I tried making two different edits to a post. The first was not approved because it "should have been a comment," which while I'm on the fence about I do feel is entirely fair because I added in text pointing towards what I edited (I like to point out potential easy mistakes to avoid in my own answers).

The post in question essentially had the correct answer, but had failed to add a parameter which was required for the way the function call was being used. It was clear what the intent was, and the edit did not change the underlying message in any way.

The second edit, which was literally just adding the correct parameter to the function call to work properly as the original post clearly intended (returning an associative array instead of an object... because the poster was using the return as if it was an associative array) was approved by two and rejected by two for "This edit changes too much in the original post; the original meaning or intent of the post would be lost" and a third for "should be a comment." I took care to explain the revision in the edit summary field.

I'm really not sure how to take that. Are edits only intended for fixing typos and adding tags, because this is barely above the level of being a typo or basic syntax error: it was just a mistake in [assuming|forgetting] the default return was the opposite format, when setting a single parameter would fix it to return as obviously (by the code) expected. Should I assume that the people who reviewed my second edit maybe didn't understand the function call being used enough to realize what was going on, and that the change I made was literally to make the code work as the original post intended? If so should I just submit the same edit over again?

It just seems a little bizarre to me. I could write a second answer with the correct function call usage and attribute the original answer, or comment and hope the original poster noticed and made the proper changes, but to me this seemed like an appropriate case for using the edit feature, even going by the guidelines shown when editing a post. If that's not the case, then I would suggest the guidelines may need revising for better clarity.

I've gone through at least a few of the questions and answers on meta with regards to edits, and found a few where substantive changes which heavily changed the message were addressed--which makes complete sense--but did not see any with at least light searching which involved questions of changes which fixed a small, easily understandable error in the answer so as to correctly fit the message. Personally I feel that's what this was, in the case of the second edit especially. Maybe that's a matter of opinion. After all, I could have suggested changes involving using the object returned by default rather than lightly fixing the call to return the expected array, but those would have been far more substantial code rewrite or not even worked with the methods being evoked on the return. It was a small, easy mistake to make, much like any typo is, and the correction of simply adding the right flag to fit the obviously anticipated return seems equally small when properly taken into context.

To add some context to my question, I would reference the answer to Was this a valid edit that I should have approved?. While the situation is by degrees different (misquoting from a text versus a small code error of leaving off a flag where the intent was clear to use the return as if that flag were set) I would personally tend to feel my edit fits the spirit addressed in the answer, namely

This was almost certainly unintentional, and the edit was in keeping with the intended meaning of the answer.

Granted, that may just be my personal opinion that my own edit (the second one particularly) was done in that same light. If I'm wrong about that then I'm beyond ok with being wrong. It's not a question of me feeling offended (I'm not), it's more that it seems inconsistent (and gawd I hate inconsistencies with a passion, hah) to me.

At the same time, I do by degrees agree with the points @Pekka has made in his answer below and the following comments which expanded on that, with regards to making sure comments illuminate issues properly for the original author. But then I would ask, what is the point of edits notifying the original author of the edit, and what is the point of the edit comment field? I realize that much of this revolves around issues of community norms, but if those norms have so far diverged from the guidelines given, then maybe one or the other needs addressing? However, I feel it would be presumptuous of me to make such a suggestion without more comments in that direction from those better informed on how the community trends and what is considered best practice for SO.

  • I just ran into this as well here and here. I had an obvious typo in the code, and someone tried to fix it, and four people voted "This edit is incorrect or an attempt to reply to or comment on the existing post." or "You are changing the code of the original question, which may change the meaning of the question asked." – Mooing Duck Mar 27 '13 at 20:01

The right thing to do in such a situation is to comment so the answerer is made aware of their error.

  • Maybe your correction is valuable information to them

  • Maybe the author's intent wasn't what you thought after all (not referring to the case you show, just making a general statement)

  • A comment is visible to anyone trying to use the code

by silently editing the code, you are taking away the author's chance to evaluate and respond to your feedback, and incorporate it into their work.

  • I might say that the same applies for language and markup edits ;) (that's a lightly teasing tone btw). I think what I would ask then, is would it not be appropriate to make the edit but also add a comment addressing it? – taswyn Dec 28 '12 at 15:41
  • 1
    @taswyn that would be an option as well, but in the case of code, I guess the general feeling is that anything is too much of an edit. Remember, edit reviewers don't see what comments you may have posted underneath, they have to judge based solely on your edit. The safest and cleanest way is to comment, IMO. Language edits carry much less danger of changing the original meaning so massively it's no longer recognizable – Pëkka Dec 28 '12 at 15:44
  • 1
    I suppose. It just seems very knee-jerkish to throw out code edits entirely out of hand when they do seem to fit the spirit of the editing guidelines... and if it's the general community feel, maybe some comment pointing towards that should be added to the edit guidelines? (obviously that's a bit rhetorical/pointless to ask in a comment, but I am curious on your opinion =) ) – taswyn Dec 28 '12 at 15:47
  • @taswyn That sounds like a fair suggestion to make – Pëkka Dec 28 '12 at 15:48
  • I'm on the verge of simply accepting your answer, Pekka, but would like to leave things open a bit longer. I did find one question which seemed by degrees similar (a code edit dealing with obvious intent) and the accepted answer to that differed from yours. Granted, the situation was also arguably more clear, but the degree of change and clearness of intent seem similar to me at least. I've added that as a section below my original question, and I think the part regarding the edit comment and edit notification versus normal comments is what I'm catching on the most now. – taswyn Dec 28 '12 at 16:08
  • 2
    by silently editing the code, you are taking away the author's chance to evaluate and respond to your feedback, and incorporate it into their work. why do you say that? I'm sure you know that you get notifications on edits and you can @reply to editors. – Some Helpful Commenter Dec 28 '12 at 16:14
  • @Some I for one frequently overlook edit notifications. I noticed maybe one over the past couple of months – Pëkka Dec 28 '12 at 16:16
  • 1
    Pekka, reviewers don't see... comments - well if there are comments that are important for understanding context, skilled editor could point that out in the edit summary. I for one have seen summaries like that in review queue and followed through the link to post to check these. Coincidentally or not, I think I approved all suggested edits summarized that way – gnat Dec 28 '12 at 16:19
  • @gnat that's a fair point. I'm just not fond of editing for this purpose because as said, I never look at my edit notifications, but I will always acknowledge a comment. Whether that's the norm or not, I cannot say though – Pëkka Dec 28 '12 at 17:47

In this particular case the first edit should have been rejected (the non-code change you added really should have been a comment) and the second should probably have been approved.

I say "probably" because it should only have been evaluated by someone familiar with the technical material, and that's not me. Now most suggested edits, almost all in fact, require little or know knowledge of the technical material to approve/reject, but this is one of the few exceptions.

In this case the clear point was that the intent of the original author was for the code to be as you suggested, and they made a simple typo and forgot a parameter that they should have used. You weren't changing the meaning of the code, it's intent, or the message it was intended to convey. There were indicators within the code pointing to the fact that this is what should have been done.

Most importantly, you mentioned all of the above in the revision comment. That's very key. Without that comment I'd have probably rejected it as well. The comment was specifically saying that, to someone with knowledge of the tag, it should be obvious that the intent of the code isn't changing and that a simple typo is being fixed. Now, I don't have knowledge of the subject material, so barring me going and doing a bunch of research, I know enough to know that I shouldn't touch it. If you're assertions about the code are correct is should be approved, if they aren't it should be rejected.

If three people with knowledge of the material feel that it is an appropriate change and approve it, but the author feels that all of you missed some key point, he still has the option of rolling back the edit (since he'll be notified of the edited post). If an edit such as this is rolled back by the OP, even if you're 100% sure you're fixing a mistake, you should still leave the post alone and either comment or post another answer. If you find it being rejected or rolled back by people other than the OP, then the right thing to do is to post on meta, as you are doing, rather than starting an edit war.

  • Thank you for clarifying with the final paragraph, @Servy, as I was just about to ask what the appropriate course is then when a non author has rejected an edit which to the best of the editor's own personal understanding and views seems to meet all of the appropriate points. – taswyn Dec 28 '12 at 16:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .