5
$\begingroup$

My answer to https://cs.stackexchange.com/q/13495/755 was converted to a comment. Why?

I believe my answer was an acceptable answer. The question was "How do you show..." and my answer described how: it said "I would suggest that you start by... it'll probably even give you some examples ... which you can apply to [this problem] ...". My answer was not a request for clarification; it was a bona fide answer that described how to solve the poster's problem.

I read through the official criteria for when an answer should be converted to a comment, and none of those apply (as far as I can see). And there are reasons why answers should remain, well, answers: e.g., questions without an answer get treated differently, comments aren't searchable, others cannot edit a comment to improve it, others cannot downvote a comment, the original poster cannot mark a comment as the accepted answer, etc.

Obviously, this case is very minor: it's just one question, and not even an important one, so big deal. I don't really care about this one case. But I do care about the general policy; I think being overly aggressive about converting answers to comments is not so great for the site.

So, why was this converted to a comment? Was it just an oversight? Or is there some general policy that I've missed that explains why this conversion was the right thing to do in this situation?

Edit: The moderators subsequently closed the question. I do not have any objection to moderators converting my answer to a comment, if they're going to close the question as well. I think this topic is now moot. Thanks to Gilles for his detailed thoughts and explanations.

$\endgroup$
2
$\begingroup$

I would suggest that you start by reading the Wikipedia article on NP and reading your textbook chapter on NP. That should explain how to do this problem, and it'll probably even give some examples of how to prove that other problems are in NP -- which you can try to apply to the Knapsack too. That will be a good test of whether you understand the definition of NP or not.

Ok, I've re-read your post, and I still don't see any part of it that answers the question. The question for reference is:

How do you show that the Knapsack Problem is in NP that is more formally
[snip: formal statement of the KP]

Sure, this is a poor question, but that doesn't excuse a non-answer. Your post is a fine comment to go together with a close vote: it explains what the asker should do instead of asking the question here.

Your post is not an answer because it doesn't answer the question. An answer would explain how to show that the Knapsack Problem is in NP. All your post does is say that the asker may be able to find the answer somewhere.

An answer can be partial. But this doesn't even begin to answer the question. It only explains where a small part of the answer may be found, and gives some very general methodology advice.

Think of it this way: if you were grading an exam question asking to show that the KP is in NP, would you award your post even partial credit?

I don't see your post in the yoda classification. It's a bit like “See here: [external link]” (which is a contested case — while Stack Exchange isn't supposed to be a link collection, there is some pressure to allow links to solutions to be posted as answer because it's a quick way to gain reputation with no effort), but the link you give doesn't answer the question at all.


More generally, I have noticed a trend towards answers that only give hints, and not full solutions. Such answers are very bad for the site. We are being too lax about answer quality. A hint-only answer is rarely useful to future visitors. Remember that when you post an answer, it isn't only for the person who just asked the question. It's also for all the people who will later find the question in the search. Stack Exchange is a questions and answers site, meant to build a repository of answers with lasting value. Leave the hints for forums.

$\endgroup$
  • 2
    $\begingroup$ "I have noticed a trend towards answers that only give hints, and not full solutions. Such answers are very bad for the site." - It looks to me like the community does not agree with you on this. References: meta.cs.stackexchange.com/q/324/755, meta.cs.stackexchange.com/a/471/755. If you think this does not accurately characterize the community's opinion, I suggest posting a separate discussion question to resolve the matter. $\endgroup$ – D.W. Jul 30 '13 at 7:54
  • $\begingroup$ Thank you for closing the question. That feels more consistent to me. I don't have any problem with the treatment of my answer, if the moderators are going to close the question. (I would have more of a complaint if moderators are going to convert my sort of answer to a comment, while leaving the question open. But since this question was closed, I have no objection or complaint. Thank you!) $\endgroup$ – D.W. Jul 30 '13 at 8:00
  • 1
    $\begingroup$ @D.W. Conversely, I would have expected that you'd vote to close. If you feel that a non-answer is the best response to the question, then you should indicate that the question should not get proper answers, i.e. that it should be closed. $\endgroup$ – Gilles Jul 30 '13 at 9:49
  • $\begingroup$ @Gilles, I understand -- but I did not feel my answer was a non-answer; I felt (and still feel) my answer is an answer, so your premise does not apply. However I get your general point, and in the future I will vote to close on this kind of question if that's the right thing to do. Thank you! $\endgroup$ – D.W. Jul 30 '13 at 15:33
  • $\begingroup$ Let me add that one way to deal with solution-diggers is to add full solutions later (i.e. after potential deadlines) and/or in a spoiler tag (>!). $\endgroup$ – Raphael Aug 27 '13 at 10:56

You must log in to answer this question.

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