Quoting from the question:
It seems to me that when a borderline question comes in, we have a few options:
- close it (with a good explanation of course)
- comment on it asking for clarification or rewrite
- edit it into an acceptable form
These options are not mutually exclusive.
Please keep these points in mind:
- Closed questions can be edited;
- Closed questions can be reopened;
- Question closers can (and often do) leave explanations;
- Question closers can (and should, when possible) leave some advice regarding possible improvements.
- Anybody with sufficient reputation can vote to reopen a question - it does not have to be the same people who closed it.
Regarding edits, every once in a while I see a question where I'm 95% sure I know what the author is getting at but s/he just didn't express it well, and in those cases I will make substantial edits and leave a comment to the original author, making sure that s/he is aware of the option to revert my edits.
Aside from those rare cases though, I only edit for readability: spelling, grammar, punctuation, formatting, and so on, and occasional "tweaks" like parenthesized temperature conversions. I will not make an edit if I feel that such an edit would be substantially altering the meaning of the question and the kinds of answers it would attract, and even though your editing privileges are yours to do as you please with, I wouldn't recommend this to anybody else, either. People are unlikely to respond any more positively to blind edits than they are to having their questions closed.
For most of the grizzled vets on SO/SE, there's a pretty consistent methodology for dealing with these questions with a few minor variations. Basically they start with these two steps:
If the question is subjective and not wiki, flag for moderator attention. Don't leave a comment asking the author to do it; many won't understand, some won't be able, all will be annoyed, and already-submitted rep-farming answers will remain in non-wiki mode.
Determine the most applicable close reason for the question. If the question can be made more objective / less localized / more coherent / etc. (whatever would prevent it from being closed), without altering the question's meaning, then edit it. Stop.
Assuming the question can't be edited due to insufficient detail (which is the case for at least 90% of closed subjective questions) then it goes down one of the following two tracks:
- Leave a comment warning that the question as written is likely to be closed.
- Suggest how the author could improve the question.
- If the author fails to respond within a reasonable time (i.e. a few hours), actually vote to close.
The other strategy, which is the one I prefer, is:
- Immediately vote to close.
- Leave a comment explaining (politely!) why you voted to close.
- Suggest how the author could improve the question, or appeal the decision on meta.
- Return to the question several hours later, and if it has been closed and also edited (improved), then vote to reopen it.
(Obviously I have to modify these somewhat as a moderator, but this was how I always did it beforehand.)
The reasons I like the second strategy better are (a) it quickly stops the flow of potentially poor answers, thus helping to ensure that the author actually gets something useful out of it, and (b) it gives the author as much time as he/she needs to read the FAQ and tidy up the question. This is much easier for the typical user who only visits once a day (or less).
And one of the things you'll often find with poor questions is that their authors are hit-and-runners or simply ignore or respond with hostility to your comments. Much of the time, they never bother to improve their question, and in the meantime every other jimbo is hurling their one-line fluff answers at the question and turning it into a useless mess. By the time an edit does happen (if it ever happens), it's already too late to separate the wheat from the chaff.
Prior experience with Stack Overflow has also proven beyond a shadow of a doubt that when people are too eager to edit certain users' poor questions, those users will simply rely on community edits as a crutch and continue to ask poor questions. This may sound like a slippery-slope fallacy but it's already happened on other communities and those communities are suffering for it.
So I am a firm believer in getting a poor or off-topic question closed as quickly as possible, to minimize not only the question noise on the site (harmful to community) but also the answer noise within the question (harmful to questioner), which is why I don't think that simply leaving a "helpful" comment really works as a long-term strategy.
My silly but hopefully educational cooking analogy:
Vague/subjective questions are like a big pile of raw ground chuck. You want to make hamburgers, but just realized that you don't have any onions in the house, or eggs, or flour, or salt, or any of the things that are normally required to make a hamburger from scratch. What do you do?
Well, you can raid the pantry and start throwing a bunch of random crap into the pan with it and see if it comes out OK once it's cooked. It probably won't, though. Or, you can freeze it, go and get the ingredients you need, and then thaw it when you're actually ready to cook it.
So that's what we're doing here. When we close a question, we're saying that it can't be answered well because it's missing details (ingredients). Once those details are added, we can pick up where we left off. On the other hand, if that never happens, and the question just sits around forever and goes rancid, then we will actually delete it, and that is generally final.
Part of our mission here is not only to give people the right answers but also to teach them how to ask the right questions, because asking the right questions is crucial to become an expert oneself and subsequently a contributing member of the community. If we can do that without being rude or unfair (and I do not believe that any of this is rude or unfair), then we should do it as much and as often as possible.