As a followup to this meta thread, I think it might be a good idea to have a timed closure method for 'very low quality' posts.
Currently, mods seem to handle low-quality posts by closing them as 'not a real question'. Sometimes there is, in fact, a question hiding in there; however, this 'close first, ask questions later' policy doesn't give anyone time to edit the post to improve it, and gives the original asker a bad impression ("Not a real question? But I did ask a question! This site sucks!").
Of course, it's possible to reopen the question after editing and improving it; however a new user may not be aware of this (and obviously can't vote for it themselves). Once the question is closed once, it's probable that the new poster may simply go away, not to come back to see the answer when (and if!) it is reopened. Moreover, this 'not a real question' closure makes no sense to a good-faith poster; they did post a question, after all, and they know it. So they have no idea what to do to fix it.
Therefore, I propose a new mechanism: When a moderator (or perhaps a certain number of high-rep users) sees a very low quality question, in addition to the usual close reasons, they have an additional 'low quality' close reason. When this close request goes through, instead of closing immediately, it adds a notice to the question itself ("This question has been flagged as very low quality; please edit it to conform to our formatting standards, or it will be closed in X minutes" or something), and starts a short timeout (say, 10 minutes) before the question is actually closed.
In the event that an answer is posted, or the question is edited (apart from tag edits), the closure is cancelled, and an automatic moderator flag placed (to remind the moderator to check whether the question has, in fact, been improved).
This, I think, would strike a good compromise between creating extra work for the moderators, and avoiding discouraging new users; the moderator only really needs to visit truly low quality (or abandoned) questions once, but valid questions that (due to cultural differences, or misunderstanding of SO rules, or poor English questions) are initially hard to read have a chance to be edited into something that conforms to policy before being closed.
Note that I'm not proposing any change to the existing close reason system - only a new close reason with this timeout.
Edit: To respond to some criticism of the idea:
- "What if people just do dummy edits to clear the flag?" - It's unclear if this will be an actual problem. If it is, there are a number of things that can be done - for example, the question could be added to the community review list for re-examination, rather than going straight back to a moderator.
- "Isn't this similar to the idea of asker purgatory?" - I'm not asking for a system to correct repeat offenders - just a bit of leniency for people who have their hearts in the right place, but are new, and possibly didn't read that wall of text that comes up when you first try to ask a question. It may be a good idea to have the delay only apply for the first question on a given account, for example - further closures would result in immediate closure (but with a note saying that you should edit the question then flag for reopening).
- "But what's wrong with just reopening in the first place?" - New users to the site are likely to see closure as something permanent. Certainly the "Not a real question" close notice doesn't give any indication that it may be fixable. And even if the user edits their question and mod-flags for reopening, they'll have a low flag score, and it may be some time before a moderator gets around to reopening the question (by which time it'll have fallen off the main page). As such, having one's first question closed within minutes of opening is likely to be very discouraging to a new user. Having a notice tacked on asking for it to be edited with a deadline, however, is far more likely to push new users in the direction we want.