What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

When someone is asking about, for example, how to add a feature, or change behavior to/of some program for which you know it is not possible, is it acceptable to post a "this is not possible" answer?

I know this is not an answer in itself, but then there is a dilemma:

  • Leaving the question without answer (in this case, the user can wonder if simply his question had no success, and he would eventually start a bounty for something which can't be answered).

  • Answering "there is no solution", and explaining why (which would provide information to the user, but at the price of discouraging further answers, and of course the risk of simply being wrong).

What would you recommend to do in such situation?

Edit: Bonus question: When there is such an answer, what about the voting behavior? Should such answer be upvoted if you agree that indeed, there is no solution? In the end, this is not giving any solution, so it looks strange to vote for it.

share|improve this question
    
Also, don't forget about good ol comments. They work great for simple responces. –  Troggy Oct 27 '09 at 15:33
1  
@Troggy, true, when it's simply to say "I dont think this is possible", without listing alternative, or giving a lot of details, it's probably a better way to do. –  Gnoupi Oct 27 '09 at 15:37
3  
A strange game. The only winning move is not to play. –  Kip Oct 27 '09 at 17:42

11 Answers 11

up vote 15 down vote accepted

Answering "there is no solution", and explaining why (which would provide information to the user, but at the price of discouraging further answers, and of course the risk of simply being wrong).

This is a great way to answer these types of questions. Provided you are confident in your answer. As long as you give a reason why, I would find that to be a complete answer. You might provide alternatives if possible or similar solutions/code/software/etc. or whatever the question entails.

As for using bounties for these kind of questions, it just brings more attention to the question and will have a greater chance of having more qualified and complete answers. That is not always the case, but brings lots of eyeballs to the screen.

Remember to first make sure the questions are being asked in the best way with complete information, links, and details and then see where it goes from there. I have been proven wrong many times by someone later that found a solution or something I have never heard of yet.

share|improve this answer
    
About the bounty, if the answer really is "no solution", it might bring a lot of eyeballs, but will only increase the chance of more "at all cost" answers. These answers, which would recommend a solution rather not matching, or at least not matching to the user, will potentially get votes (typically if citing a really loved program). In case of a bounty, it might finally lead to have a wrong solution being accepted. –  Gnoupi Oct 27 '09 at 15:56
3  
These "at all cost" answers are awful. They really add to the general noise quantity... In a question that has a bounty I would feel cheated if one of these answers became the accepted one. I'm ok with it's not possible as long as it actually isn't possible; however, I'm not ok with a mediocre answer that's almost irrelevant. –  alex Oct 27 '09 at 16:03

Yes.

"There is no solution to this problem" is logically different logical situation than "The existence or not of a solution to this problem has not yet been established" (and, of course, than "I don't know."). By saying "There is no solution", you rule out (unless you are wrong, of course :) the possibility that further waiting would result in someone's investigations getting to find a solution to the problem.

share|improve this answer
1  
Indeed. Probably in most cases, such answer should include precise details about the "situation" (system, version, etc.), to be actually right. That's the problem in an IT environment, it is hard to be 100% sure that no, there is no solution and won't be, as this environment evolves (and quite fast). –  Gnoupi Oct 27 '09 at 15:52
2  
I'd say that "The existence or not of a solution to this problem has not yet been established" is also an answer; for example, it's an excellent answer if the request is a halfway efficient exact solution for the Travelling Salesman problem. –  David Thornley Oct 27 '09 at 15:53
    
@Gnoupi, @David, yes, of course, I agree with you. That answer gives useful information to the reader. The only answer that really shouldn't be posted is "I don't know." –  Daniel Daranas Oct 27 '09 at 15:57
2  
I liked the description in the first few pages of Garey & Johnson about NP-complete problems: "I don't know how to do it, and neither do any of this really long line of other smart people." –  David Thornley Oct 29 '09 at 13:52
    
@David - It's smart people all the way down :) –  Daniel Daranas Oct 29 '09 at 20:25

Answer that there is no solution, and explain why if you can.

If you can stop someone from going on a wild-goose chase, why wouldn't you?

share|improve this answer
1  
Sometimes the chase is more important than the destination. If we never did the wild-goose chase, we'd never learn anything. :-) –  Brian Knoblauch Oct 28 '09 at 13:05

While in theory, I'd like to say "Yes", I'm forced to say "No" after observing... People are saying "No solution" to questions that indeed there ARE solutions to. They're just not as smart as they think they are...

Perhaps some sort of middle ground, such as listing potential problems with it that you don't know how to solve (and indeed may be unsolvable).

share|improve this answer

How about responding "There may be no solution to this problem, and here's why:". Followed by a detailed list of the probative data as to why it's impossible/unsolvable.

share|improve this answer
    
I have seen answers suggesting that a problem might be NP-complete. –  David Thornley Oct 29 '09 at 13:53
2  
"As far as I know" is also a good way to preface such an answer. –  Tim Post Oct 29 '09 at 14:33

If there really is no answer, then it should be perfectly acceptable. However, try to avoid answering with "there is no answer" if the solution is simply advised against.

share|improve this answer

If there is no solution, ask about the question's general context - what is the person who asked the questions trying to achieve... maybe there is a possible alternative way.

share|improve this answer
2  
Assuming we know what the questioner really wanted. Too many times people ask how to do the impossible, unlikely, or what's generally a really bad idea without mentioning what they want to accomplish. –  David Thornley Oct 29 '09 at 13:54

I would say that to an extent kind of depends on how the question is phrased, but answering a question with a negative response (i.e. "There is no solution.") is acceptable as long as the answer provides a valid explanation of why there is no solution as well as suggestions as to what might be done as a workaround.

For example, if someone where to post asking for a polynomial time solution to the subset sum problem, then the best answer would be that the problem is known to be NP-Complete and that a polynomial time solution is not known to exist. This could then be followed by providing some suggestions in regards to deal with the problem if it has to be solved anyway (e.g. brute force it if the input size is limited).

share|improve this answer

There is no solution to this problem. :-)

share|improve this answer
    
downvote bait :-) –  beggs Dec 7 '09 at 2:43

Hate bringing up old posts uselessly, but sometimes when you know an answer might be out there, but just difficult to implement or find, the ding to your acceptance rate can be really annoying, for instance, this actually just happened to me here.

I mean, sometimes, there's no obvious answer, but for instance, if someone asks how they can make 1+1=3, the most obvious answer, is you can't. But at second glance, in programming, and several languages, this is very easy to do. You could create an Add(1,1) function that returned a value of 3, you could create a custom number class, and override the addition operator, you could use an alternative numbering system...

I mean, some answers are ridiculous, and impractical, and can be explained as such, but you don't necessarily know the requirement or why someone needs it; so a provided answer, no matter how ridiculous or impractical it might be, as long as its explained properly; I believe is better than an answer that says, nope, can't do it.

share|improve this answer
    
I don't think there is anything remotely wrong with your 89% accept rate. But I understand your point, still. –  Andrew Barber Dec 5 '11 at 17:13
    
Aye, I just feel it discourages more people from doing research or actively helping you too, cause they look simply, and see "no solution" and either start posting comments/answers saying they agree, looking for an easy upvote from the masses; or they move on. –  Brian Deragon Dec 5 '11 at 17:17

I did the same :).

share|improve this answer

You must log in to answer this question.

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