Maybe it's just me, but it's quite annoying that I can't edit my comments.

If I reread what I wrote and see a typo, a mistake or something else that need fixing, I delete the comment and rewrite it, but this means new notification to everyone related and a mess in the order of comments and the timestamps.

Shouldn't comments be editable for longer than 5 minutes?


2 Answers 2


Comments are meant to be second rate citizens on the StackExchange network. They are not indexed or searchable and they don't bump questions to the front page. It is usually assumed that most comments are temporary, to be deleted once the relevant discussion has been integrated into the question/answer. The site is not intended for discussion, unfortunately. This is an especially unreasonable restriction on metas.

On cstheory, we are much more permissive (or even encouraging of comments), but we cannot overrule all the global features/bugs of the StackExchange network. It is unlikely that the SE overlords would be inclined to change the comment editing rules.

  • $\begingroup$ Thanks @Artem. That's unfortunate :/. I think many questions here are much more likely to be solved via discussion. $\endgroup$
    – R B
    Mar 23, 2014 at 20:21
  • $\begingroup$ @RB I think something is wrong then: if that is the case, an answer should be posted from the comments, or the relevant information that is in the comments should be moved to the question/answer. $\endgroup$
    – Juho
    Mar 29, 2014 at 11:50

As Artem Kaznatcheev notes, comments on Stack Exchange sites are really designed to be an ephemeral "scratchpad" for improving questions and answers. They provide very limited markup options, they're not searchable (using the internal SE search; Google and other external search engines do index some of them), and they're subject to deletion at any time for any or no reason.

The idea is that, once the question has been solved through discussion in the comments, the solution should be written up as an answer (or incorporated into an existing answer), crediting the contributors to the discussion as needed. At that point, the "scratchpad" comments are no longer needed, and may be deleted.

No, it doesn't always work that way in practice, but that's the way it's designed to work.

One thing you can do, to try and get a little closer to the ideal, is to occasionally go over your old answers and see if any of them have comments (written by you or someone else) that add value to the answer. If you spot any, edit the answer to incorporate the information directly into it (with appropriate credit, as needed). This ensures that, even if the comments are later deleted, the useful information in them will not be lost.

You can also go over your old comments (available in your user profile, under "activity" → "comments") and see if any of them might be usefully turned into edits or stand-alone answers. Also, if any questions you've asked have been effectively answered via comments, you can either summarize the conclusions into a self-answer (again, with appropriate credit, of course) or just ask the user who posted the most useful comment(s) to turn them into an answer.

In some cases, once the comments have been turned or edited into answers, you may also want to delete them (if they're yours) or flag them as obsolete (if not) immediately. However, that's completely optional.

  • $\begingroup$ Very nice ... but given that the policy regarding editing discourages editing immediately, and given the fact that waht is not done now never gets done, where do we go? $\endgroup$
    – babou
    Mar 25, 2014 at 17:42
  • $\begingroup$ @babou: I'm not sure what specific policy you're referring to, but, ideally, you should edit your answer once the discussion has reached a useful conclusion -- no sooner, and no later. Of course, this can be problematic in cases where small useful tidbits just gradually "dribble in", and the discussion just dies out once nobody has anything further to say. That's why I suggested occasionally going over your old answers and seeing if there might be something in the comments worth summarizing into the answer itself. $\endgroup$ Mar 25, 2014 at 17:51
  • 1
    $\begingroup$ Your criterion sounds very similar to the halting problem for Turing machines. I do go accasionally over old answers but ... I do not get the feeling that going over old answers is much rewarded in this system. I get that feeling from answering old questions. The fact is that early visitors have more voting power than late visitors. Hence improving old text is a matter of self satisfaction and possibly unrewarded citizenship. Considering further that editing the text will turn it into community wiki, I think there are more discincentives than incentives. Not enough respect for contributors. $\endgroup$
    – babou
    Mar 25, 2014 at 18:33
  • $\begingroup$ @babou: Editing old answers can actually be highly rewarding (at least as measured in rep), since it bumps the question to the front page. If the answer is any good (or if your edit just improved it a lot), that's likely to result in more upvotes. In fact, combating this strong incentive to bump old threads is exactly why excessively edited posts are automatically made community wiki, and why too frequent editing of your posts is frowned upon in by community norms. However, as long as you don't keep making dozens of tiny edits to your answer to repeatedly bump it, you should be fine. $\endgroup$ Mar 25, 2014 at 18:40
  • $\begingroup$ Ps. Note that multiple edits made within a span of 5 minutes only count as one edit for purposes of bumping and auto-CW-ification, so if you make an edit and realize that you want to amend it, it's generally better to do it immediately than to wait a few minutes. Also, if you find your post turned into Community Wiki, even though you don't think you edited it excessively, you can flag it for mod attention and ask them to un-CW it. (FWIW, that's never happened to me so far on any SE site, even though I feel I often edit my answers a lot without specifically counting edits or anything.) $\endgroup$ Mar 25, 2014 at 18:43
  • $\begingroup$ Two points in your answer. Easy first. While not allow unlimited no-bump editing. I would do a lot more because I hate errors or even typos in what I write. Actually, not knowing about the bumping, I used to edit a lot, and learned that my contribution were no longer considered mine. I was not trying to cheat and the moderators knew it, but they refused to give me back my answers, which they could have. I was a very new user, not even knowing how to flag. I got the message a bit too strongly. Not enough respect for contributors, as I said. I was only trying to improve for others. $\endgroup$
    – babou
    Mar 25, 2014 at 18:56
  • $\begingroup$ Regarding the rep for improved answers, I do not much believe it. There are some questions on which I am a good specialist. They had lousy answers, pedestrian or downright ignorant. I tried to give a good answer, state of the art. It was pretty much ignored (on three different SE sites). Maybe my preferred topics interest few people. But then, that's me, and I suppose it can apply to other answers I may give. Oh well, I may get a rep increase every other month for it. I am my own real incentive, ... when not bullied. $\endgroup$
    – babou
    Mar 25, 2014 at 18:58
  • $\begingroup$ I use the five minutes span whenever I have the opportunity. I do occasionally get caught, because I finish editing too late. It should at least stop the clock. I do not understand "auto-CW-ification". Lots of thing could be improved on SE. $\endgroup$
    – babou
    Mar 25, 2014 at 19:02
  • 1
    $\begingroup$ @babou: I agree that it's somewhat unfortunate that the first warning to a new user that they shouldn't edit their own posts quite so much may be seeing the post turn into CW. If you try to make a very small edit (e.g. fixing a single typo) to someone else's post as a new user, the system does warn you that "edits should be substantive" and refuses to accept it, but there's no such warning when editing your own post. You might want to suggest such a feature, or, better yet, vote up this existing suggestion. $\endgroup$ Mar 25, 2014 at 19:23
  • $\begingroup$ I did .., and answered too. $\endgroup$
    – babou
    Mar 25, 2014 at 23:19

You must log in to answer this question.

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