 1st 2011

    I too appreciate the attempt that Paul Siegel has made regarding how the community is treating new questions. I would suggest an alternate approach to the matter. First I have a couple of assertions on which I welcome challenge as well as verification.

    One (part of a combined) assertion is that the community is divided in many respects, especially on what questions are appropriate for MathOverflow and how inappropriate questions should be handled. I also think the community is unclear on the issue of who should be served by MathOverflow, as well as what should be served. To borrow from a recent pirate movie, the rules are really more guidelines as to what should be acceptable on this forum. To speak of standards at this point is good, for some will be needed; I think it is premature to assume this community has something as formal as standards as it is still defining itself and trying to determine things about itself and about the forum. Until a body of members steps up and says this is how things should be formally and they are expending effort to ensure that, I think MathOverflow will continue to operate informally. (I think the issue of making things formal needs to be addressed soon; the impression I have is that the powers that be prefer the informality for now while they are dealing with other current issues.)

    I also assert that the current thread lacks a clear goal. While Paul Siegel has expressed concern about how newcomers are treated and wants to address that, I think it would be more effective if he asked the postdocs and graduate students he mentioned why they are reluctant to participate in MathOverflow; that might get better results than the current discussions on how to treat handling of new questions. I suggest that we make the goal of this thread a way of encouraging more activity on this site of a kind we like; this will involve people stepping up to the plate and describing what MathOverflow means to them and what they would like to see. Paul has made a good effort in that direction, but I think using the notion of "community standards" will work against the goal.

    Along these lines, I would like to see the following: a set of templates for asking and resolving reference requests (with the intent that anyone following such a template is assured by community action that the question gets a minimum time to stay open); a community-approved set of examples of questions that were closed and why, accompanied by a "citizen patrol" composed of community members who agree to use such a list in recommending closure on future questions; a community-approved set of templates for handling certain common issues such as "read my thesis chapter" requests, career-related advice requests, poorly worded knowledge requests ("I'm in field A, studying field B, and I have a homework-like question I want to ask..."), solving someone's thesis problem.

    At this point I could (and may) write up such a series of templates; I prefer not to do so until there is some community support and agreement for such a list and what should be on it. Along with that, I encourage community discussion and consensus, as much as possible, on making the purpose, intent, and mission of MathOverflow more clear and formal, while still allowing the flexibility to grow and change as we figure out what this thing called MathOverflow should really be.

    Gerhard "Ask Me About System Design" Paseman, 2011.10.01

    • CommentAuthorquid
    • CommentTimeOct 2nd 2011

    Gerhard, I am quite surprised to see the following as a bad example: "I'm in field A, studying field B, and I have a homework-like question I want to ask..."

    In my opinion this is a really good introduction (if A is some other mathematical subject). And, typically after such an introduction the question won't be that bad either. Perhaps not great, but reasonable. And, if it turns out to be really simple, chances are the questioner will be the first to admit.

    Indeed, it is my opinion that there is perhaps a lack of such reasonable questions (moderately advanced graduate level, things that are standard knowledge for experts but might well not be known outside this circle,...). Caused by the fact that those that would like to ask them are scared-off in one form or the other. (Not only new user, perhaps also existing users that do not wish to ask something too simple.)

    Tastes are different: but I prefer by a large margin a precise and specific questions with a clear purpose that is just somewhat standard (if it is presented in a reasonable context; not such that one can almost see the homework assignement from which it was copied over) over a pseudo-sophisticated big-picture question.


    I'm in favour of allowing reasonable questions of the form "I'm in field A, studying field B, and I have a homework-like question I want to ask..."

    From the beginning, we've wanted MathOverflow to be useful to researchers realizing they need a result from an unfamiliar field. We don't want to get in the way of this. It should be no shame for mathematicians to ask for help, especially outside their usual expertise! Indeed, if MathOverflow is going to encourage collaborations (not an explicit purpose of the site, but certainly desirable) we should be happy to see this happening. If you're at a big department, think about times you've casually asked someone a question at tea, expecting that they'll have a ready answer. To the extent we can reproduce this dynamic we should!

    I add a vote in favour of questions like "I'm in field A, studying field B, and I have a homework-like question I want to ask..."
    • CommentAuthorgrp
    • CommentTimeOct 2nd 2011

    Indeed, I put in an example of a knowledge request so that my meaning would be clear. I did not put in an example of a poorly worded request because I trusted others to come up with their own poorly worded examples, e.g. "I'm studying field B and I have a homework-like question I want to ask." Further, I would think that reasonable requests might include beginning as well as intermediate undergraduate level questions, whereas this forum is littered with examples showing other community members do not think such questions are reasonable. What is accepted is what the community accepts: there is no current (and from my view, nontrivially characterizable) standard for accepted or reasonable questions on this forum, despite what is stated on the web pages. This is primarily due as much to the looseness and informality of the operation of this forum (which has some major advantages) as it is due to the variety of experiences and interests of those participating in and partially moderating the forum.

    Indeed, we may never arrive at a characterization of what is acceptable. However, if we have a community approved list of what it is we would encourage, such as well formed knowledge requests (e.g. "I'm an architecture major who is studying set theory, and am wondering if I can use the axiom of choice to help clients in making design decisions; by-the-way, why would certain formulations be weaker, say given a surjective function that there exists a right inverse might be equivalent, while given an injective function there is a left inverse is not equivalent to AC?") and properly worded reference requests, well hey, some people just might follow them. I see it as an opportunity to optimize some of the functionality of information transmission, as well as potential for creating a searchable database from the attempts.

    Anyway, since the knowledge request example generated some thumbs ups, I will return with a few more suggestions to post on the how to ask pages. I encourage other readers of this thread to submit examples of what is desired.

    Gerhard "Ask Me About System Design" Paseman, 2011.10.02

    • CommentAuthorWill Jagy
    • CommentTimeOct 2nd 2011
    I had an idea yesterday, it still seems worth saying today. One desired under-represented population is graduate students and postdocs, fear being a factor. Fair enough. Part of MO history is the Secret Blogging Seminar. Just after I had my Berkeley orals (early 1980's) students began having mock qualifiers, where peers held an oral exam of the victim. The idea was to reduce nervousness, possibly to point out weak areas. I would hope there has been a reduction in students fainting during orals.

    So that is the idea, have students discuss possible questions, including the wording thereof, with peers before posting on MO. I would guess the discomfort of doing this is comparable to the discomfort of a bad experience on MO. For a person who cannot discuss a point of confusion with peers (or professors), MO is not actually the main problem, nor is it the best cure. I can add that if I had not been too proud to participate in some of the study groups that began about my time, things might well have gone better for me.

    So, I think this would work for some people we would like to help: "My buddies and I were drinking beer and wondering about the category of skeezix-related pontoons, and we were able to prove subacidity, but we were unable to decide whether all these were babdaroon. In particular Fred, before hurling, suggested Voltaire's best-world argument, but we could not make it work. What do you think?"

    There is no obvious way to disseminate such an idea, but perhaps some meta readers could give it a try.
    • CommentAuthorquid
    • CommentTimeOct 2nd 2011

    WIll, I am sorry, but I do not understand the intent of this:

    So, I think this would work for some people we would like to help: "My buddies and I were drinking beer and wondering about the category of skeezix-related pontoons, and we were able to prove subacidity, but we were unable to decide whether all these were babdaroon. In particular Fred, before hurling, suggested Voltaire's best-world argument, but we could not make it work. What do you think?"

    • CommentAuthorWill Jagy
    • CommentTimeOct 2nd 2011 edited
    Quid, sorry, it was a pretend question by a hypothetical graduate student about some mythical topics. Skeezix was a character in American cartoons, I know of no use of the word pontoon in mathematics, nor of the word acid or subacid(not a word I think), and so on. I made up the words. Voltaire's character talked about his being "the best of all possible words," and "babdaroon" is a name that Lord Dunsany realized (too late) that he should have used for the name of a mythical city, the name he actually used being "Babbulkund."

    Anyway, a pretend MO question about pretend mathematics, the point being it was posted only after discussion with other students, therefore a better constructed question than would otherwise have been posted.

    Anyway, thanks for asking. Usually I get no feedback on my big ideas.

    @Will, I think something along the lines of your suggestion might be very helpful to add to the "How To Ask" page. I think maybe taking out the literature and pop culture references will help in readability, though.


    My opinion is that graduate-level is exactly the right level for MathOverflow.

    • As a questioner: If I have a question in my own field, then I probably know a better place to ask it than MathOverflow. Also, the argument that it's good to have these questions in public for everyone's benefit has less weight since fewer people would be able to understand it, or be interested in it. A not-very-good example might be

      On the other hand, if I have a question in another field, then I can probably get up to the level of a graduate student without outside help, or (if I'm lazy), I could get the graduate-level background to understand the answer to a question without further help. Again, a not-so-good example might be

    • As an answerer: If I see a question by an expert in a particular field, my default assumption is that I'm not going to be able to contribute. If I see "Bill Johnson" on a functional analysis question, or "Allen Knutson" on a topological one, then oftentimes I don't even bother to look at the question! Unless the question happens to be in my exact area, if it's above graduate level then I'm unlikely to be able to answer. Or if I am, then I'll have to give it more thought and effort than I would like (though, to offset that, if it's near to my area then I'm more likely to be interested enough to give it that extra effort), and be less sure of the correctness of my answer.

      A graduate-level question is still useful to me when I answer it. It will remind me of the basics of my field, and bring to the fore things that I'd let slip to the back of my mind. Anything lower than graduate, and that benefit rapidly vanishes. Yes, there will be the odd question at a lower level that will have that effect, but it becomes increasingly rare and the downsides of wading through the rest vastly outweigh the benefit.

    • As an observer: graduate-level questions (particularly in fields other than my own) are those where I'm still able to appreciate the question and the answer.

    Since the dangerous side is the lower side, I'll often emphasise (here) my desire for higher-level questions on this site. But it's not that I set the boundary at graduate-level and regard that as "acceptable on sufferance"; no, it's that graduate-level is the best level, but the boundary just below it is sharp.

    Of course, "graduate level" varies a bit from region to region so this isn't so well-defined, but I think it's defined well enough to be workable.

    I'd also enter a plea for not making the boundary too distinct as there's never going to be a line that we can all agree on, so I'd rather have an area that all know is a bit grey (and act accordingly) than a distinct line that almost everyone disagrees with.


    I commented on Math.SE when I answered (, that I thought it fell in the gray region acceptable at both sites.

    That's a somewhat interesting example because it's literally an exercise from a book. You could definitely argue that it's senior level undergrad rather than first year grad level, but still it's a fun question that I think would be fine at MO.

    • CommentAuthorquid
    • CommentTimeOct 3rd 2011 edited

    Following Francois's posting, a concrete suggestion:

    The following is more a technical point, but it comes up relatively frequently (indeed, just now again), so it might be worth mentioning explictly somewhere.

    The problem I hope to address by the text below:

    There is a fair number of question that are written in what I would call classical excercise/homework style. That is: "Suppose this and that. Let something. Prove a statement." And nothing or at least not much else.

    Sometimes I believe this is due to the fact that it is really just copying over homework or alike, but sometimes I believe it is not due to this. In particualr, up to a certain point in ones mathematical training/career one could get the impression that to write in such a wat iis the way to write a mathematical question. After all one saw many of them written by experienced people, and perhaps not much else. And, in some contexts it is a good way.

    Yet, to write in such a way on MO seems not a good idea to me. First, it raises 'homework-alarm' red-flags for some people. Second, some do not like the imperative style. Finally, for me the main problem, there is no motivation. In some sense all this is addressed in FAQs / How to ask. Yet, perhaps it is worth making it more explicit.

    A draft of what I would write to avoid this [perhaps for the Motivation-section; at the moment the closest analog seems to be in the 'Be precise'-section].

    "An MO question written in the style of an excercise in a book or a problem on a homework assignment is tyically too terse to make a good MO-question. Please keep in mind that, for example, an instructor in a course has fairly precise knowledge of the background of her/his students and likewise the students know in which context the problem is posed. By contrast, there is a priori no joint background and context between the person asking the question and the persons potentially answering the question on MO. Thus it is important to create some minimal common basis by mentioning some information in addition, even if in a strict sense the question would be self-contained. For example, you could mention what motivated you to ask the question on MO. This should not be misunderstood as meaning that there has to be a severe or urgent reason to ask the question. Something along the lines: 'I read the similar statement [details on it] in [book or paper]. This seems like a natural generalization, but I could not figure out whether it is true.' is a fine motivation."

    • CommentAuthorWill Jagy
    • CommentTimeOct 3rd 2011 edited
    Alrighty, I printed out How To Ask. As Ryan suggested, here is my little proposed section, less about technical issues and more about personal/social:

    **Consider asking your peers to help with formulating your question**
    MO is interested in attracting more questions from graduate students and postdocs. We are reliably informed, however, that many are fearful of being treated harshly as MO newcomers.

    One approach, available in existing university programs, is to discuss a potential question with one's colleagues before posting. A group of students might well be able to suggest improvements in the formulation of a question. Meanwhile, a person posting may feel less exposed knowing peers have contributed to the post. For someone accustomed to working alone, such personal discussions may be a valuable introduction to the world of give and take that is embodied in the journal referee process.

    In another vein, an excellent way to find out what makes a good MO question is to answer several, over the course of a week or two. Answers could be composed with one's colleagues as well. Note that several groups contribute solutions to the problem section of the M.A.A.'s American Mathematical Monthly. One might also answer questions at
    Dear Paul:

    I would like to address the "improving the questions" aspect of your post, as it comes up pretty often.

    This question strikes me as a typical example of a question with room for improvement:

    Yet, I don't think improving it is appropriate. As a mater of fact, I think it seldom is: modification should be reserved for trivial details.
    I think the MO community should make a point of answering the actual questions of users, not the questions we wish we saw, the questions we could make out of a given post.
    If you modify the question too much, who can tell you if your new version is what the OP had in mind? Especially in the case when the OP is already unclear on what their question is (examples are legion here), even they would not really know.

    And MO is research-level: don't we want to promote the idea that asking the right question is often more important than getting an answer?

    As it is, I think the question that I linked to has exactly what it needs: a good, simple example to show the OP that there is really something non-trivial going on, as well as enough indication that a more complete answer requires a thorough clarification of the question first. We can hope that the OP now realizes that the question was not as innocuous as they imagined, and I suspect the OP might not even want to know more than that right now (as it would involve delving into algorithms and decidability deeper than they might like to, not to mention precising in what ways a vector space can be constructed). I don't see what would be gained for the OP if we rewrote the question and answered it.

    (PS - I never really realized how careful one had to be around "improvable" and "unprovable" until now.... Still, I think the point is made.)

    As a minor point on Thierry's post. There is a third way. If the question needs significant improvement before it becomes the question you would like to answer, you can always post a whole new question; link to the old and explain how the new question is a refinement (or whatever) of it.

    (So, to be clear: I agree with Thierry about not modifying the original question to a whole new question, and I'm pointing out that it is possible to have ones question (cake) and answer (eat) it.)