8
\$\begingroup\$

I'm not really sure whether this question where OP has added a lot of feedback about one of the reviews actually qualifies for a rollback.

The code is not changed and OP also hasn't added any improved version of their code so it does not match any of the standard reasons for an undo listed in What should I not do? but it somehow still feels wrong to put the comments there.

What would you do in this case and what do we do next time?

\$\endgroup\$
  • 3
    \$\begingroup\$ IMO it's the same as updating the code. There's comment sections for responses, and you can post a new question with improved code/knowledge. It's also riddled with 'thanks' and other too chatty comments. \$\endgroup\$ – Peilonrayz Mar 7 at 19:56
  • 1
    \$\begingroup\$ As it has reached the Hot Network Question list, I have rolled it back for now until we've finished discussed it here. \$\endgroup\$ – Simon Forsberg Mar 7 at 20:09
  • 1
    \$\begingroup\$ Rule of thumb: the moment an answer arrives, you stop editing the code and you don't respond to answers. Answering answers gets messy, really fast. That's what follow-up questions and comments are for. \$\endgroup\$ – Mast Mar 8 at 14:37
13
\$\begingroup\$

The state of the question as seen in https://codereview.stackexchange.com/revisions/214917/6 (before my rollback) literally says "In response to Peter Taylor's answer".

It makes the question confusing, as you can no longer read from top to bottom. You instead read: Question, response to answer, answer. The response to the answer would work better in comments to the answer, and - in case of lengthy comments - even better in chat.

This particular question also became one of the Hot Network Questions, and has had 311 views in around 7 hours. How do we want to show our questions on Code Review, do we want to encourage editing replies to answers into the question? No thanks.

The OP responded to Peter Taylor's answer by editing the question, how would Peter Taylor respond to the OP's response? First of all, Peter Taylor is not notified by the edit to the question so he might not even be aware of the response. Then if he would respond... would he....

  • Edit his answer? That would make a mess when reading the question as you don't know which is response to what and when.
  • Edit the question? THIS IS NOT A WIKI!
  • Add a comment to the question? Again, confusing.
  • Add a comment to his answer? The OP would not be notified by this.
  • Edit his answer? Yeah that would probably be the best thing, but then that would invalidate the edit to the question.

The whole thing would be very messy so the best thing to do is to rollback.


The auto-comment that we frequently use for answer-invalidations:

Please do not update the code in your question to incorporate feedback from answers, doing so goes against the Question + Answer style of Code Review. This is not a forum where you should keep the most updated version in your question. what you may and may not do after receiving answers

Should probably be updated to:

Please do not update your question in response to answers, doing so goes against the Question + Answer style of Code Review. This is not a forum where you should keep the most updated version in your question. Please see what you may and may not do after receiving answers

\$\endgroup\$
  • 4
    \$\begingroup\$ Updating the question with clarifications is often OK. Updating the question to add rebuttals is not OK, I think. \$\endgroup\$ – 200_success Mar 8 at 0:31
  • \$\begingroup\$ At first, I was a bit hurt about the rollback, but I agree with the overall reasoning of your points here. Interesting case! I'll restrict extended responses to chat in future, or at least keep responses more terse. \$\endgroup\$ – Dan W Mar 8 at 2:47
  • 2
    \$\begingroup\$ When the question no longer reads as a code-review request, that's a bat edit. Like other SE sites, what we're building is an archive of Q+A, not back-and-forth discussion. \$\endgroup\$ – Toby Speight Mar 8 at 9:50
  • 2
    \$\begingroup\$ @DanW Comments on answers and questions are usually fine and is a good place to discuss such things, but they might be deleted after a while (and possibly converted to chat if it's a lengthy discussion). Most importantly though, I'm happy that you asked a question and even happier that you got a good answer to it :) \$\endgroup\$ – Simon Forsberg Mar 8 at 9:55

You must log in to answer this question.

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