I’ve just again come across a post where the answers didn’t seem to fit the question well, and a look in the edit history showed that the question had been substantially altered after the answers had been given.
As this is not the first time I’ve seen something like this, I wonder if there could be technical measures to prevent this.
Obviously blocking all edits after an answer has been posted would not be a good idea. It would block all possible improvements of the question, including those that don’t invalidate the answer.
If the distinction between large and small edits gets implemented, one might consider allowing only small edits after an answer was posted.
Another possible measure would be that the threshold above which edits no longer need review is higher for questions with answers, than it is for questions without answers. Possibly with a grace period, so that question edits generally don’t fail because someone posted an answer while the OP was already editing the question.
It’s entirely possible to make even major edits to a question without invalidating answers, but I do think it would be worthwhile to require review of edits to questions with answers without regard for whether it’s the OP editing or someone else (but still allowing sufficiently trusted users to edit with fewer or no reviews as usual).
That said, I doubt we’ll need this for quite a while; it’s enough for MVP and some time beyond to simply have a site culture of not allowing chameleons.
I don’t think there is a technical solution for this. People will have to judge whether an edit changes the question too much. At most the system can provide a flag for that reason.
Quite simple - we force edits on questions with accepted answers to always require review (and you are unable to review your own edits). The only problem I see here is people ignoring the question itself and simply accepting the edit, but that’s an unrelated problem that I’ll likely make a new topic about.
Disagree. The few times I did actually ask a question, it was useful to be able to update the question to include more information in response to comments.
The way to deal with this is to make the rules clear, and have there be consequences when they are broken. I’ve certainly seen cases where questions have changed out from under existing answers, including my own, but that’s rather rare.
Accepted answers requiring a review isn’t a bad idea IMO.
Something else (which is more bothersome) is updating old questions or answers. As is being discussed elsewhere, sometimes a newer version of things needs to be addressed without being a duplicate, and digging through answers for the “latest” version isn’t the right solution. Maybe requiring review for questions after a certain number of days/weeks/months would be beneficial as well.
This could work out incredibly well for Dealing With Obsolete Content - #25 by Corsaka also. Every 6 months/1 year, we can send questions to a review queue where they’re checked for inaccuracies or problematic edits.