An edit was suggested to an answer for a somewhat old question: Delete UserName from Git repository? (both asked and answered on June 5, 2011).

The suggestion was approved by all the reviewers. The suggestion isn't quite a formatting improvement or updating an outdated link. (An user of git is likely to be aware of --system, --global, --local, --file options.)

Was it appropriate to approve the suggestion?

It is worthwhile to note that neither of the three reviewers (one, two, three) have a positive score for .

enter image description here

  • 10
    Oh dear, 3 robo-approvers all on one suggestion :/ (or so it seems) – Doorknob Oct 27 '13 at 14:05
  • 2
    It would have been better off as a comment for sure (and the user can’t post comments) and it may be useful and it’s kind of inappropriate but it’s also not the most harmful thing in the world and nobody appears to be a total robot, so… oh well. – Ry- Oct 27 '13 at 14:22
  • 3
    @Doorknob Do the reviewers deserve a break? – devnull Oct 27 '13 at 14:31
  • 1
    Or all 3 reviewers checked out the edit, found it to work and have different opinions to us as to how much content can be added to a post – Richard Tingle Oct 27 '13 at 14:39
  • 4
    @minitech all three reviewers have no experience whatsoever in the relevant git tag so they just clicked "Approve" without knowing if it's correct or wrong. Robo or not, they should leave it to those who know. – Shadow9 Oct 27 '13 at 14:42
  • @ShaWizDowArd Correct. Maybe I should have had added that information to the post. – devnull Oct 27 '13 at 14:42
  • @devnull yep, that will be useful. If possible they deserve at least a warning. – Shadow9 Oct 27 '13 at 14:43
  • @ShaWizDowArd Added the info with links. – devnull Oct 27 '13 at 14:48
  • @ShaWizDowArd Tarkus has Github experience – Richard Tingle Oct 27 '13 at 14:52
  • 1
    @RichardTingle but no git at all, so I'm still in the opinion he shouldn't make decisions on such matters. – Shadow9 Oct 27 '13 at 14:57
  • 2
    Steady on. All three might have been using git since nursery school, for all we know. – Michael Petrotta Oct 27 '13 at 15:09
  • 1
    @MichaelPetrotta Indeed, all this talk of "warnings" and "breaks" seems based on a lot of assumptions. Not that I'm wildly comfortable with the recent spate of "calling reviewers out" on meta either. – Richard Tingle Oct 27 '13 at 15:23
  • 1
    @Richard, yeah, I've been watching the recent kerfuffle with... uncertainty, I guess. Bad reviewers are a real problem, and public shaming might help. Witch hunts won't. I have a bunch of experience with localization, but you couldn't tell on SO - I don't seek those questions out to answer because I'm sick of that domain. – Michael Petrotta Oct 27 '13 at 15:34
  • This may help: "When should I make edits to code?" – jmac Oct 27 '13 at 23:24

This is not an appropriate edit. It is adding entirely separate content that the author never added to the post. Such information belong as a comment, or in another answer. Editing is designed to make the existing content of an answer shine, not to add something entirely new that was not there before. (Note that if the answer's author provided such content in comments, moving it from comments to the answer would be fine, but that's not the case here.)


It's fine to add to an answer in small ways that improve it and make it a better, more complete answer for those who find it. The real question is if it's accurate or not, and that is what reviewing an edit like that should be based on.

We actually have a mechanism, that if it happens enough times to one post, it becomes Community Wiki, and the community nature of the post is displayed.

  • 4
    Reviewing suggested edits, like all other moderation, should not be dependent on evaluating the technical correctness of the content being added. – user102937 Oct 27 '13 at 17:24
  • 1
    @Robert, that seems a bit...short-sighted... If an edit is technically correct (for instance, fixing braces or semi-colons in the code of an answer), you cannot effectively evaluate -- as Shog9 said "Learn to love that skip button". In the same vein, accepting an incorrect edit simply because it "looks good" will hurt the quality of the answers on SE as well. Could you please explain your justification for reviewing solely on "feel" rather than understanding of whether or not the content being reviewed is correct? – jmac Oct 27 '13 at 23:23
  • @jmac Who said anything about "feel?" Skipping is the correct thing to do if you don't know whether or not the edit is correct. But if you don't edit other people's code, it's difficult to go wrong. – user102937 Oct 28 '13 at 14:27
  • @Robert, apologies if I misunderstood, but you said "reviewing suggested edits...should not be dependent on evaluating the technical correctness..." If technical correctness is irrelevant, then it implies that it is okay to review without knowing. Can I rephrase your comment as, "If you aren't sure about the technical correctness skip the review, but if the technical correctness is irrelevant to the quality of the edit, no problem"? – jmac Oct 28 '13 at 14:32
  • @jmac: While the community seems to be divided over whether or not code edits are OK (in an answer), they almost universally agree that changing the meaning of someone else's answer is not OK. When you change code, there's always a high risk that you are also changing meaning. One of the general principles of moderation is that you shouldn't need to be a subject matter expert to do it, and the level of technical rigor required by multiple reviewers would seem to preclude these kinds of edits. It is for all these reasons that I encourage comments instead of suggested edits to code. – user102937 Oct 28 '13 at 14:38
  • @jmac: As always, if people have a problem with this, then they should earn enough rep where they can get unilateral editing privileges. – user102937 Oct 28 '13 at 14:44
  • @Robert, perhaps you disagree with the content, but I made a faq-proposed about editing code, if you disagree with the content, please feel free to edit that. – jmac Oct 28 '13 at 14:44
  • @jmac: I've read it; I think it's fine. You don't need a ton of technical knowledge to indent code, nor do I necessarily categorize that as "editing code." Indenting code is more or less the same thing as fixing people's grammatical errors; it doesn't change meaning (unless the language is Python). Note that indentation is not the issue being discussed in this meta post. – user102937 Oct 28 '13 at 14:47
  • @Robert, I am definitely talking about more than indentation in regards to answers there. Questions I say grammar-equivalent only (due to the high risk of fixing the problem the question is about, or otherwise making it more difficult to answer), but for the answers, definitely making them work should be a priority. – jmac Oct 28 '13 at 23:32
  • @jmac: It's the responsibility of the person writing the answer to make sure they are posting code that works, not the community. If the answer has broken code in it, write your own answer that contains the correct code, or post a comment below the broken answer telling the poster what is wrong with their code and how to fix it. – user102937 Oct 29 '13 at 1:57
  • 1
    @Robert, as explained in that question (and connected discussions), creating a virtually identical answer, using the same code logic, but only correcting a misspelled variable name, or a missing bracket, should not be a separate answer -- that is what edits were made for. SE is collaboratively edited for a reason. Feel free to discuss there where it is a better fit. – jmac Oct 29 '13 at 2:41
  • @jmac: This isn't something that I just pulled out of a hat. There is a significant percentage of folks in the user community that feel very strongly that code in an answer is the poster's responsibility and should not be edited by anyone else, unless the poster has marked the answer Community Wiki. I don't care how it happens, just so it doesn't involve changing someone's answer to mean something else. This isn't Github. – user102937 Oct 29 '13 at 2:45
  • let us continue this discussion in chat – jmac Oct 29 '13 at 2:47

You must log in to answer this question.

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