I was one of the reviewers who rejected the edit. This question has stirred a lot of discussion, so I'll try to refocus this answer a little to better address the issue that's been raised.
###New vs. old questions
Most questions get their answers soon after being posted. That's the most useful timeframe for the OP and the period in which the question is most visible on the site and attracts the most attention. It is often the timeframe in which the subject of the question is the most relevant. Improving a new question can help to get it answered, so any edit that makes it better should be approved.
There's nothing wrong with trying to improve an old question, but that leads far less often to new answers. The main benefit to improving old questions is usually to make them a better resource and easier to find in searches. As I'll describe below, there can be a "price" for bumping old questions in terms of reviewer time and reducing exposure for new posts. So in deciding whether an edit to an old post is an improvement, there is an argument for considering whether it will accomplish something useful, like:
- helping to attract answers, or attracting better or more current answers
- making the question clearer or easier to find if it is something people will be looking for and it could be a useful resource.
###General Philosophy
Before getting to the specific edit mentioned in the question, consider edits in general and the SE philosophy. As Braiam's answer and comments cover, the SE philosophy is to, in general, approve any edit that makes a post better. I have no argument applying that to new posts. For old posts, we need to consider the definition of what constitutes improvement.
At least for me, improvement relates to the utility of the question--does the change make the question a better resource, not does it improve compliance with English rules. These aren't always the same. Whether or not an edit improves a question's utility depends somewhat on the context of the question, itself. Consider these two extreme examples:
A canonical question. This is a site standard that will be frequently referenced. In this case, I would say that virtually any change that polishes it in any way is probably useful and benefits the site.
An example at the other extreme: an ancient thread about a problem specific to one version of some long-obsolete software. Nobody uses the software anymore, the problem was fixed in a subsequent release, none of which is currently available. Maybe there's some historical value in case some hobbyist is dabbling with it. It attracted several marginal answers that offer conflicting and mutually exclusive solutions, so even if someone is interested in the question, the thread isn't a good resource. Even at the height of the question's currency, it attracted little general interest. It has been harmlessly buried on the site for years.
We could talk about potentially deleting it, but that's a different matter. For this hypothetical example. maybe that would happen later, or maybe a decision is made to leave it alone. But in this example, we have a new user who figured out that he can run a site search on a common misspelling or slang term to identify a bunch of easy edits. This long-dead question was in the list and he has proposed an edit to fix that.
I submit that nothing would really improve a question like this in terms of its utility. That's especially true for a trivial or cosmetic change of this nature, which provides no real benefit. Worse, engaging in this kind of serial trivial edit proposals on old questions eats up the time reviewers have available. If the edits are approved, they flood the main page, wasting readers' time and shortening exposure for new questions. So there is a benefit to the site and to users in general to reject non-useful edits so as to not encourage them and to educate new editors on useful priorities.
In between these two extremes, there is a continuum.
One school of thought says what I described for the canonical question should apply to any post that is not at the extreme lower end (and maybe even those); the context of the question, itself, shouldn't influence whether an edit is an improvement.
The other school of thought is that "improvement" should be measured in terms of utility as a site resource rather than as mere compliance with English rules; context should be considered in assessing whether a change is an improvement.
###Evaluating old questions
I look at edits in three "classes" of questions:
- Canonical and similar high-value, high-usage questions (and new questions): accept any edit that makes it better in any way.
- Old, useful questions: apply a little judgement to the above rule; generally reject trivial changes that don't really improve anything in a meaningful or useful way.
- Old flotsam: generally reject edits unless they make a substantive difference in the question's utility.
So how do you decide which category a question belongs in? These are the kinds of indicators I look at:
How old is the question and when is the last time it had any activity?
Note that age is not a criterion, only an indicator in the sense that the older a question is, the more likely it is to have age-related factors that might make it currently irrelevant.
Is it still relevant (is the software or hardware still in use; are answers applicable to current versions if it is; etc.)?
How useful was it when the question was fresh (votes, views, favorites, comments)?
If a question attracts little interest when it's fresh, it's rare to attract a lot more interest or usage when it's old. But if the subject matter is still relevant, look at whether the reason it might have attracted little attention was due to problems that can be fixed with an edit.
Are the answers useful and consistent or did it attract a motley collection of low-quality or conflicting answers?
If the question is unlikely to attract new answers, and the existing ones aren't useful, what's the sense in making the question easier to find?
###Evaluating the edit
To what extent could the question use improvement and how much of the obvious improvements does the edit accomplish?
If the question needs a lot of work but the edit fixes only one item of low-hanging fruit, the editor should be encouraged to do a more thorough job. That encouragement, or lack thereof, will affect the quality of their future edits.
Does the edit improve the question's utility?
Might the edit attract new answers?
Might improving the question make it a significantly better resource if the information is "timeless"?
###The edit in question
Bringing this back to the edit mentioned in this Meta question, that question struck me as a very dated and localized issue with little current or general relevance. It was already answered and wouldn't benefit from bumping. The edit made the title more precise, but I would argue that this "improvement" had negative utility in this case.
Consider that the question was specific to the OP and the thread wouldn't generally be helpful to other users today. The original title wasn't helpful for searches, which was maybe a good thing in this case. The new title is more precise for the question, but still generic. Now we have a bumped question with a current date and a title that could seem to apply to different, current issues. So a currently irrelevant question that was previously buried, may now be wasting time for people who see an apparently current question and think it might be applicable to them. Not every improvement is really an improvement.
###Example of title edit approved for old question
To the question of whether a title-only edit on an old question is always bad: no. Here's an example of another old post (6 yrs old), where I approved a title edit only: http://superuser.com/review/suggested-edits/625545.
The question attracted 14 upvotes, 8 favorites, and almost 13K views. There is an accepted answer that attracted 18 upvotes. There has been recent activity. The question could use a little polish, but has no issues that prevented it from being useful. The title was awful, but the thread managed to be useful and attract a lot of readership despite it. This thread is a resource that appears to still be relevant, and just a title edit made it more useful.