13
$\begingroup$

In this answer a user suggests that "answers should almost always be posted as answers”, which I quote

If the question exists elsewhere on the site, and you have an answer for it, it should go on that question. if the question doesn't exist, you should write it up and self-answer it.

This is the best fit for our format.

The primary goal is to get feedback about your solution. And the way we get feedback for solutions is to post them as a solution, and then people have at them.

(FYI: the tag is now identified to , so I am using modern day terms.) This suggestion gives arises to two questions. A stronger one is:

Should solution verification (always) be posted as self-answers?

And a weaker one is:

Can solution verification be posted as be posted as self-answers?


There are a few scenarios that I can imagine, in which I would like to ask a question:

  1. I finished the proof, and a something in the middle is contradictory to a result that I found in the other book.
  2. I finished the proof, but I am not confident that the conditions to plug in a theorem are indeed justified.
  3. I finished the proof, and realized that my methods also applies to a significant generalization of the original problem. Sub-cases includes:
    1. I was going to do a college-level exercise, but indeed proved a famous unresolved problem.
    2. I was going to prove $p\wedge q \to r$, but indeed proved $p\to r$. That is, some seemingly critical conditions are not used.
  4. I originally had a problem with my proof, and the problem resolved as I typed in that question. But I am still only 85% sure that the whole thing is correct.

Because of the premise of (1), (2) and (3.1), I think the answer to the stronger question is “no”. But this still leaves the weaker question open.

On the one hand, it looks fruitful to post answers as answers. Because if the answer to “Is my proof stated as … correct?” is likely “Yes”, such questions are unlikely to receive a well-written answer (specifically, a well-written answer in MSE. A related opinion. How to answer proof verification questions?).

On the other hand, I haven’t seen anyone doing that. In fact, I haven’t seen many self-answers. Occasionally people write up their solutions after being hinted and/or encouraged in the comments, but the actions of simultaneously posting questions and answers are, in my perspective, rare. Moreover, it seems difficult to do that. It appears that the community has a high-quality standard for self-answers (source: e.g. this). Also, providing context can be hard. (Is self answer considered as adding sufficient context?)

I would like to know under what circumstances posting as self-answers is allowed/encouraged, and if there have been successful examples of people doing so.

$\endgroup$
6
  • 4
    $\begingroup$ In short “can you..?” Yes. “Should you…?” Apparently there is not a consensus yet. What I really wanted to impress upon a reader of that solution is all the negatives of wadding a question and answer together, and to give suggestions about how they can be constructively pried apart. $\endgroup$
    – rschwieb
    Apr 10 at 12:12
  • 5
    $\begingroup$ Despite my preference that solution verification should be avoided, I would not say ALWAYS simply because an absolute proclamation like that would be bound to be wrong sometimes. I’m sure there are some borderline cases in which it is the best way to go. $\endgroup$
    – rschwieb
    Apr 10 at 12:15
  • 1
    $\begingroup$ In the list of four things you gave, applying the position on my post would say that 1 and 3 might still be good candidates for posting the solution in the answer, because you have definite cause for confusion and you cannot see the resolution. 2 and 4 look more like “I have a solution which for all I know is completely correct, please proofread for me”. I think the ones like 2 and 4 should be worked on by the poster longer until they turn into ones like 1 and 3, or else it is ready to itself be a solution. $\endgroup$
    – rschwieb
    Apr 10 at 12:20
  • $\begingroup$ @rschwieb Thank you for your response! I have seen your answer elsewhere. I think it is interesting and innovating, but I haven’t ever seen people putting that into practice. Also, as I pointed out in the question, this can be hard, and I haven’t heard of any guidelines for contexts of such Q&As. Therefore, I started this discussion in seek of more community opinion. $\endgroup$
    – fantasie
    Apr 10 at 12:43
  • $\begingroup$ @rschwieb By the way, “applying the position on my post would say 1 and 3 might still he good candidates for posting the solution in the answer”: do you mean “in the question”? That is, a traditional solution-verification question. $\endgroup$
    – fantasie
    Apr 10 at 12:46
  • 2
    $\begingroup$ oh yes of course: I meant “solution in the question”. Indeed, if what I was suggesting was common practice I would not have to mention it. For someone who just wants to write a post It is simply easier to write a post with whatever you want in it rather than to evaluate what it means for the site quality overall and doing a question plus solution. For those of us who have experience seeing lots of posts over the long term, suggesting something less obvious than common practice becomes more attractive. $\endgroup$
    – rschwieb
    Apr 10 at 13:45

1 Answer 1

2
$\begingroup$

It's weird that you don't try to address the points made in my answer on that linked thread, which explain why all your scenarios fall under the category of solution-attempts that should not be posted as 'answers'. To emphasize again the main points:

  1. This is not a forum, but a Q&A site, so the notion of "answer" should be reserved for those who are confident of their mathematical contribution to the question at hand. I've seen many many Math SE 'answers' posted by users who think they have a proof but actually don't, and many of them lie around unchallenged. People who have a very weak foundation in mathematics should not be dumping their so-called 'proof attempt' of any mathematical problem onto an existing thread asking about that problem! This is the reason for having the solution-verification tag, with the goal of distinguishing clearly between actual answers and attempts.

  2. Proof-verification questions exist to serve the needs of students who need guidance in writing proofs, perhaps because they were not taught basic first-order logic, and thus cannot quite put together a logical argument from their ideas. This definitely covers all your scenarios.

  3. Proof-verification questions can be closed as duplicate if roughly the same approach had been asked about before. It is the responsibility of the asker to do sufficient research to determine whether his/her approach has been tried before, and post a question only if he/she cannot find it (at least on Math SE). This probably does not apply to you.

  4. A proof-verification question will not just be a request to check some proof. Rather, it would specifically identify something that distinguishes it from other known proofs (including those the asker have found on Math SE). Mistakes are also more easily identified; if the asker takes effort to highlight how his/her proof differs from the known ones, experts can much more quickly tell whether the difference is inessential or not. This criterion greatly enhances the value of the question.

Concerning the issue of answering a solution-verification question, I don't see why quid's answer is problematic to you, especially in light of the above criterion for the question to be a good one. In fact, requiring the asker to pinpoint what he/she finds most troubling often causes the asker to resolve their own question without even asking, and if they still need to ask it will indicate some significant conceptual misunderstanding that will certainly not be answered by a trivial "yes" or "no". So this approach solves that problem too.

$\endgroup$
10
  • $\begingroup$ Thank you for your respond! I agree that it is better to address your answer, e.g. in the “On the other hand” paragraph, but there were a lot of information going around and I just didn’t notice that. (Maybe I will include them later.) But I don’t think there is a direct implication from your answer implies all these scenarios should not be post as ‘answers’. In fact, it is being debated, and after years the two answers supporting/opposing that idea still get even scores! $\endgroup$
    – fantasie
    Apr 22 at 16:39
  • $\begingroup$ As per quid’s answer, I think a solution-verification question can be answered only when there is indeed some points worthwhile to add. Take this question for example. I asked it six month ago and the answer from comments is “yes”. No one has written an answer, and I doubt if anyone will. But I did intend to believe it was my fault instead of the book prior to those comments. So, I opinion is that “that will certainly not be answered by a ‘yes’ or ‘no’” is a very strong argument. $\endgroup$
    – fantasie
    Apr 22 at 17:15
  • $\begingroup$ @fantasie: Just because you haven't gotten an answer to a question doesn't imply anything about whether it is good. Also, my post here assumed that the source of the problem was correct and didn't have any flaws, so it doesn't apply to your example where the source added a redundant condition that caused your doubt. =) $\endgroup$
    – user21820
    Apr 22 at 17:24
  • 1
    $\begingroup$ Also, debate is good, but nobody has given any solid reason to deviate from the current practice of reserving the answer box for people who have the sufficient competency in the relevant topics to generally ensure correctness. Encouraging people to post unsure attempts in the answer box is just silly in my opinion. It is completely different if the asker has enough competency to verify correct proofs, and at first couldn't solve the question, and later post an answer when he/she has found a proof. The latter is definitely encouraged. The former is not, and should never be. $\endgroup$
    – user21820
    Apr 22 at 17:31
  • 1
    $\begingroup$ the current practice of reserving the answer box for people who have the sufficient competency in the relevant topics to generally ensure correctness. Current practice? You are the only person I have ever seen attempt to use the concepts of the two groups "people who are competent to answer" and "people who aren't." Is there some policy you can point to that reflects what you are saying? $\endgroup$
    – rschwieb
    Apr 22 at 18:47
  • $\begingroup$ Also, I think it's a little weird to oscillate between "surety" and "competence" in that last comment. "Surety" is one's self-assessment of their competence. The thing is one's surety is not a reliable indicator of their competence, nor is their lack of surety a reliable indicator of their incompetence. I think the best we can hope for is an earnest attempt in good-faith, and moderation is there to handle the bad-faith. Maybe you can provide some refinement on this point. $\endgroup$
    – rschwieb
    Apr 22 at 18:50
  • $\begingroup$ Also who encourages people to post unsure attempts, whatever they are? $\endgroup$
    – rschwieb
    Apr 22 at 18:51
  • $\begingroup$ @rschwieb Well I also think that those who are going to post an answer, is because they are sure about the correctness of the answer. The attempt in good-faith could be either in a comment or maybe in a discussion, say in a chatroom. Once they're sure, then it can be posted in the answer box. $\endgroup$ Apr 22 at 19:15
  • 1
    $\begingroup$ @rschwieb: Your linked post does encourage people to post unsure attempts as answers. In particular, there are 3 main kinds of posts: (1) by competent author who thinks it is correct; (2) by incompetent author who thinks it is correct; (3) by incompetent author who thinks there is a problem. We don't disagree on (3); you also say these should go into the question. We disagree on (2). I didn't mean to confuse you regarding surety and competence, but I said ( encourage posting unsure attempts is silly ) and ( only competent users should post answers ), and yes most contention is in-between. $\endgroup$
    – user21820
    Apr 22 at 19:32
  • $\begingroup$ For example, this post explicitly states that the author is unsure. There are many other examples, but this was the easiest to find in a few minutes. The biggest problem is that there are many users who post unsure solutions without that disclaimer. $\endgroup$
    – user21820
    Apr 22 at 19:33

You must log in to answer this question.

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