17

Actually this time I do not have a question. Instead, I would like to discuss why I do not agree with the persons who declined my flags.

I have flagged this and this post as NAA. Both flags have been declined with the following reason:

The question asks for external resources. This answer provides a link to an external resource. Thus, it's an answer. The problem here is the (now closed) question, not the answer.

No. In both cases, it is NOT an answer. The first one is by all means a link-only answer. If you bother to answer an off-topic question, at least put some value to it. As for the second one - it is a plain thank-you/me-too post.

While on the topic, I have also recently flagged this post as NAA. The result was:

declined - flags should not be used to indicate technical inaccuracies, or an altogether wrong answer

The upvoted comment below the post reads:

this merely repeats point made and explained in prior answer that was posted several months before this one

I kindly ask everyone to read the comments before making the review.

11
  • 7
    Both answers would have been NAA if it weren't for that question. See the question. It is off-topic, a resource request... so the answers are actually answers to that question. While reviewing you need to keep these things in mind. Commented Apr 30, 2021 at 4:50
  • 2
    @Yatin, Read again what I have written.
    – scopchanov
    Commented Apr 30, 2021 at 4:51
  • 3
    "It allows to visualize changes and selectively merge specific differences in files and folders." -- what else could they have added to it? It is definitely an answer to a resource request. The 2nd answer does look like a "thank you" answer on the surface but it also links to another tool WinDiff. (tbh I am bit on the edge with the 2nd one, I guess it could be converted into a comment) Commented Apr 30, 2021 at 5:01
  • 3
    For the 3rd one read this: meta.stackoverflow.com/a/342216/11573842 Commented Apr 30, 2021 at 5:05
  • 3
    @Yatin - btw, do we have an actual policy on not flagging NAA if the question itself is off-topic? P.s. Makyen, Machavity - I don't have anything against the approach, but maybe we should update the FAQ then? Commented Apr 30, 2021 at 5:06
  • @Yatin Here is how I understand "giving some value" stackoverflow.com/a/47531458/5366641. For me they both could have been comments.
    – scopchanov
    Commented Apr 30, 2021 at 5:06
  • 5
    @Oleg Valter: Yeah, I'd like a source de jure too. I just know it as a de facto policy - if the question itself invites NAAs, you're just wasting flags and time, and polishing turds by flagging the very NAAs they're soliciting.
    – BoltClock
    Commented Apr 30, 2021 at 5:36
  • 1
    @BoltClock - this sounds reasonable - if the whole is bad, no need to individually process the constituents. If there will be no vehement objections to the practice in the next couple of days, I (or if someone beats me to it) will edit this into the FAQ. Better have a clear guideline available for anyone, methinks. Commented Apr 30, 2021 at 5:43
  • 10
    @OlegValter There is a subtle source in How should I get started reviewing Late Answers and First Posts? "Is link-only (consists in essence of nothing more than a URL of the actual answer, without paraphrase, amendment, or tips included):" It says to "Check whether the question asks for only a link. If so, Flag the question as off-topic (asking for a offsite resource) — no matter how old it is and Skip the review"
    – Scratte
    Commented Apr 30, 2021 at 7:11
  • 1
    @Scratte - argh, I feel like a part-time pro bono lawyer these days :) This one surprisingly clashes with the distinction provided by the FAQ on NAA flagging. I think we are in dire need of unifying terminology. Or, better yet, codifying (I think I've seen a semi-official dictionary floating around) at least some of it. Commented Apr 30, 2021 at 7:14
  • 2
    "The first one is by all means a link-only answer." There, you said it yourself. It's an answer, just not a good one.
    – Michael
    Commented Apr 30, 2021 at 19:43

2 Answers 2

19

My answer is going to say similar things to what Boltclock has already said in their answer here. I'm posting an answer, rather than just upvoting Boltclock's answer and adding a comment, because I'm the moderator who handled the first two flags you are asking about and one of them is "mea culpa".

The first answer you mention is, barely, an answer to the question. It's not a great answer. I wouldn't even call it a good answer. However, the issues with that answer are the purview of downvotes and/or editing, not flags. While it looks like it may be link-only, A) the question is asking for resources, so links to such resources tend to be answers; and B) the name of the product, Araxis Merge, is actually in the URL, which means that even if the current URL dies, you could still find it from a search. Yeah, it's not clear in the URL, but it is there and would be the first, and probably only, possibility I'd try if searching.

On the other hand, for the answer mentioning "Beyond Compare" and "WinDiff": Yes, I should not have declined this flag. I'm sorry about that. It's really not an answer. At the time, I believe I was thinking that it was mentioning WinDiff, which isn't covered in other answers. I was probably just skimming and felt that the answer could be edited to remove the "thank you" portion and leave just the WinDif mention. That is, however, not what's appropriate for what's there in that "answer". You're right that it is, at best, a comment and an NAA flag is appropriate here. I've converted it to a comment on the original "Beyond Compare" answer.

While I didn't handle the flag on the third answer which you NAA flagged, it is an answer to the question, so NAA and/or VLQ flags are not appropriate. NAA flags are only for things which are clearly and obviously not answers. To report more complex issues, like the fact this answer is, probably, just restating what's in the accepted answer, you need to use a "in need of moderator intervention" flag and include an explanation of the issue (i.e. it needs a custom mod flag). [Note: Actually knowing that it's only a restatement of the top voted answer requires knowing that the only possible entries for "Target Framework" are .NET versions. Without domain knowledge, I don't know that. I consider it a moderately reasonable assumption, but it's not something I know and it's not stated in either answer or the comments on the answers.]

You mentioned that there's an upvoted comment on that answer which provides some supplementary information as to why you NAA flagged. Frankly, comments on the post are commonly not even seen by the moderator when evaluating named flags. Comments are not shown in the default view in the flagging interface. What is shown in the default view is the unformatted first few lines of the post. For named flags, the decision on the flag is commonly made based on that view. The moderator can, of course, get more information, but the named flags have very narrow uses such that seeing more information is commonly not necessary.

The underlying reason that the named flags need very narrow definitions (i.e. why reporting things like the above "late duplicate answer" issue needs to be a custom flag and not NAA or VLQ) is that moderators can't spend the time to do a full investigation of every flag looking for all possible problems that might exist. If we were required to do that, then it would increase the time it takes to process those flags by somewhere between 10 and 1,000 times. It's just not possible for Stack Overflow to have enough moderators for that level of time/effort on every flag. Thus, we need the people who are flagging to do the first pass at sorting issues into categories. The named flags are for things which obviously fit into the class of issues for which the flag is named. Things which don't obviously fit into those categories need custom flags, so that A) we know what you think the issue is (i.e. so we don't have to investigate everything while re-doing work you've already done, and still possibly miss the issue you see), and B) we know that we need to spend the time to take a more detailed look.

14
  • 3
    Regarding the first answer, it's nicely explained in the answer to the 'When to flag an answer as “not an answer”' FAQ, under "Links to an answer." It's the difference between "Try function [x](link to documentation about x)" (not NAA) and "Try [this](link to documentation about x)" (NAA) when you do the "strip the markup" test, as in the former case you still know what function to use, in the latter you don't so the post doesn't actually contain an answer to the question (just a link to a possible answer). Commented Apr 30, 2021 at 11:59
  • 7
    Isn't getting bad posts flagged more important than the way they are flagged? More than once I've had a flag declined because I didn't pick the right type of flag, but the moderator action I requested was performed anyway. More than once, that decline was "Familiarize yourself with the Standard Flags" (an ironic, and slightly insulting response to someone who was a former moderator for many years), to which my reply is always "Which standard flag should I have used here?" Commented Apr 30, 2021 at 15:22
  • 1
    If someone raises a flag, the message that they're sending is "something doesn't seem right here." It seems unfair to ding people for doing this, especially if a reasonable person would look at the same post and go "Yeah, this just doesn't look right." I would expect a little latitude and courtesy from moderators in this regard, and maybe just a bit of gratitude for raising the issue, even if the manner in which the issue is raised is not "perfect." Commented Apr 30, 2021 at 15:27
  • 1
    @RobertHarvey Well, the way flags are raised is pretty important, but I agree that the phrasing of the declination is not as friendly as it could be when the action is taken anyway. Perhaps "Declined: While the issue you flagged has been resolved, please raise the appropriate standard flag for such a situation so it's more likely to be correctly handled." would be better? Ideally, the message would also indicate which standard flag should have been raised instead, but that would increase the mod workload, so I think that's too much to ask for.
    – cigien
    Commented Apr 30, 2021 at 20:18
  • 1
    @cigien: The lack of "friendliness" doesn't bother me; the lack of communication and indifference to flags that pointed out a problem but did it in the "wrong" way does. Commented Apr 30, 2021 at 20:24
  • @RobertHarvey I see. Do you have any suggestions for how the communication could be improved, other than rephrasing the message?
    – cigien
    Commented Apr 30, 2021 at 20:34
  • 3
    If I were master of the universe, I would start by chucking the whole "Familiarize yourself with standard flags" thing entirely. It's dismissive, does nothing to educate people about how flags work, and gives the mistaken impression that the standard flag meanings are not disputed or a source of confusion. Commented Apr 30, 2021 at 20:37
  • 3
    Then, I would include in the next moderator training the idea that flags should not be declined unless folks are misusing the flagging system. If a person has reasonable cause to believe "something's not right here," they should get the benefit of the doubt. Commented Apr 30, 2021 at 20:38
  • 1
    @RobertHarvey Oh, so you're actually suggesting that flags should not be declined unless they're wrong. It's a tempting idea for sure, and I would love if there were just a single flag type that indicated "something's up", and where we could leave an optional comment. But as mods have explained, there are simply way too many flags that need handling for them not to be chunked into buckets. At least until we have enough mods so that every flag can be investigated fully, flaggers need to raise the appropriate flags, and if they don't, then I'd say it is a misuse of the system.
    – cigien
    Commented Apr 30, 2021 at 20:43
  • Same thing with "they should get the benefit of the doubt". That would be great, but from what mods have said, there are a lot of incorrect flags raised across the site, and giving every single one the benefit of the doubt is something we simply don't have the resources (i.e. sufficient mods) to do.
    – cigien
    Commented Apr 30, 2021 at 20:45
  • 1
    The point here is that, if the moderator took action on the flag, by definition the flag was useful. Commented Apr 30, 2021 at 20:46
  • 3
    @RobertHarvey No, I don't entirely agree that it was useful. The incorrect flag being raised wasted the mods time by requiring them to investigate an issue they shouldn't have had to. In other words, the flag wasn't as useful as it could have been and that's what the declination signals. Like I said, the language in the message could certainly be softer/clearer, but declining the flag does seem like the right thing to do.
    – cigien
    Commented Apr 30, 2021 at 20:49
  • 3
    It would be really nice if moderation were this straightforward, but it isn't. Judgement calls are always made, from the moderators and the flaggers alike. I think there's a chilling effect on flagging posts when a moderator essentially makes the assertion that "I agree that there's a problem with this post so I'm going to act on your flag, but I don't like the way you cast it." That's not a good look. Commented Apr 30, 2021 at 21:15
  • 3
    @RobertHarvey I agree completely, it's not a good look. I can confirm that I've felt quite frustrated when a flag of mine is acted upon, but declined. But at the same time, it's clear that mods would simply be unable to handle the volume of flags if they are not chunked into categories, and so to some extent users have to cast the right flags for the system to work.
    – cigien
    Commented Apr 30, 2021 at 21:47
12

Yeah, you're right. The one mentioning Beyond Compare was a thank-you answer with only a tiny bit of exposition that's better suited as a comment. While it qualifies as NAA, I don't think flagging it was the best choice as NAA flags tend to be handled in bulk — if they don't stand out immediately as thank-you or some other form of NAA, the flags are likely to get declined, unfortunately. Even I don't feel confident I would've handled this flag correctly, although I do take the time to read every post (and comment) before handling flags of any kind.

Link-only answers are not NAA, especially when the question itself is asking for links in the first place, and the one that you flagged even provides a short description of what the tool being linked to does. Yes, we don't allow tool recommendation questions because readers have to download stuff from elsewhere for the answers to be useful, but you're better off finding link-only answers to things that aren't inherently link recommendations in the first place.

That last one? Hmm. Yeah I guess 17 non-deleted answers is a lot. I guess it wouldn't have hurt to remove that answer and declutter the question, it's old enough that the author isn't going to— wait, you flagged that as NAA too? That's not right. Late answers are answers, and a custom flag should've been used here. You seem to be using NAA as a catch-all flag for any answers that should be removed for a variety of reasons. NAA is to be reserved for specific types of posts only, so that it can be safely handled in bulk (Stack Overflow gets hundreds of these every day and mods only see a fraction of them). There's a reason it's called "not an answer". Flagging things that are answers contradicts this entirely. So it's no surprise that it was declined in this particular instance. It doesn't matter that someone pointed out the nature of the answer in a comment. It's still your responsibility to use flags correctly.

25
  • 2
    Apparently the last one was custom-flagged as a late answer shortly before gnat's comment, and the flag was declined. Can't explain that one. It's been over six years, I don't think the one who declined the flag would remember either.
    – BoltClock
    Commented Apr 30, 2021 at 5:31
  • 3
    "Link-only answers are not NAA" So now we differentiate between "links to an answer" and "link-only answers" (based on the FAQ), with the former being something like "you can try [this](link to documentation about function x)" and the latter something like "you can try function [x](link to documentation about function x)"? The "link to an answer" being NAA (unless the question is asking for links), the "link-only answer" not. Just asking so I get the terminology right. Commented Apr 30, 2021 at 6:12
  • 2
    @Jeanne Dark: My rule of thumb is, would the answer retain any useful information (to the question, preferably) if the links were stripped or otherwise unusable (e.g. by printing the answer)? If the answer text is meaningful as-is, it's not a link-only answer. This is why various usability guidelines recommend giving hyperlinks meaningful labels at the very minimum.
    – BoltClock
    Commented Apr 30, 2021 at 6:15
  • @JeanneDark - I think we always did? That's literally what was in the guideline. But that does not mean link-only answers should stay, just means that we have other mechanisms for it other than flagging NAA. Commented Apr 30, 2021 at 6:17
  • @BoltClock - I think Jeanne referred to the distinction between "link-only" and "link to an answer" that can literally be found in the FAQ, the former stated to be ineligible for NAA flagging and not to what should be considered a link-only answer :) Commented Apr 30, 2021 at 6:18
  • @Oleg Valter: What we currently call link-only answers are what Jeanne is proposing be called links to an answer. I don't know what the best name for these things are, to be honest.
    – BoltClock
    Commented Apr 30, 2021 at 6:20
  • @OlegValter Yes, I saw this answer saying "link-only answers" are not NAA and I also know the FAQ (and your latest edit) with the apparent distinction between "link-only answer" and "link to an answer." To me it has become a bit more confusing now actually. Commented Apr 30, 2021 at 6:23
  • 1
    I'm not proposing it, I'm referring to the FAQ answer: "Links to an answer Essentially this:... Notice that this is not necessarily the same thing as a "link-only answer" (although there is much overlap). In particular, answers where the link itself is the answer to the question are excluded and should not be flagged." Commented Apr 30, 2021 at 6:25
  • 1
    @BoltClock - I propose renaming these to "toad answers" :) Commented Apr 30, 2021 at 6:27
  • 1
    @Jeanne Dark: Oh. I had never bothered looking at that FAQ before. That's new (?) to me. I don't know, I don't really have the time to look into this right now :( All I know is my rule of thumb, and that using NAA for these things isn't the best idea.
    – BoltClock
    Commented Apr 30, 2021 at 6:28
  • 1
    @JeanneDark - I think your definition from the first comment is correct: a pure link to an external "this" is a "link to answer" (literally, the princess is in another castle). That's NAA. A link that contains some info (like method name) is still an answer, but of low quality (and the mechanisms range from downvoting to getting to SOCVR for help) Commented Apr 30, 2021 at 6:29
  • 1
    @OlegValter That's also why I thought that NAA made sense: The "answer" on SO doesn't contain an answer to the question, merely a link to it (that could even break one day). With SO being the place where to find answers to questions, not just links to answers. Commented Apr 30, 2021 at 6:32
  • 1
    @JeanneDark Ok, now it is starting to feel like a Monty Python dialog :) Which of the 2 answer types are you talking about? An answer that only links to an external resource ("check this blog") does not attempt to answer the question -> NAA. An answer that provides some info, but mostly is a link to a resource is an attempt, albeit a very poor one -> downvote. P.s. Isn't that also what you said in the first comment? Commented Apr 30, 2021 at 6:43
  • 1
    ^ your logic is perfectly fine. The FAQ, admittedly, doesn't use the best terminology by calling the former "links as answers" and the latter as "link-only answers". But here we are, the guideline was always there: link-only answers are not NAA. The wording of the whole thing, though is... confusing. Commented Apr 30, 2021 at 6:59
  • 1
    In my understanding it has always been the case that an answer like "Check out this great article" was NAA and should preferably be flagged as such unless the question was asking for links in which case flag / vote to close the question and that's also the guidance in the FAQ explaining the differences between link-only answers. I personally have the habit of flagging such answers as VLQ (eg. see my MSO bug report). Commented Apr 30, 2021 at 8:09

You must log in to answer this question.

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