19

The answer to that question may be obvious to some, but I see lots of great comments on questions (and answers) that could clearly stand as answers to an OP's question. Sometimes they are pointers to the Documentation or a web site. I've also seen one-liner code solutions posted as comments. And some comments encourage people to post their comment as an answer. The tendency to post an answer as a comment probably affects the answer-to-question ratio for this site.

What is your view on when it is appropriate to post an answer as a comment?

2
  • This question is especially important when the site is in beta, so is judged on the number of unanswered questions (and number of answers per question). Of course, that does not mean we should accept low quality answers!
    – Simon
    Commented Jan 20, 2012 at 9:12
  • btw, this is a common problem. SO, math.SE, ThPhys.SE, etc...
    – Simon
    Commented Jan 20, 2012 at 9:20

6 Answers 6

Reset to default
17

I don't think it is (almost) ever appropriate to post an answer as a comment. Since this might discourage other people from writing a similar answer that can be voted on and possible accepted (for other reasons, see [1] [2] [3] etc...) If the comment is just a partial answer posted as a pointer for others, then it should stay as a comment.

As for comment sized answers consisting of just a link:
1) If the link is to a website that can be trusted to remain available and static (e.g., Mathematica documentation, another stackexchange site, a particular version of a Wikipedia page) then a simple and polite comment sized

This question is addressed in this part of the documentation link

is probably acceptable as an answer. Although I would not feel comfortable accepting rep for posting just a link and would probably make it a community wiki. This also encourages other people to expand on the answer if they want to. On the other hand, it does not take much effort to post extra comments or code samples along with the link, so that should be encouraged.
2) If the link is to a website that can not be relied on to remain static, then the relevant content should probably be moved into the answer and the link given as a reference.

Finally, if there is a comment that you think answers the question well, there are a couple of things you can do about it. (Taken from https://meta.stackexchange.com/a/90270/156389)

  • Write a comment that requests the comment to be posted as an answer. (Be sure to ping the person using @UserName)
  • If that gets no response in a reasonable amount of time, then copy (and maybe expand) the comment as an answer. If you don't feel comfortable getting rep for it, then set it as a community wiki.
11

Issac's answer is actually the correct solution here. Comments should not contain answers, especially good answers. I can understand someone suggesting an inefficient (perhaps obvious) solution as a passing remark, but useful information should stay in the comments. I guess the reason why some people do that (from observations on other sites) is because they (rightly or wrongly) feel that such answers are not worthy of having their name attached to — and that's perfectly fine.

However, everything that anyone posts here is CC licensed, and as such others are free to reproduce it as long as it is attributed properly. It is perhaps good practice and common courtesy to request them to post that as an answer first and not rush to post an answer (perhaps they were wondering if the OP had already thought of it?). But if they don't answer within a day (assuming they've been active otherwise) or indicate that they don't want to answer it, then it should be perfectly reasonable for anyone to write that down as an answer.

I'd also point out that there is no necessity for Community Wiki as suggested by Issac — you're free to earn rep from someone else's idea. What difference is there between saying "X said Y in the comments" and getting rep, versus "Michael Trott/Stan Wagon/etc. said Y in their book" and earning rep? Upvotes/downvotes should never be for the person, but for the answer. If you would've voted for it if X posted it, consider voting for it when someone else posts it and attributes to X. That said, people are free to vote as they wish and mark it CW as they wish.

0
10

I'll go a step further and suggest that if you see a comment that is more properly an answer to the question, you might comment asking that comment's author to post it as an answer and if they don't, copy/paste it into a community wiki answer, giving attribution to the original commenter. This is especially true when there aren't other answers or when whatever other answers are there haven't been upvoted, since a question is considered unanswered and periodically resurfaces to the front page if it doesn't have an upvoted or accepted answer.

8

Comments should not contain answers to the question. The best "answer" to Elegantly pairing up mismatched lists is currently in a comment. It's annoying. In this case I'm waiting a bit longer since approaches are still trickling in.

6

It might just be me, but I get the impression that the comments on questions on the main site are quite often useful answers in their own right. Is this because Mathematica is terse enough to encourage people to answer questions in comments, or because they don't want to be seen to be giving trivial answers, or because they don't want their reputation score to increase too quickly? I don't think it has become less common since the above answers were given.

7
  • 5
    Personally, I try not to write an answer if it is a very simple question and instead opt to give a hint or point them in the right direction in the comments. It's generally an invitation for newer users (not Mr.Wizard) to answer it (with or without the hint in the comment) or for the OP themselves to answer it. That comes from my view that it is important that we grow our base and get more people (no matter their skill) to feel like they're contributing to the site, just as it is important to retain the top/experienced users's interest by quickly removing low quality stuff from the site.
    – rm -rf Mod
    Commented Nov 6, 2012 at 15:56
  • I don't want anyone to feel like they're just moochers (i.e., only asking questions) and this gives them an opportunity to start pulling themselves up, both in rep and sense of contribution to the site in terms of helping others out. Often, I give a partial (maybe full) answer in the comments either because I'm busy and don't have time to flesh it out in detail or because I'm not near an mma to test it out. Other times, I only comment because the question is marginal at best and probably should be closed, but I don't want to slam it shut with one vote.
    – rm -rf Mod
    Commented Nov 6, 2012 at 15:56
  • I think several others, such as JM, belisarius, Leonid, bgatessucks, Sjoerd, etc. feel the same. I must note that I don't always manage to follow this... there are times when I'm so bored that I'll gladly take the low hanging fruit and even fruit that's fallen to the ground. But that's increasingly uncommon, given my tightening schedule.
    – rm -rf Mod
    Commented Nov 6, 2012 at 15:58
  • Cormullion, @rm-rf My urge definitely increased recently to write comments like "Why not post it as an answer?". Usually I restrain myself, but I have the feeling that these comment-answers rarely encourage new users but rather put them off. It seems a bit like marking the territory.
    – István Zachar
    Commented Nov 6, 2012 at 21:42
  • @István I don't fully agree that it discourages new users. I've seen it help new users a fair bit: 1, 2, 3, 4 If it's a concern for many, I can stop commenting. As I said in my answer, if someone has posted an answer/hint in the comments and doesn't want to post it as an answer, it should be fair game for anyone else to use it. I think the FGITW by top users is more harmful than hints in comments.
    – rm -rf Mod
    Commented Nov 6, 2012 at 22:12
  • @rm-rf Yes, partially I agree with you, that is why I rarely nag at others. After all, contribution here should be fun, and judging whether a comment is a hint or a perfect answer is indeed tricky. FGITW OTOH is a different matter, this shouldn't be a comparison of the two phenomena.
    – István Zachar
    Commented Nov 6, 2012 at 22:57
  • I do remember at least three cases where I was able to coax the OP into answering his/her own question after a subtle nudge from yours truly in the comments, so I think it works out. Also, it's just that I'm not too enthused about getting rep for stuff that doesn't require much flexing from my part, as well as believing stuff that should be noticed by other people, but not sufficiently substantial (with respect to my internal standard) to stand as an answer, is best left as a comment.
    – J. M.'s missing motivation Mod
    Commented Nov 7, 2012 at 0:52
4

Part of the problem is that people are reluctant to post what they think is a partial answer or just a pointer to a resource as an answer. Obviously we would want answers to be more than just a link (see here), but the high quality and in-depth nature of many of the answers here (yes Leonid and Mr.Wizard I am looking at you) might lead others to think that shorter answers are unwanted or less likely to get upvotes.

I think the main thing we can do to avoid comment-answers is encouraging people to post their comment as an answer, perhaps with a little more explanation added. We already seem to be doing that, so I think the situation warrants monitoring without getting too worried just yet.

3
  • 1
    I think a bigger reason is that people are reluctant to post seemingly minor improvements, when the question begs for an elegant/intuitive approach. For example, in Brett's question, J. M. suggested using ArrayPad[...,1] instead of Flatten[{0,...,0}] in the comments. I understand why he did that – the rest of it was identical to Brett's and he must've thought that Brett merely forgot about ArrayPad and conceptually it was not a very different solution. So I'd say the wording of the question and what the answerer perceives to be a good answer to the question also plays a role in it.
    – rm -rf Mod
    Commented Jan 23, 2012 at 1:53
  • @R.M Note that on that question I eventually created an answer with J. M.'s modification and accepted it. (I had considered putting a bounty on it, but I like having enough rep to edit things, and with public beta coming up...)
    – Brett Champion
    Commented Jan 23, 2012 at 3:44
  • @BrettChampion Yeah, I saw that – that's the right way to do it. The OP must accept what they feel satisfied their needs/criteria/etc, irrespective of what the votes say. I was observing that I can sort of understand why J. M. didn't post it as an answer
    – rm -rf Mod
    Commented Jan 23, 2012 at 3:46

You must log in to answer this question.

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