Obsolete comments: To flag, or not to flag?

Question

^{↑ That ↑ is the question...(Sorry, couldn't resist.)}

I've flagged a couple comments as obsolete here, and had both flags declined. These comments were suggestions to correct errors in OPs, which were then fixed in editing. Specifically:

1. The first comment here. "R²" was changed to "adjusted R²" in response to a corrective comment:

   i think you might be confusing {R-sq} and {adjusted R-sq} here.

2. The last comment here. The clarification was copied into the body of the question, as suggested:

   Please edit your question to elaborate it (not in a comment)

I inferred from this that CV doesn't want to delete such corrective comments even after corrections have been made, but I can't find actual discussion of any such unwritten rule, so here's a place for it. If I've inferred too much from two declined flags, I'll be happy to resume trying my luck on new cases as they arise...but I see these sorts of comments hanging around on quite a lot of old questions too.

^{Edit: For the record, I just flagged a third comment as obsolete (a request to edit in data, which was done), and had it marked helpful. This weakens my statistical evidence against the null hypothesis of no difference in policies between CV and SE in general. (Of course, it was extremely weak in the first place.)}

I'm fine with leaving corrective comments regarding resolved errors if that's how we want to do it here, but that would appear to be idiosyncratic policy, and should have an on-record decision here on the meta-site if we want to go against the network-wide grain. In personal experience, I have flagged comments as obsolete on Cognitive Sciences and had them removed. On Math Educators, I found a useful meta-question (that actually inspired this one): "Do we want to delete “obsolete” comments on this site?" Its accepted answer ("Yes.") also links to some useful questions on Meta Stack Exchange:

• Should moderators delete obsolete and resolved comment threads? (Some actual debate here.)
• Is it OK to casually flag comments as obsolete? (Pretty one-sided here: "Flag away!")

Should CV have its own policy? If so, what should local policy be on comments like those I've linked?
If they're not obsolete enough to delete, what clearly would be? And where should we draw the line?

Answer 1

Personally, I favor deleting obsolete comments, and I think this should include the comments I flagged.

• I see some value in preserving comments like the first one I linked for educational or cautionary purposes, but I think it's telling that the Math Educators community doesn't find enough of this value in keeping corrective comments around once the errors in question have been edited out.

  • There are much better places to look for substantive education on subtle differences for general purposes: e.g., How to choose between the different Adjusted $R^2$ formulas?
• Comments on matters of policy or general advice like the second one I linked shouldn't be left as pointers for other users unless they link to relevant meta-questions. We should probably have a new user advisory meta-post about making clarifications via edits instead of comments (if we don't already), as this is a common faux pas, but it seems naive to imagine isolated comments having any effect on the incidence of this behavior. Maybe the same would be true of a meta-Q. Still, I don't think users should inadvertently volunteer their posts as hosts for policy announcements by making initial mistakes that they then correct...At most these posts should host comments as signposts pointing to independent policy statements. Actual meta-questions can at least be edited, downvoted, responded to separately, and protected better from deletion.

• I agree that these comments are somewhat distracting. I also suspect they may make a question or answer look worse than it is at first glance after its original blemishes have been removed.

  • An author who never makes errors initially might be more credible than an author more like myself, and that info might be contained in obsolete comments to some very limited, unreliable extent, but not enough to justify tempting any readers to judge credibility by this criterion instead of more valid ones.

These points being made, I don't yet see a need for a CV-specific policy on obsolete comments, but I would welcome any suggestions, counterpoints, or competing answers.