I feel like I'm partly responsible for this, having brought special attention to the question and all, so I'd like to throw in my two cents.
The question itself is a useful resource, but that's only because the answers are of a reasonably high quality. Luck factored into this one; questions phrased the way this one was originally phrased have so much potential to end up like a reddit comment thread, and usually, they do. What's wrong with it?
- It explicitly asks for opinions ("disagree with");
- It doesn't ask for evidence, or imply that answers should be backed up;
- It doesn't even ask for any elaboration.
In short, it fails the poll test. Fortunately, it happens to be a subject that many programmers are passionate about, and so many of them took the question seriously and answered it with the requisite amount of detail and care. I still think that luck and timing played a major part as well; when I open a question with a title like that, I expect to see dozens of one-liners, and it's kind of miraculous that there weren't any here (or maybe there were, and they got edited/deleted).
The quality of answers clearly indicates that it should not be deleted - but that doesn't necessarily mean it should be open. Closing a question sometimes means "this question is going straight to /dev/null", but sometimes it also means "this question needs an overhaul so it doesn't start attracting crappy answers". In this case, I felt the latter would have been an appropriate rationale, albeit a sudden and unexpected one.
I've put in a major edit to the question representing what I think would have been a more constructive version; I have a feeling it's going to get rolled back within minutes, but at least I tried, and anyone can view the edit history to see what I'm getting at here.
Now, off the subject of the specific question and onto the larger moderation issue, there are a few points to note here:
The more open-ended a question is, the more people become compelled to upvote it because they enjoyed the question, and not necessarily because it was useful and clear as upvotes are intended to be used. Popularity has to be taken with a grain of salt for certain types of questions, and moderators have to walk this tightrope frequently, sometimes overruling the community.
There are dozens if not hundreds of questions on Programmers.SE that were asked during or shortly after its "free for all" phase (before the "constructive" criteria were introduced), and it's unreasonable and unfair to expect all of the high-ranking community members to slog through the entire history and use up their close vote quotas. I actually spent one day here going through what I felt were the worst questions and flagging them; I got exhausted after about 10 of them.
The improved flagging system allows lower-reputation users to flag for moderator attention based on a specific close reason. This strongly indicates that, yes, it is OK and expected for moderators to occasionally intervene if the community is not pulling its weight. If the community was perfect, if it always did the right thing, then we wouldn't need moderators, right?
Although Programmers.SE probably has the highest number of 3k+ (close-rep) members of any of the Stack Exchange communities, there are still only about 70 of them. That number is good, but it's not huge. The trilogy sites have hundreds or even thousands of these privileged members, so it's natural to expect a high level of autonomy there; the same is not true of Stack Exchange sites where the number of these members is relatively low and many of them are still learning what Stack Exchange is all about.
As Jeff says, Programmers.SE is a bit experimental, and has a lot of checkered history to cope with. Since many of the rules are new and have not been followed or enforced consistently in the past, members here don't have a great example to follow, and they're left scratching their heads and saying "well, this question is pretty open-ended, but it's not as bad as this one over here..." and of course, there be dragons. Moderators are uniquely qualified to do the cleanup work necessary in order to make the site more consistent and set a strong example for the community itself.
Honestly, if I could sum all of this up in one sentence it would be that Stack Exchange sites tend to need leadership, not just moderation, and the diamond moderators are in the unenviable position of wearing many hats. Once the sites mature, then the role of diamonds will become less and less important over time.
I don't know how many people will agree, but I hope I've done a good job of representing the position and responsibilities of P.SE moderators and really all SE moderators here.
