I have a privilege to review pending edits for answers. One edit changes the answer considerably - not only providing minor fixes, but adding external knowledge and changing the answer to the point that it can't really be called the answer of user X, but a collaborative effort. Should I allow such change, or decline it because it modifies answer too much?
3Is it good information? Does the edit make the answer clearly better?– terdon ModMay 1, 2016 at 19:07
1I've occasionally wondered the same thing. Interested to hear the community's thoughts.– Jeff Schaller ModMay 1, 2016 at 23:30
The Stack Overflow (site, not company) policy on this is to reject any edits which change the content, and not form, of the answer.
I call this dumb. (I've had this debate there.)
We are not here to protect the sanctity of post ownership. We are here to maintain a pool of quality knowledge. And Stack Exchange does note when a post a has been edited, and provides a revision history to see who edited and what they changed.
There are, generally, two situations where massive edits to a post are warranted:
- The post was correct, and has become outdated and is now incorrect or even dangerous. An answer could be posted separately, but the visibility of a highly voted answer renders this ineffective.
- The post was correct, but lacking in details. It's just a line of explanation and a command or two, or maybe just a link pointing to the correct answer. And somebody edits it to add substantial details. In this case, too, the editor could add a new answer, but many of us don't really care about rep, and some would like to show the poster how it's done by improving their post.
For 1, I check the votes and number of answers to the post, and the invalidity of the old answer, and the correctness of the edit, before approving. It's a judgement call on how much visibility I feel the new solution needs, and how problematic the original. Also, I'd rather the original was left untouched, and the new method added to the top or bottom.
- Subpoint to 1, it could be that the answer was very general, and attempted to cover a range of categories involved in the question. The edit aims to keep the post updated with changing times.
- For 2, as long as the actual point of the answer is retained, it's approve all the way.
Both cases require that I have enough expertise in the matter to judge the additions. If I don't, pass. Maybe mention the edit in chat so someone else can judge it.
Somewhat imaginary examples:
- An answer dealing with init systems had only sysv and Upstart. Someone adds in systemd. Approve.
- Someone had posted an excellent answer in a (somewhat) dupe. Since the posts aren't similar enough to be merged, and the parent post has just a barebones answer, they chose to edit in major parts of their answer (screenshots guiding what needed to be done) to the parent post's answer. Approve.
2I'm not sure if that's SO policy either. I think (and hope) it's just the opinion of the users who happened to read and vote on your meta post there. If it is policy, it's absurd and is certainly not policy here.– terdon ModMay 2, 2016 at 11:39