I think the title says it all. Personally, as someone who uses the "newest questions" as the front page, I'm not keen on people tweaking questions or answers that seem to have outlived active interest.
Added 2013-07-08: I see we are now getting nit-picking corrections of English. Well, more defensible than mere reformatting, I guess, but in my view often just as misguided.
I think we should make sure all questions and answers display correctly, even if old.
[Edit: By this, I mean, necessary changes, that add clarity, should be welcome. Adding spaces, correcting trivial typos, and the like, should not be the focus: They clutter the front page, and at least currently, seem to be happening simply as a means by low reputation users to gain reputation. This seems to be an abuse of the system and I think they should be discouraged. End edit.]
Hopefully the questions here are and will be useful well beyond the few days when they were on the front page. (Isn't this lasting usefulness part of the point of the site?)
Editing in moderation should be fine.
Here's an idea for a new feature; depending on what people say here I can post it as a feature-request either here or at https://meta.stackoverflow.com/.
A "minor edit" checkbox, along with the parenthetical text "minor edits must be reviewed by another user, but do not bump the post on the list of active questions".
It would then do what it says. Since we already have the review mechanism deployed, it seems like it might be quite easy to achieve technically.
I think all editting should be fine (it improves the database of questions and answers), but editting should not automatically lead to the question being bumped. Maybe the question should only be bumped if the person doing the editting explicitly clicks on a "bump" button, or something.
One of the founders of MO told me in person that he thinks it's actually a good idea to bump up your old questions every year or so, and he suggested finding a minor edit to make specifically for this purpose...
In general I am in favor of editing things on MO and I rather thought this happened too little in the past. However, via this edit-approval feature I just was made aware of two potential edits that rather do not seem necessary to me and fall, in my opinion, plainly into the category Yemon Choi mentioned; they are purely tweaking the display of formulas that are alright (though not good), anyway, of it seems old questions.
They would be an improvement, yet in my opinion the improvement of just having parenthesis of a better size, for example, is too small to warrant the noise.
I did not vote "reject" but rather "skip" as momentarily I was unsure what is consensus and did not want to reject something just so, also since the descriptions says for reject that the edit is wrong (which it is not, in an objective sense, only in a meta-sense in my personal opinion). But personally I am rather inclined to reject such things in the future, in particular if they get frequent.
Added: About a week has passed. Turn out was not overwhelming but not as minimal as the score could suggest. Thanks to everyone who participated. At the time of writing we have +4 vs -4. So, opinions are mixed.
My personal conclusion: since there is no clear consensus against this edit (if I were to include my opinion in the count there would be a minimal majority against it though), I intend to approve (somewhat against my very personal opinion) such edits in the future (I did not reject much during the last week either), and by extension most suggested edits (except of course if there is some actual problem with the edit). The reasoning being that sometimes such edits being approved and sometimes such edits being rejected should be an annoyance for those suggesting them in good faith.
If somebody would still like to express their opinion, feel free to continue voting. I will continue to watch the score as I am generally curious about existing opinions.
I would like to revisit this matter with an example. (I am not sure if I rather should edit my existing answer or ask a new question but anyway.)
It seems to me there is no real consensus reached at all. There is one highly voted answer, however, it is very unspecific so that everybody can read it their way. (I did not get any reply for my request for clarification on what the poster, Andres Caicedo, meant; let alone finding out what the voters mean.) The bandwidth of in my opinion reasonable interpretations ranges from editing should be done if an only if something displays in an actually wrong way to every improvement however small is an improvement and should be done.
Yet, from my observation and his comment-reply to Yemon Choi I think that Andres Caicedo's and my views are actually not that incompatible, somewhat in the middle of the bandwith in that actual correction should be done and the rest if there is significant room for improvement (increasing readablity and clarity in a noticeable way), the answer does however leave much room for interpretation. Thus this follow-up. So, now, for the specific example.
I reviewed the edit of Polynomial bijections from Q x Q to Q now improved to "Polynomial bijections from $\mathbb{Q} \times \mathbb{Q}$ to $\mathbb{Q}$" (this is the full improvement!) I rejected it as 'too minor' one of the default reasons, however, it was approved by two others so it took effect.
As said, the edit consisted in changing only in the title, for the rest it was not necessary as already done long ago, "Q x Q to Q" to "$\mathbb{Q} \times \mathbb{Q}$ to $\mathbb{Q}$". In my opinion, an edit can hardly be more minor than this (also the question was old, also in the sense of not having been active very recently).
Moreover, in a recent discussion on the board now know as tea, the sentiment was expressed by some, not including me I was rather on the opposite side of the debate, that "texifying" is not universally appreciated, in particular not in irrelevant cases (where there is no risk of misunderstanding or decrease of readability), also mentioning loading time as concern. (I do not fully agree with this, but specifically for titles I agree, as the commulated longer loading for lists of question can be annoying.)
So, to cut a long story short:
Should edits like changing "Q x Q to Q" to "$\mathbb{Q} \times \mathbb{Q}$ to $\mathbb{Q}$" once when it is clear the rationals are meant be made even if it bumps the question?
Perhaps let us say vote up for: no, they should not be made, and down for yes.
(I hope it is alright to use answers in this form and induce voting in this way; I CW it though not sure this is relevant on meta.)
I was just browsing through the new badges, and it seems there are two which reward editing questions more than 6 months old (excavator and archaeologist). Perhaps this is why we've seen the spike in old questions being bumped?
Personally, I think it's a bit ridiculous to reward editing ancient questions/answers. I'd be happy to see these two badges disappear. If you agree, leave a comment or something. If people are in favor, perhaps we can make it official via a feature request.
