Very minor edits are often frowned upon because they don't do "enough", but as long as they improve the post somewhat, I never understood why they are problematic for anything but the most significant posts on the site.
The biggest gripe people tend to have is when a post is riddled with severe problems, and someone takes time to submit an edit suggestion, but only fixes 1% of the problems. Hey, if you're going to help, can you at least apply your help to the entire post? Even if that only means cleaning up grammar and markdown throughout the post, that's still a whole lot better than fixing one typo when there are 18 more typos and 4 markdown issues remaining, to say nothing of content problems.
On the flip side: like Wikipedia, we tend to care about and curate our famous/well-known posts, and we care a lot when anything happens to them -- kinda like a "soft" protection on a Wikipedia article. I could see someone justifying to reject an edit as "Too Minor" if someone tries to edit a very popular question or answer with just a small one-character fix. For one thing, it replaces the OP's profile link next to the question/answer with the guy who edited it most recently.
My guidelines would be to accept the edit, if you can answer "Yes" to all of the following:
- Does the edit suggestion objectively improve the post's quality, without interfering with the OP's intended meaning?
- Is the post being edited fairly obscure (i.e., less than 20 upvotes), OR a Community Wiki, OR is the edit very substantial?
- Is the post not in the "Closed" or "On Hold" status?
If you answer "Yes" to the first bullet, "yes" to at least one of the disjuncts in the second bullet, and "Yes" to the third bullet, you're good to go for accepting the edit.
For the third bullet, I specifically want to emphasize that you should be very wary of approving trivial edits on Closed / On Hold questions or answers thereof. The reason is that, when these edits are approved, the question is automatically chucked into the Reopen queue. If a question is on hold and the suggested edit does not completely resolve the reason for the question being put on hold, you should Reject the edit.
If one person is repeatedly flooding the edit queue with bad edits (incorrect edits, or edits that are way too minor) and it's interfering with the front page, mods may superping them to ask them to slow down.
I don't think it makes sense to punish "innocent" users who may not be aware of the ideal best practices for editing, by rejecting edits that are "Too Minor" all the time. Instead, what I'd do is to flag an edit if you repeatedly see the same user making edits that don't resolve all problems in the post (or at least, not even a whole category of problems, e.g. grammar or markdown).
OP says "Teh qcik; brwon fxo jmuped ovre teh lzay dgo."
Edit attempt #1: "The quick, brown fxo jmuped ovre teh lzay dgo."
Edit attempt #2: "The quick, brown fox jumped over the lazy dog."
Edit attempt #3: "The quick, brown fox jumped over the lazy dog because his Windows laptop had a BSOD."
We would always (gladly) accept Edit #2; no ambiguity there, and they fixed every grammar issue with the post.
For Edit #1, we would accept this conditionally. If it's a new user who's just starting out and he makes that sort of edit, I'd accept it, personally. But if I notice the editor has many edits and they're all of that type, I'd flag it, or ping a mod to superping the editor.
For Edit #3, since content and grammar are edited, you have to be careful and make sure the added content jives with the rest of the OP's prose and that the editor isn't introducing new information or changing the OP's statement to mean something different. Other than that, good edit.