I asked a question and got a helpful, though not complete answer. I had to put in a bit of work to get it to the point it completely answers my question. I would like to post the complete answer, which is heavily build on the other person's answer. I also would like to give the other person credit. I'm not sure how the credit system works regarding accepting answers.

I see two options for me:

  1. Write a new post with the complete answer in which I write at the top that it is build on the other person's answer and accept my answer.
  2. I edit the other person's answer and accept this.

My questions:

  1. Does accepting a person's answer give the person rep directly or only the up-votes that might follow from the acceptance?
  2. If I do the second option and the answer to my first question is there is no direct rep involved: Do we both get rep?
  3. On the one hand I'd like to give rep, on the other hand editing seems to be not so polity. What do I do?
  • 3
    Short (incomplete) answer comment: If you accept an answer the relevant user will get 15 rep. You will get only 2 rep for accepting. If the answer is incomplete but showed you the path to the right solution, I would accept it. – user31729 Sep 28 '15 at 3:35
  • Ok, that answers one point. How about the other aspects? I only want to except the complete version: So should I edit the other person's post? – Make42 Sep 28 '15 at 9:57
  • You can self-answer you question and accept that if you want. You can still upvote the other answer and even start a bounty for giving reputation points to the other person. – clemens Sep 28 '15 at 14:31
  • My rep is far from shareable ;-). – Make42 Sep 28 '15 at 14:57
  • 2
    I would edit the answer. I've done so a few times, and I've never experienced negative reactions from it. You won't get points for doing so, but then again you didn't ask your question in order to get points anyway. So my answers to your questions: (1) Yes. (2) No. (3) It's fine. – Sverre Sep 28 '15 at 15:40
  • 2
    @user49283 I have too much. If you need a bounty let me know and I can open one – percusse Sep 28 '15 at 18:40
  • 3
    @Sverre: If you would edit my answer you would get a reaction ;-) Editing answers should be done (in my point of view) only on typos, but otherwise leave a comment – user31729 Sep 28 '15 at 20:46
  • 2
    @ChristianHupfer I disagree. If there's a near-complete answer to a question, then I think it's better to make the answer complete rather than post a second almost identical answer or to add comments to the near-complete answer (I see the latter all the time on TeX.sX, but in most cases the person who wrote the answer never updates it to make it complete). I fail to understand why someone improving on your answer would bother you, but I'll try to make sure not to do so with your answers if that's the case. – Sverre Sep 29 '15 at 9:04
  • @ChristianHupfer Why, as long as the edit is clearly marked as such, and the answer is accepted? Sometimes it's too long to be mentioned in comments, or it's just tedious... If the answerer has left some details blank because he didn't test his answer, or because he didn't have time to post a full answer, and the asker does fill in the blanks, why repeat the work? If your triggered "reaction" is "rollback, then edit yourself", there is no harm in terms of content, but isn't it wasted time? – T. Verron Sep 29 '15 at 12:00
  • On github you have a pull request, you ask the owner of the repo to pull your changes. I expect the same from another user here on TeX.SX: Leave a comment and i will consider adding your changes. Correcting typos/markup is ok though. – Johannes_B Sep 29 '15 at 14:11
  • 2
    @Johannes_B Isn't the whole idea of this site based on the idea of people editing to improve? SE is not Git Hub. The site does not recognise answer ownership in the sense you have in mind, even if there is a natural tendency to think in these terms. – cfr Sep 30 '15 at 2:11
  • @T.Verron: I would rollback then – user31729 Sep 30 '15 at 13:55
  • @cfr Some qustions ask for strange things. If i leave such a strange request out in an answer of mine, because i would never advice anybody to do that practice and somebody improves my answer to include what i don't want. Clearly against my will of giving a good answer. <- My pov – Johannes_B Sep 30 '15 at 16:20
  • 1
    @Johannes_B But then, assuming your judgement is correct, it is not an improvement. Obviously non-improvements and active damage are bad, even if the editor believes they are improvements. I've rolled back edits of this kind to my answers - the person meant well, but the changes were problematic so I reversed them and left a comment explaining why. – cfr Sep 30 '15 at 21:05
  • 1
    If somebody needs points I can give it to them, this looks like a hassle to me. Why do I even need to mention the user name etc. ? What do they do with points? – percusse Oct 5 '15 at 22:16

I suggest an option 3: Accept the answer that gave you the solution, and post your own answer. We have many questions here with multiple valid answers, that's desirable.

If there are flaws in the answer you followed (rather than it just not going far enough) then use the comments (and possibly edit).

Elsewhere on SE I've accepted the answer that led to the solution and self-answered with some more detail (which may be more specific to my case anyway). The answer was what I was looking for, not the rep, but I gained some upvotes anyway. This approach may or may not suit your case of course; it's particularly suitable when the extra part you added is very specific.

Two advantages of this method:

  • Accepting the answer shows others that the problem is solved as soon as it is solved (i.e. without waiting for you to have time to write up the answer).
  • A full solution to the exact problem you were having is published.

Rep is a means to an end, you'll gain more by playing nice than playing the system. Not only that but self-answering and accepting doesn't gain anyone any rep.

  • 1
    I disagree with option 3: you should accept the answer that gives the best solution to your problem. If you feel the need to post another answer to your question, with the exact solution that helped you, you should definitely accept your answer. Accepting is not only about reputation attribution, it is about visibility: it is assumed that an accepted answer contains a solution. If you want to give reputation to the user, upvote and give a bounty for his answer. I agree wholeheartedly with the first sentence of your last paragraph, and I believe that it serves my point. ;) – T. Verron Sep 29 '15 at 12:05
  • 1
    @T.Verron when the fine details of my problem are the bit that need additional work, accepting the answer that addresses the more general case (though may not go far enough this time) seems to me to be a valid outcome. Especially if the localised details didn't make it into the title, even more so if they only came out as a result of comments. – Chris H Sep 29 '15 at 13:17
  • If your self-answer only includes localized details, is it really worth posting it? I mean: if the details matter only to you, it probably won't benefit any reader. If the answer requires those details to work (or even only as an example of how to transform the general solution into a specific one), usually, it still won't make it different enough to make it a new answer. In this case, you probably should make these details known to the answer you accept (either by commenting or editing). (cont'd) – T. Verron Sep 29 '15 at 13:42
  • (cont'd) Now if you have found a completely different solution following a suggestion from an answer (e.g. if your situation allows to great simplifications of the process), then it does make a new answer, and you can and should post a new answer. But in my opinion, you should also accept your answer. – T. Verron Sep 29 '15 at 13:44
  • @T.Verron it depends how localised but of course things may be different in different systems. – Chris H Sep 29 '15 at 13:51

Apparently I'm the only one noticing questions in the question. :)

  1. If the other person is not you, it gives him 15 reputation, and you 2 reputation. Otherwise, 0.

  2. As said above, you gain 2 reputation points. Additionally, if you don't have editing privileges, you'll have to suggest the edit, and if it is accepted it'll be another 2 rep for you.

  3. Reputation is primarily "given" by bounties.

    Upvotes are also a way to award some reputation, but it has a visibility impact. If you can't give a bounty for an answer but you really want to, arguably, you could look up the users' top-voted questions and answers and give some upvotes. A post with high vote-count is usually deserving of these votes (at least on this site), that's why I believe it is arguably acceptable. It is abusing the voting feature, technically.

    Accepting an answer has the effect of making it the first visible answer in a thread. This should not be used merely for giving reputation to a user.

About politeness, opinions vary as you may have seen in the comments. I have chosen option 2 several times, on sites where I didn't have edit privileges, and my edits were welcomed every time. In my opinion, you're giving the answerer two favors: you upvote and accept his answer (immediate rep benefit + high visibility => more upvotes) and you improve on his answer (higher quality => more upvotes).

You should probably leave a comment explaining the modifications, and maybe suggest that the answerer includes them himself if he prefers to. And in any case, I suggest that you clearly mark the editted part as such.

But you shouldn't be ashamed of choosing option 1 instead, if the modifications are so drastic that the edit would change the whole answer. In other words, if your edit would be writing a new answer, write a new answer instead. But if the modifications are just about filling some blanks in the process, showing the complete code or answering some doubts formulated in the answer, commenting and/or editting is probably the way to go.

  • People are not suggesting accepting merely to give reputation. The acceptance would recognise that the core solution is found in the accepted answer. A second answer showing how to apply that solution or adding details should not be complete - it should reference the core solution in the first answer. At least, this is how I have approached this kind of case. I certainly wouldn't accept my own answer if its core was provided by another - even if my answer had to include that core in order to make sense. I'd accept the core solution. – cfr Sep 30 '15 at 2:23
  • I was addressing question 3. of the OP, where he specifically asks about giving rep. And my point is rather for the case when the core is not applicable without the additional details of your answer. If you make another solution, even referencing the core, it effectively becomes a better solution than the one provided. In this case, accepting the incomplete one sounds dishonest to me, but I see now that two concurrent visions of this situation exist. – T. Verron Sep 30 '15 at 5:54
  • 1
    +1 just for completely reading my question post! I feel that on SE nobody reads my questions, instead so many questions are marked as "double" or "irrelevant", in so many situations where this is only explainable by assuming that the moderators did nor read my post and/or the post they assume to be a copy. – Make42 Sep 30 '15 at 8:17
  • @user49283 That's a whole other story, and it's much more subtle than "SE vs the outside world". Here on meta, people are discussing as well as answering, which may lead to some subquestions being forgotten in the heat of the debate. On the main site, if an answer or dupe flag seems out of place, usually it means that the users believe they understand the question better than the OP. In some cases, it is true, and the OP is better served by trusting these answers/flags if he's in doubt. (cont'd) – T. Verron Sep 30 '15 at 8:23
  • (cont'd) In other cases of course, it is blatant misunderstanding, due to some people not reading the questions well enough, and/or some people not explaining their problem well enough. In my experience it doesn't happen that often on tex.sx though. – T. Verron Sep 30 '15 at 8:24
  • Sadly it did happen to me as well on tex.sx: tex.stackexchange.com/questions/269296/… It took a while until my esteemed reader noticed the little word "comparison", which was not discussed in the proposed duplicate, but to be fair: the word "comparison" was easy to miss. ;-D – Make42 Sep 30 '15 at 11:32

You must log in to answer this question.

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