11
$\begingroup$

I couldn't agree with the comment by Michael Galuza on my answer to Algorithm to answer existential questions - Reduction, but I can see that if the question is "is my understanding right?" and my answer is "yes", then the answer doesn't add much value to anyone who may (very reasonably) have doubts about my qualification to comment. In this case, the OP had done all the work, so there was nothing more to add - all I had to do was to read the paper she was concerned with and convince myself that her reading was correct. Any thoughts on this would be appreciated.

$\endgroup$
  • 3
    $\begingroup$ The ideal case to me is that someone confirms that the OP is correct in the comments, and then the OP transfers his or her solution/explanation from the original question to the answers section. This could feasibly be done by others (posting the answer as community wiki), but it seems best to give the OP a chance to do so first. $\endgroup$ – Peter Woolfitt Aug 5 '15 at 19:47
  • $\begingroup$ I did not check that question in detail, but a cursory inspection shows some imprecision. It thus would seem possible to give a more detailed answer. (I left a comment on main pointing out what I noted.) $\endgroup$ – quid Aug 5 '15 at 21:03
  • $\begingroup$ @quid: It's obviously always possible to give a more detailed answer. Did you take the trouble to read the paper the OP was asking about. Or to try to answer her later question about it? $\endgroup$ – Rob Arthan Aug 5 '15 at 21:20
  • $\begingroup$ @RobArthan no I did not as I made quite clear. I merely browsed the question and found a point were the description of OP is wrong, suggesting they are confused about a point there. This can and should be corrected in an answer. $\endgroup$ – quid Aug 5 '15 at 21:31
  • 2
    $\begingroup$ @quid While it may be possible to find minor tangential points that one can use to expand an answer, I think it is ill-advised to recommend that an answerer spend their time doing such merely to satisfy some arbitrary length constraints on an answer. In many cases the (limited) time of the answerer could probably be used more valuably elsewhere. That said, I often do the things that you recommend in your answer, when I have the time, and when it adds sufficient value. $\endgroup$ – Bill Dubuque Aug 6 '15 at 1:17
  • 1
    $\begingroup$ @BillDubuque It is kind of you that you to endorse the approach I propose. Even the minor discrepancy seem not like real one's, but in the interest of saving everybody's time I will not further continue my arguments. $\endgroup$ – quid Aug 6 '15 at 8:07
12
$\begingroup$

I disagree with quid's answer. Many questions end up unanswered because people take this approach and answer such questions in comments. Of the two risks of leaving the question unnecessarily unanswered on the one hand and answering only the question without giving unsolicited additional information on the other hand, the former weighs more heavily in my view. If the OP had wanted further information beyond confirmation, they could/would/should have asked for it.

Certainly one should pause to think about what one might add to make the answer more useful, and point out small errors or possible improvements in the presentation being confirmed (and apparently in this particular case there was an opportunity to do so), but if none come to mind, one should simply answer the question as posed, and one should also not feel deterred from answering the question as posed if one doesn't have time to search more intensively for improvable details.

$\endgroup$
  • $\begingroup$ "and one should also not feel deterred from asking the question as posed if one doesn't have time to search more intensively for improvable details." I am not sure I understand what you want to say there. Possibly something got garbled. If you meant that one should not be deterred from answering the question as asked, then I would ask you if you think this is a valid approach for other yes/no questions too, and if not where you draw the line? For example "Q: Are there examples of phenomenon P? A: Yes." seems, more often than not, not like an overly helpful answer to me. $\endgroup$ – quid Aug 6 '15 at 17:20
  • $\begingroup$ Further there is also the risk of "gamble answer" (by which I mean users answering without actually knowing an answer). Things like this do happen with some frequency (at least it looks like this). There was no risk in this specific case. But if you propose that it should be generally accepted to answer just "Yes" then you should be aware of the risk of somebody just posting "yes" on a 50/50 chance with a refund via deletion in case of an initial "loss." $\endgroup$ – quid Aug 6 '15 at 17:21
  • $\begingroup$ @quid Sorry, yes, I meant "answering", not "asking" -- I've corrected that. $\endgroup$ – joriki Aug 6 '15 at 17:25
  • 1
    $\begingroup$ @quid: I don't agree with either of your two points. On the first one: This is a whole different case. That question is only superficially asking whether there are examples, and no well-meaning respondent would think that simply writing "Yes" would answer it as intended. In the unlikely case that someone really did just mean to ask literally whether there are examples and wasn't at all interested in learning about them, "Yes." would be a perfectly acceptable answer, much preferable to the question remaining unanswered. $\endgroup$ – joriki Aug 6 '15 at 17:30
  • 4
    $\begingroup$ @quid: On your 2nd point: I thought we were talking about what well-meaning people like you and I and Rob should do. That won't have any effect on trolls who post "yes" without knowing whether that's the answer. So I guess what you're considering here is a norm that should be enforced to stop the trolls, with the unwanted side effect that it will also stop Rob and me from posting because our "yes" can't be reliably distinguished from that of the trolls. That, I think would be a bad idea. I think this problem is sufficiently addressed by reputation -- trolls typically don't get high reputation. $\endgroup$ – joriki Aug 6 '15 at 18:00
  • $\begingroup$ A post is to be evaluated on its own merits and not based on who wrote it. I consider it as quite risky to say some users can give such answers because "we" believe them while others cannot. I was not really talking about trolls. I presented it as gaming the system which is not trolling but if anything a form of cheating, But more importqntly there is a huge gray area of users that might say in good faith "This looks right to me." The value of such an answer can then be very hard to judge, what does a vote even mean there? "I agree, and convinced myself." or "Thanks; I believe you." $\endgroup$ – quid Aug 6 '15 at 21:16
  • 3
    $\begingroup$ @quid I didn't say that some users can give such answers while others cannot; in fact that's a rather serious distortion of what I said. Everyone can give such answers. You raised the potential problem that someone might give such an answer without knowing whether it's true, and I said that this problem is sufficiently addressed by reputation. That has nothing to do with preventing people with lower reputation from giving such answers; it's just very natural, and one of the intended effects of reputation, that the answer will have more weight if the respondent has more reputation. $\endgroup$ – joriki Aug 6 '15 at 21:22
  • $\begingroup$ How exactly is it addressed? So there is a "yes" answer by some user of unclear quality: What exactly should happen next? $\endgroup$ – quid Aug 6 '15 at 21:27
  • 3
    $\begingroup$ @quid: Your argument seems to boil down to this: Absent any remarks about potential improvements or minor errors in the presentation to be confirmed, there's no way to judge whether the "yes" answer is reliable. I do think that reputation provides some guidance on this, but still I certainly agree that providing such additional clues about the reliability is desirable. I wasn't arguing against providing such clues (I try to provide them myself), but against discouraging people from answering if a) there are no improvements or minor errors to suggest or b) they don't have the time to do so. $\endgroup$ – joriki Aug 6 '15 at 21:27
  • $\begingroup$ Yes exactly. The specific question here is really quite a special case as via a latter edit a question for an explanation was then effectively turned into a self-answered question. And I think Peter Woolfitt made a good suggestion. The subject of yes/no answers was discussed some time ago. I assume we can agree that in the end there is a spectrum of possibilities. For the specific question, due to several circumstances, the short answer was unusually fitting. $\endgroup$ – quid Aug 6 '15 at 21:31
  • $\begingroup$ But in general I continue to think it should be discouraged (not forbiddend), The side-effects of having a culture that encourages such yes/no answer-posts in my mind is a lot worse than having some questions that are only answerd in comments, which at the end of the day is merely a minor nuissance. There always can be special circumstance. I wrote what I think (ideally) should normally happen. $\endgroup$ – quid Aug 6 '15 at 21:32
  • $\begingroup$ @quid: What I don't understand: You say "discouraged (not forbidden)" -- but to my mind, your questions about cheaters / system-gamers / gray area inhabitants / ... only make sense if you're suggesting a (potentially enforceable) norm. As long as we're just talking among well-meaning people how we should best act, the cheaters are irrelevant, because they're not here reading this post and they don't care; otherwise they wouldn't be cheating. $\endgroup$ – joriki Aug 6 '15 at 21:35
  • 1
    $\begingroup$ @quid: I don't see any problem with the solution of adding a short comment that you agree if you feel the answer would otherwise be considered unreliable. $\endgroup$ – joriki Aug 6 '15 at 22:04
  • 1
    $\begingroup$ @quid: Perhaps it comes down to actual experience. I don't remember (literally) ever seeing an answer of the form "yes, your work is correct" that seemed doubtful to me, or was even challenged or disproved. If your experience is that such answers occur with a frequency that makes them a real problem, then I'd agree that feeding the cheetas with a supportive comment might exacerbate the problem. Whereas I've never seen this problem occur, I've seen hundreds of questions remain unanswered because people had answered them in the comments even though the answers were perfectly correct and helpful. $\endgroup$ – joriki Aug 6 '15 at 22:15
  • 1
    $\begingroup$ @quid: We agree on all of that. Those latter hints are unhelpful, and proofs that offer room for improvement should be improved. $\endgroup$ – joriki Aug 6 '15 at 22:35
-1
$\begingroup$

Such a question is similar to a proof-verification question. I wrote up my opinion how one should answer those in some detail at How to answer proof-verification questions?

In brief, I think normally more detailed answer can and should be given. A simple "This is right." should better be given as a comment.

Such detailed answers can include:

  • Giving a high-level summary of the argument.
  • Highlighting points were details are omitted.
  • Putting the argument in a larger context.
  • Suggesting alternative arguments.

More details on my opinion around this can be found at the linked to question.

$\endgroup$
  • 2
    $\begingroup$ Yes, but this is about a proof in a published paper. The paper contains the high-level summary and the OP had worked through the details and got them right and had copied copious excerpts from the paper into her question. Putting the argument in a larger context and suggesting alternatives is out of scope given the nature of the paper. $\endgroup$ – Rob Arthan Aug 5 '15 at 21:09
  • 1
    $\begingroup$ I wrote what an answer can include (not that each answer must include all of these); my answer is written as a general one not only for this specific question. Granted, in some cases one or the other point might not be appropriate. For the specific case: as I remarked the presentation given by OP seems not quite correct. One could focus on points of imprecision and explain the details there. $\endgroup$ – quid Aug 5 '15 at 21:29
  • 1
    $\begingroup$ I try to avoid the cursory inspection approach that you disdainfully apply. To answer the OP's question, you need to read the paper she was asking about. Have you done that? $\endgroup$ – Rob Arthan Aug 5 '15 at 21:30
  • 1
    $\begingroup$ For the second time, no I did not, but neither did I answer the question. (But I answered some question or maybe several of that user some time ago.) I do not know what is disdainful about pointing out that contrary to what OP wrote not each element of $F[t, t^{-1}]$ is a power of $t$. OP explicitly asks if there are errors. To identify this error I do not need to read the paper. OP making this false assertion, suggest they are confused about the approach there. An answer could and should address such errors. Why didn't you point out this problem? $\endgroup$ – quid Aug 5 '15 at 21:52
  • 1
    $\begingroup$ because it was of little concern to the OP's understanding of the paper. $\endgroup$ – Rob Arthan Aug 5 '15 at 22:01
  • $\begingroup$ Even if you believe this is the case, to discuss this would seem like a good way to expand you answer a bit, and in doing so you'd address the concerns, the expression of which seems to have prompted your post here. $\endgroup$ – quid Aug 5 '15 at 22:10
  • 1
    $\begingroup$ I don't really agree, but I guess it is a matter of approach. My inclination is to answer a question as asked and see what the follow-up is. On reflection, I think I now appreciate your point of view which is (forgive me if I am wrong) to try to anticipate other concerns that the OP might have. Shall we agree to differ? (PS. I didn't downvote your answer and have just upvoted it.) $\endgroup$ – Rob Arthan Aug 5 '15 at 22:17
  • $\begingroup$ Yes, I agree. Typically, I'd try to give at least some bit of additional information, for OP or at least other reader. But related inquiries come up with some frequency. So it is certainly a situation where approaches do differ. $\endgroup$ – quid Aug 5 '15 at 23:15
  • 2
    $\begingroup$ many thanks for your input. It is much appreciated. $\endgroup$ – Rob Arthan Aug 5 '15 at 23:19

You must log in to answer this question.

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