It would be nice to allow users to mark a change in an answer (or question) as minor (e.g. for a simple spelling or grammar correction). A change marked as minor wouldn't push the question on top of the list of questions on the home page or in the feeds. It wouldn't trigger any notification. This could be implemented with a "Minor change" checkbox when editing an entry (a la Confluence). What do you think?
It would be good if this request were re-evaluated, in light of modern edit reviewing technology.
Paŭlo Ebermann speaks correctly about this issue when it pertains to slower sites. A site that only gets 50 questions per day is still a good site. But if you make 20 edits, you're pushing good questions off the top.
I see this happen on Gamedev.se all the time, when one of the mods goes on a re-tagging spree. It's important to do that kind of maintenance work. But it shouldn't clutter the front page.
This isn't an enforced thing. We're not even asking for the option to be default. Just to have it be there for those who need to make a minor change to a post that doesn't warrant it bumping other content off the front page.
We have tools to see people editing posts. We have tools to see people making edits to old posts. We have tools to prevent these people from making malicious edits.
In short: Robert Cartaino♦'s argument about exploitation is simply no longer valid.
Implementing any sort of "don't bump" functionality would also delay accountability and transparency for those edits.
Notifying users of edits in the system allows them to take a look at the content and make sure there isn't something fishy going on. Imagine if people could make changes to the system without anybody noticing. That is very exploitable.
(I just was typing this for a question which then got closed as a duplicate before I finished. I'll post it here, thus.)
There are several points to think about, here.
The edit bumps questions to front page feature has these goals:
- If a question is edited, maybe it is now easier to answer => People should look at it, and maybe upvote/answer it.
- If an answer is edited, it could now be a good answer => People should look at it to upvote it.
- Maybe an edit made something worse. => People should look at it, and maybe revert the edit, if necessary.
On the other hand, bumping many old questions to the front page has the effect that people don't see the new questions anymore.
For example, in the last some days I edited all questions with "question" in the title on tex.stackexchange.com. After about 6 such edits I got a comment request to slow down ... since my edited questions came faster than new ones. I then changed to a "one per hour" rhythm (on average).
(Some of these questions were further edited then by other users, and I think some even got new answers.)
This is particularly bad for small sites - Stack Overflow has no such problem (for normal editing speeds), though it might affect people who follow the tags which I'm just editing.
So, if you simply don't do this too often, the bumping is not a real problem. (Of course, if multiple people do the same, it again gets a problem ...)
I'm not sure how to solve this - maybe have a kind of "approve" feature so another user will confirm that this edit both
- is not malicious
- does not make the post so much better that it should be bumped
Or, as Michael said, simply put a size limit. But some edits (like adding/removing a "not") can quite change the meaning of a post.
As I understood Mr. Cartaino's answer, this suggestion was declined, because misusing the "minor edit" function could make easier to make destructive or invalid changes below the radar. This is right, but the Wikipedia has a solution for this, and a similar solution is what I suggest:
On the list of the recently changed questions, a "without minor changes" checkbox could exist.
On my opinion, this would solve the problem of the danger of the misuse of the suggested functionality.
Apologies if this has already been mentioned (there are several more-or-less similar proposals, with multiple comments and answers each, but I have not found the same combination).
How about the following?
An editor can click a different (adjacent) button to save as a minor edit. (Single click, not a check box plus click to save.)
Such a minor edit would not add much to the editor's rep, since less time, effort, understanding etc. is needed. It should count for something, however.
Likewise, a review of a minor edit should count less for the reviewer rep, for the same reason.
A reviewer can choose to see minor edits also or only non-minor edits, the latter by default.
When a reviewer views a minor edit it is clearly labeled as such.
A reviewer can still decide to reject a minor edit because it is too minor (not helpful), but will typically be less likely to do so. The edit is not trying to pass itself off as a significant change that is worth much. It is trying to pass itself off as an improvement, even if minor, and it can be rejected if the reviewer thinks it adds nothing.
The advantage for an editor is that the edit is less likely to be rejected for being too minor. S?he will get some points for helping a little, but not many points since helping only a little.
The advantage for SO is that cosmetic edits will get done instead of being avoided, improving readability and thus communication. There might be fewer comments asking for clarification etc., as users help each other by cleaning things up in "insignificant" ways.
An editor can decide at the time of saving whether s?he thinks the edit is worth normal points (and thus risk being closed if a reviewer thinks it is too minor) or worth only minor-change points. IOW, the editor judges the editing result when done, and makes a bet wrt its worth.
#4 might take care of the problems mentioned wrt minor edits being sent to the top/front and crowding out normal edits and creating extra work for reviewers. A reviewer would see them only by explicitly asking to see them.
Minor changes might not get accepted as quickly, since some reviewers might not bother with them, as they are worth less. And that might not be altogether bad -- they are after all less important. But at least they would have a chance at being accepted -- they are some help, even if not a great help.
Again, apologies if these suggestions have been made before. I didn't find this combination suggested.
Postscript: Perhaps another possibility would be to have reviewers of an edit click a new button,
OK as Minor, which would accept the edit as a minor edit, with as consequence reducing the points to be awarded to the editor.
I really think it's a shame to reject edits that improve things, even if in only a minor way. Such edits are being rejected not because they reduce the quality but only to fit the SO rewards system, which in this case is fairly coarse-grained: significant improvement or none. That's not ideal.
Instead of being denoted manually, some sort of heuristic could be used - e.g. if fewer than 1% of the post's characters have changed, it's a "minor edit". It might be tricky to get right though.
I am a volunteer Wikipedia editor and it annoys me that I cannot make minor copyedits here. People should be able to make edits of less than 6 characters long.
Recently Jeff repeated his opinion about not supporting a trivial edit checkbox:
I definitely do not support a "trivial edit" or "hidden edit" flag. All edits need to be vetted by the community, and hiding them is not the right way to accomplish this goal.
(This matches Robert's answer, and I agree, even though nowadays there's peer review for those <2k.)