What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 127 Stack Exchange communities.

I often see comments that relate to something in the question or answer that is subsequently changed in light of the comment. This can make the comment appear odd to users who are looking at a revision of the question/answer and not the original that the comment related to. For example, a comment might point out "You wrote foo when I think you meant to put bar" and the author then corrects it. The question/answer no longer contains a reference to foo yet the comment remains and users may take a while before they realise that foo must have been removed.

How about some kind of mechanism to identify a comment with a particular revision so you can tell at a glance that it may be out of date? This could either take the form of a hard link between the comment and the revision or simply some form of shading/colour coding/other identifier to indicate that the comment relates to a previous version without necessarily indicating which one.

Assuming you don't already have a relationship between a comment and a revision logged behind the scenes, it ought to be possible to work it out for existing comments based on timestamps.

share|improve this question
add comment

3 Answers

The comments already have timestamps which will tell you how far back they go. It is possible to take that timestamp and place it to a particular revision.

If a comment says "4 days ago" I am going to look back at the revision history and see what revision(s) "4 days ago" is referring to in terms of a particular revision. For items that are more recent you can be even more accurate.

Trying to add a hard link for every comment just to point back to a particular revision wouldn't make a whole lot of sense, especially when you take into account the fact that answers don't even link to a particular revision.

share|improve this answer
    
Sure, I'm just trying to make the fact a bit more prominent to save having to crosscheck every single comment manually. –  Luke Bennett Oct 15 '09 at 13:04
1  
Why is it so important? If I see someone who makes a comment and it makes no sense in relation to the post and then I see that the post has been revised, I'm probably just going to shrug and not think about it more. I don't know why you need to apply so much importance to the comments when there isn't even a way (like you are requesting) to trace a particular answer back to a particular revision of a question (which is probably MUCH more important because a mismatch of an older answer and a newer question can lead to people downvoting the once-accurate answer. –  TheTXI Oct 15 '09 at 13:08
    
I didn't call it important. It's just one of those minor usability enhancements that can provide a benefit without being intrusive. The point is not that there is currently no way to figure it out, but that it's not immediately obvious. New users in particular won't necessarily realise that the q/a is newer than the comment. They may not think to check the dates - it's not even the first thing I think of as someone who uses the site a lot. Why does SO provide alerts via the top banner? I can find out for myself by looking in the right place. Point is, it makes life easier and clearer. –  Luke Bennett Oct 15 '09 at 13:28
    
I do agree by the way with the need to link answers to revisions; I proposed this question having just spent some time trying to unravel the revision history of a couple of SO answers and their comments so this scenario was freshest in my mind. I would fully advocate the same mechanism for answers -> question revisions as well. –  Luke Bennett Oct 15 '09 at 13:28
1  
I think it's somewhat important. The fact is, most people aren't going to take the time to cross reference revision histories before responding to an answer. I know I've downvoted an answer or two without realizing the question was changed. And then, if I don't realize before 10 minutes are up I can't do anything to correct it. It seems like some indication to make it obvious that the answer may be outdated would be helpful. –  Cogwheel Jun 17 '10 at 16:19
add comment

When I comment, I create my own context. I answer as if the answer weren't even there because it could change.

For example, I will not say,

Your answer is wrong because ...

I write:

It's best practice to do this ...

Or another example:

Your query would return duplicate results because you're using UNION ALL.

Should be:

UNION ALL returns duplicate results. Use UNION if you don't want duplicates.

The comment will always be true, regardless of the answer.

If they modify their answer so that it agrees with my comment, my comment will simply appear as consensus, otherwise, it will appear as a caveat.

This has the added benefit that the comment is less likely to be perceived as a personal attack. Instead, I am merely adding value to the answer.

Of course, we should do the same for comments on questions.

If everyone did this, the situation you describe wouldn't be as much of an issue, though it might still fare better with some cleanup.

share|improve this answer
add comment

On such occasions as pointing out minor things that are later rendered obsolete/updated in the post, wait for the commentor to delete it and keep things clean for everyone, or kick down that point in there being no trouble to flag it.

flag this comments as noise, offensive or spam

Don't really care for comments being made that much more feature rich. If they're that good, they should be answers in their own right.

Unless they're off-topic picks at the Pundit badge. In which case, its comedy or salience really doesn't need to be that heavily tied down to a revision point in the answer's history.

share|improve this answer
    
Removing noisy comments is a fair point. However, oftentimes the comment maintains some kind of value that you don't want to lose eg: "Never invoke Foo like that, it'll hammer your CPU." –  Luke Bennett Oct 15 '09 at 13:31
    
Comments like that foo hammer don't really need to be tied down to a revision point though. Last sentence of this answer agrees. –  random Oct 15 '09 at 13:41
    
But what if the author has updated their answer to invoke Foo correctly? To the casual observer it will now appear like the correct implementation of Foo is being criticised. –  Luke Bennett Oct 15 '09 at 13:44
    
If it was someone with high rep (who could have edited it better), that would be a lazy comment. But if the originator of the comment now lets it stand, thinking ahead that their critique might be acted upon, it does make them look silly. –  random Oct 15 '09 at 14:11
    
Assuming they're aware the author made the changes... I don't have time to go back and review all my comments on a regular basis so I have no idea whether they're all still contextually applicable or not. –  Luke Bennett Oct 16 '09 at 8:04
add comment

You must log in to answer this question.

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