81

Having recently reviewed a few suggested edits, I've found that not being able to see the comments on the post is hindering my understanding of the context of some edits.

In particular, there have been a few posts where someone has made a substantial edit to a question or answer which significantly alters its meaning. Typically I'd reject that as "too radical", since it can't be known that the new meaning of the post was what the original author intended.

However, in some cases it is clear from the comments that the original author is aware of the intended change, and as such the editor is not in fact misrepresenting the author's intentions. In these cases, it would have been useful to see the comments when making the review decision.

As an example, a common cause of this is when an inexperienced user is asked to clarify their question, and ends up writing a comment that would be better off as an edit to the question (due to length, excess code content, etc), presumably due to lack of awareness that the question can be edited. Often, other users will helpfully edit the question to include the information the OP put in the comment (see here, here and here for examples). In this situation, it is clear that the edit reflects the OP's intentions, but this information is only present in the comments which are not visible in the review interface.

Following is the suggested edit screen showing the suggested edit. It seems indentation is increased and one line is moved a few lines upwards and some empty lines added to increase readability. But it look like an invalid edit since changing the related line may cause alteration on the runtime. So not looks lie a valid edit. Better open the answer and examine it

Checking the Q&A itself reveals that a talk had taken place between the OP and the answerer and revision done by the OP with the permission of the answerer. So it was indeed a valid edit.

In short: is there any particular reason comments are not displayed in the suggested edit review interface? I think it would add useful contextual information when making review decisions, and presumably this would be trivial to implement.

Note: this is not a dupe of this post, as it does not concern comments posted after the suggested edit.

11
  • 3
    @Bart: I don't disagree - in an ideal world, the original author would make the edit. It's not an ideal world though, and the situation I describe does happen. Even if we discount this particular circumstance, I still see no reason why comments can't be included in the review tool, and think they would be of benefit if they are included. It's about making the most informed decision possible, which can't be the case when comments are absent.
    – Mac
    Oct 11, 2012 at 23:27
  • 3
    Yep, other review queues already show comments so it will be consistent with existing behavior. Feb 25, 2013 at 8:24
  • @ShaWizDowArd: thanks, if that by itself is not a good enough reason then I don't know what is. Thanks for the bounty too: I was hoping my edit yesterday would get some attention (it has), but a bounty certainly will help!
    – Mac
    Feb 25, 2013 at 19:42
  • 1
    Cheers Mac, my pleasure and here on Meta I'm using my rep to try attract attention to what I think deserves it. :) Feb 25, 2013 at 20:58
  • This would definitely help sometimes, when users edit from the comments.
    – hjpotter92
    Feb 26, 2013 at 23:46
  • Possible solution: Allow linking (suggested) edits to one comment Aug 22, 2013 at 8:25
  • Another helpful related suggestion: Editing: add a “require author approval” checkbox Aug 22, 2013 at 8:35
  • 3
    The edit guidelines state "To include additional information only found in comments, so all of the information relevant to the post is contained in one place" as a valid reason, rather silly given comments aren't displayed in reviews (at least not on answers).
    – Jason C
    Nov 8, 2013 at 18:08
  • As an alternative approach, review shown with comments as hidden with a clickable show comments link. When clicked, hidden div became visible and shows comments. So, comments will fill the screen as the reviewer displayed the suggested edit and will be reachable as easy as possible.
    – Mp0int
    Jan 28, 2014 at 10:17
  • 1
    +1 This would be particularly useful on CodeGolf.SE, where we tend to reject edits which improve another submission, unless it's clear from the comments that the author consents to this. But I think in general there will always be edits whose validity depends on some context given in the comments. Jan 23, 2015 at 23:09
  • It seems this got silently status-completed?
    – starball
    Dec 9, 2022 at 21:29

5 Answers 5

13
+100

(I was going to ask this as a question, but found this dupe while I was typing, so I'll add it here as an answer instead.)

I rejected this suggested edit as "Too radical" because, on the face of it, the editor has interpolated their own information; they've added data that is nowhere in the original question. The reviewer comment is "formatting for command line, add referenced properties, fixed spelling".

I was a little too hasty, though, because if you look at the actual question, you can see that for some reason the OP provided this information in a comment to their own question. That's what the editor meant by "add referenced properties", I suppose, but I missed that.

So I'm here to add my vote for this feature: can we please see the comments on an edited question or answer, to help us understand the context of an edit? I realize yeah, I can get to the info if I want, but sometimes, like this time, it's just not obvious that there's a reason to do so.

2
  • 6
    Yep, what you describe is exactly the kind of situation that prompted me to ask this question in the first place.
    – Mac
    Jul 19, 2013 at 0:41
  • 1
    "I have this problem, too!" Aug 16, 2013 at 8:40
11
+75

My approach to this is to reject edits that add material or change code UNLESS the edit revision comment says something like "incorporated background from comment by question-asker" or "adding code from answer by question-asker" or otherwise makes it clear to me that there is more going on here than meets the eye. Then I will right-click-opn-in-new-tab to take a look at the question in context, with comments, other answers etc.

If you react to all additions of material with a trip to the original question you're a better person than me. And maybe you're encouraging edit-suggesters not to explain themselves properly? The revision comment is there for a reason...

11
  • 7
    I do exactly the same. I really feel I shouldn't need to open another tab just to see the comments though. And, clearly, not everyone does this - the second edit I linked to in my question was rejected as "too radical", presumably because the reviewers didn't bother to do this. I suspect it's just too much bother for most, and having comments available would help address that.
    – Mac
    Feb 27, 2013 at 0:48
  • I would approve that because the comment is Added manifest from comments for OP - don't know why the others rejected it Feb 27, 2013 at 1:03
  • 3
    I see three reasons why the other reviewers might have rejected the edit: (a) they didn't bother to read the edit comment, and decided it was too radical based only on seeing a mass of added code, (b) they saw the edit comment, but decided the edit was too radical regardless, or (c) saw the edit comment, but couldn't be bothered opening another tab to check the comments to verify. The point of this question is to address scenario (c), and reduce the likelihood of scenario (a). (Scenario (b) is more a difference in review philosophy, which is a whole different discussion.)
    – Mac
    Feb 27, 2013 at 2:30
  • Big -1, sorry. I personally do not see how adding useless noise is any help. You are supposed to do quality review or none when deciding to act on the suggested edit, and not skip. If you do not take the time to do it properly, just skip it. It will be better for everyone. Otherwise, you might reject a completely valid bug fix. You are not obliged to review things that you do not have time for. May 19, 2014 at 7:55
  • 2
    It is the task of the editor to provide a case for the edit. That's why the edit summary is there. Use it to present a clear reasoning for your edit. If you don't, you risk a rejection.
    – Bart
    May 19, 2014 at 7:56
  • @Bart: the reasoning is the whole thread and the deep knowledge in the area. Even if you could (which you cannot due to the size limit) repaste, it would be silly. The url to the thread exists for a reason. It is pretty destructive behavior to reject a valid fix to the answer because you are careless to understand the thread and make a responsible decision. Just do a favour for the quality of the site, and skip the suggested edit. Let the experts handle it who have the knowledge and time, i.e. can make a good job as opposed to you. May 19, 2014 at 7:58
  • @LaszloPapp suggested edits are not for bug fixes. They are for grammar, spelling, bringing images inline, typos, and changes made as comments that should have been edits. A review comment is never useless noise, and one that explains serious additions or changes by referring to comments left by the OP is definitely not useless noise. Why do you think that field even exists? May 19, 2014 at 12:40
  • @Kate: OK, it will be easier if we discuss a concrete issue where the "reviewers" rejected a valid bug fix. I find it destructive behavior. Looking at the "reviewers", they have no track record in the given tag. One "reviewer" even specified "Too minor". The user then submitted the same and the same (gold-badged) "reviewer" approved that time. I think that comment is self-explanatory, and if one does not have the domain expertise, s/he should learn the whole thread to grasp the idea anyhow. I realized this, but it could sneak on.. May 19, 2014 at 12:49
  • this is an important distinction that you have not been told. Reviewers should not need subject matter expertise. If someone wants to suggest an edit that requires knowing Perl or C++ or jQuery to assess, it should not be a suggested edit. Changes to the correctness of an answer are bad suggested edits. The suggester should write their own answer instead, that can be voted on. May 19, 2014 at 12:59
  • @Kate: it would be bad to duplicate a long and almost perfect answer. If something changes, it will have to be changed at two places, sync'ing up, etc. I do not think this is good practice. I consider bug fixes (erratas) even more important than some minor grammar issues as bugs can leave an otherwise perfect answer broken. That is not the case with minor grammar issues. Also, note that how that answer is rejected as "too minor", etc without giving proper explanation to the OP. That is another issue therein. By the way, if you reply to me, please use my name so I can get notified. May 19, 2014 at 15:19
  • 1
    @Bart, I agree that the editor needs to justify their suggested edit, but often making this case in the edit comment is not sufficient. In the situation Raff linked to in the bounty description, I used "Incorporated important information from OP's comments into his answer itself" as the edit comment. Despite that, my edit was rejected because the reviewers didn't realize that the new material had come from the OP's own comments. I'm new to SE, but from my experience today it seems that many reviewers may not be as thorough as you and Kate. Dec 2, 2014 at 22:36
8

The editing guidelines at https://stackoverflow.com/help/editing specifically state:

Common reasons for edits inlclude ... To include additional information only found in comments, so all of the information relevant to the post is contained in one place.

This directly conflicts with the suggested edit review UI, which does not show comments for answers (although it shows them for questions).

As a consequence, I generally see all such edits to answers rejected as too-radical or attempt-to-comment. The review UI quite handily directs all reviewers away from following that guideline, and comment display needs to be added to answer reviews if you truly want the community to follow current editing guidelines.

33
  • 2
    "you generally see all such edits to answers rejected as too-radical or attempt-to-comment" <citation needed> A proper revision note solves this issue reasonably effectively.
    – Servy
    Nov 8, 2013 at 18:09
  • 1
    I would still be time consuming for the reviewer to check. Clicking the link to load the actual question is not what's going to be time consuming; reading through all of the comments is what's going to be time consuming. I don't see the need to reiterate all of the other points I made to your other question here.
    – Servy
    Nov 8, 2013 at 18:39
  • 1
    Think, too: A user who makes an edit like this where it's rejected as too-radical isn't likely to actually associate that with the fact that they provided a poor note -- They probably won't go back and resubmit with a better note. Instead, I believe they'd more likely see the reason was "too radical" and simply come to believe that edits like that are generally unacceptable, and stop making them. Then, in a larger sense, that hurts the community as good edits are lost and new < 2k users get accidental negative reinforcement against merging comments into answers.
    – Jason C
    Nov 8, 2013 at 18:39
  • 1
    @Servy It will only be time consuming on questions with lots of comments. If the comment being merged has upvotes and is truly worth merging, it will probably be visible. There is a need to reiterate, actually. For the benefit of anybody reading this that has no context from that other discussion, which ultimately will be the majority of people reading this. If you weren't prepared to provide context here you shouldn't have revived the conversation here. :)
    – Jason C
    Nov 8, 2013 at 18:40
  • 1
    If an editor doesn't suggest edits appropriately, has them rejected, and as a result decides not to make such edits, then that's not the end of the world. Ideally they'd take a look at the editing guidelines to understand what they should/shouldn't do, but at the end of the day I'm sure they'll be able to find plenty of ways to contribute positively to the site, even if editing comments into questions isn't one of them.
    – Servy
    Nov 8, 2013 at 18:43
  • 1
    So to be clear, you are against adding comments to answers in suggested edits (recall, that is the topic here) because you believe that it's "not the end of the world" if a reviewer doesn't have the info they need to make a proper review? Well, that's just you I guess. Also one editor having their edit rejected, sure, no big deal. Having that happen consistently introduces a systemic, broader issue (we are parts of a whole). I hope you can see that. A community is collectively defined by individual behaviors, and common, reinforced individual values become larger community values.
    – Jason C
    Nov 8, 2013 at 18:47
  • 1
    No, that's not what I said at all. I said that if an editor isn't willing to make proper edits, then having those edits be rejected, rather than having reviewers go out of their way to try to compensate for the failings of an editor is worthwhile. If the editor wants to have these types of edits approved they should be making a proper revision notation (something they should be doing in general, but to which this case is particularly important). If an editor is unwilling to take the appropriate actions when submitting those edits, then they're better of just not making them.
    – Servy
    Nov 8, 2013 at 18:50
  • 1
    When I said "adding comments to answers in suggested edits" I intended to convey "showing comments on answers in the review interface". It may have been a poor word choice, I apologize for that. I meant "adding" as in the developers would add them to the UI. I am referring to the topic of this question.
    – Jason C
    Nov 8, 2013 at 19:07
  • 1
    Take it as a response to the answer, not your last comment. "What would this proposal (displaying comments in the review queue), if implemented, change?" I agree it comes out a bit weird in context.
    – Servy
    Nov 8, 2013 at 19:23
  • 1
    Well, for that, see the OP. The OP is a request to show comments in the review interface, it lists benefits and reasons, all of which I agree with, personally. My answer agrees with the OPs request. I endorse it and present an additional justification for it (the edit guidelines).
    – Jason C
    Nov 8, 2013 at 19:25
  • 1
    It's a matter if convenience I suppose, and also depends on your review habits. The review UI shows question comments, which I personally find very useful. With the way I use the review feature, I would find them useful on answers as well.
    – Jason C
    Nov 8, 2013 at 19:42
  • 1
    In all general cases described in the OP. As I said, I agree with the OP fully. I also agree with comments on the OP regarding consistency. I'm not going to copy and paste the OPs text here because you are unwilling to scroll up. I have no further reasons to add to the OPs clear reasons, other than the one in my answer, that being guideline consistency. If you'd like to comment on the reasons for the OP, then post in the ongoing discussion in the comments on the question above. You will get higher visibility and more varied responses from people who are not reading this particular answer.
    – Jason C
    Nov 8, 2013 at 20:56
  • 1
    If you are unable to extend the OPs general case beyond the specific example, there's not much I can do to clear it up, as other reasons are nuanced and case-by-case. If you would not find this feature useful because it does not help you in your use of reviews, that's fine, but "it shouldn't be done because I wouldn't use it" is short sighted and close minded. If you would find harm in it, sure. Otherwise you're just being contrarian. I obviously can't convince you to change your review style, nor would I want to. I will not take additional time to clarify the OP for Servy's benefit.
    – Jason C
    Nov 8, 2013 at 21:06
  • 1
    In other words: The usefulness is clear to the type of reviewer who would benefit from this feature. If the usefulness is not clear to you, your style would never benefit from this feature no matter how many examples I give. So what do I gain from spending time doing that? Any reader of this question who would gain use from this already understands the usefulness without me giving you an outline, which is generally too large to fit in these comments anyways.
    – Jason C
    Nov 8, 2013 at 21:10
  • 1
    "We?" Are you a site developer that would be responsible for the change? If so I apologize sincerely. If not then go review suggested edits, and observe ones marked as too radical. Once you work through that for a few days, come back if you still need justification.
    – Jason C
    Nov 8, 2013 at 21:15
3

I fully endorse this request and would like to take it one step further:

Add the option to "suggest* edit based on this comment" next to comments, which would specifically highlight that comment below the suggested edit, such that edits that otherwise would be deemed "too radical" at first sight can be put into a better context. On approval, said comment would also become linked-to by the revision comment.

Since it makes sense to upvote such a comment due to it providing information justifying an edit, that link could replace the flagging-flag (which is anyway hidden after an upvote) with the edit-pen that indicate edited comments, like so:

mockup of this functionality

(mockup based on a true story, coming to cinemas never)

edit in fact, for the editor's convenience, the clicked-on base-comment might also be added to the bottom of the post and preselected (e.g. for one-key deletion or conversion to a quote)


* The "suggest" would not be shown for users with the required edit privilege (<2k rep or OP), yet this option would allow to link an edit to a helpful comment

2
  • The reason I propose having to upvote a comment first is a) it is relevant (though it might be flagged obsolete after the edit got approved) and b) one-line comments would take to much space if the pen were also shown in addition to the other two icons... Aug 16, 2013 at 8:44
  • turned into feature-request: meta.stackexchange.com/questions/194277/… Aug 22, 2013 at 8:24

You must log in to answer this question.

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