32

Sometimes you can see a question like How to do <something> in <some program>, asking for advice how to replicate a feature of competitive program or how to do something that OP expects to be possible, but couldn't find out how to do it.

If the feature just isn't available, like in Is it possible to crop and resize in one step in GIMP?, is it acceptable to post "It's impossible" as an answer, or should it be a comment?

Why it should be an answer:

  • It actually answers the question if OP asks "Is it possible to do <something>".
  • If it's wrong, it can be downvoted.
  • If it's correct, it can be accepted. Otherwise the question looks like it's unanswered.

Why it should be a comment:

  • It doesn't solve the real problem. If OP asks "Is <XYZ> possible", he really means "How do I do <XYZ>".
  • In most cases, short answer is a poor answer. For example alternative solutions or workarounds could be provided.
  • Such answers shouldn't receive points for upvotes or accepting, as they aren't helpful. Getting rep for not being helpful is bad.

Alternative solutions:

  • Adding a Unsolvable close reason would both indicate that the question is kinda answered and prevent getting rep for unhelpful answers. But it would block the question from being answered in the future if the feature is ever implemented in the mentioned application. Also, it looks like an overkill to me.
  • 8
    In that particular example, something like @iglvzx's comment "I can write up an AutoHotkey macro for this" qualifies as proof that yes, it is possible to do. You could also make a GIMP plug-in to perform the two actions in one UI operation. It might not be supported in a default installation, but that is very different from it being impossible to do. – a CVn Sep 18 '13 at 12:07
  • 1
    Nothing is unsolvable. Sometimes it just requires changing the (often assumed) constraints. – JamesRyan Sep 19 '13 at 9:40
  • 2
    @JamesRyan many things are unsolvable. Many others, yes, can be solved changing constraints. – o0'. Sep 29 '13 at 8:25
51

Some things are simply impossible - however, it's useful to say why not (or better yet, how it could be done, but how incredibly difficult it is).

If we had a question like

I need to alter my turbo enfrobulator using the degobulation widget on frobster 2.0 - how do I do this?

You can say

Enter image description here

Or you can explain that you can't enfrobulate from the degobulation widget - it only handles degobulation related settings - you need to do so using the enfrobulation command line tool.

The former doesn't help. The latter does. If you need to say it isn't possible - it's useful to say why it isn't possible.

  • 4
    That is just so true. There's many answers only telling "I believe that it's impossible". Sad thing is that some of those answers are written in complete lack of knowledge and outcome for this is upvoted misinformation. – Sampo Sarrala Sep 18 '13 at 20:57
  • 5
    I'd upvote twice if I could; "no, you can't" is bogus, while "no, you can't, because reasons" is quite useful. – Aaron Miller Sep 18 '13 at 21:08
  • 1
    Typo: degubulation, replace an "u" with an "o". – nalply Sep 24 '13 at 10:05
  • 1
    heh, remind me to add that to spellcheck ;) – Journeyman Geek Sep 24 '13 at 10:53
  • 1
    @SampoSarrala agreed. I've seen multiple cases where a question got an initial "Impossible" answer only to have someone post instructions on how to do it months later. – Dan Neely Sep 24 '13 at 14:15
  • Some things truly are impossible, though. For example, someone has recently asked "How can I get Lotus Notes to do an autoresponse based on subject line?" Well, Notes doesn't doesn't have that feature, so it's near-impossible unless you want to write some bit of inordinately obscure code. So perhaps one could say "It's impossible", but I'd prefer "Notes doesn't have that capability, and coding to do it might well be far out of balance with the benefits you'd derive." Maybe. On the other hand, it's also obvious that sometimes someone says "impossible" in response to a solvable problem. – Debra Sep 27 '13 at 4:15
  • 3
    +1 for sad cat. – GlassGhost Sep 28 '13 at 21:58
25

If what the OP asked for is clearly impossible, it is acceptable to state so in a full answer, as long as a reasonable explanation is provided.

If you just want to say that it's impossible, it's not worthy of an answer as it is little more than noise. At Stack Exchange, we want to keep the quality of answers high, so an answer that states that the desired action is impossible needs to be backed by an explanation.


For example, let's say someone asked the question, "What is 4 divided by 0?"

The answer I would give would be like this:

You cannot divide by zero. It is considered undefined.

Division by zero is undefined because you are effectively trying to solve for x in the equation x ∙ 0 = 4. As there is no value x that can satisfy this equation, it is not possible to define division by zero (given a non-zero dividend). In the case of zero divided by zero, any value of x will satisfy x ∙ 0 = 0, making this operation indeterminate and therefore undefined as well.

For more information on division by zero, see this Wikipedia article.

11

Those are always edge cases. I would be hesitant to provide a simple rule on how to deal with these kinds of answers, although one thing should be clear: "You cannot do it" can be a valid answer.


However in this particular instance it's clearly this:

It doesn't solve the real problem. If OP asks "Is possible", he really means "How do I do <XYZ>".

In general, it doesn't make a lot of sense to ask whether something is possible. Asking how to achieve it is always preferred, and if the answer is, "it's not possible", then so be it. Clearly, in this case, the OP needs to do something very specific – a workflow in GIMP – and just telling them you can't do it is

  • not particularly helpful
  • maybe not even true (who knows, maybe it's possible but the answerer hasn't found a way to)

So, this question here can't be compared with others where you can actually prove that something cannot be done. Unless you can technically argue that it's not possible – and will never be – then that's a full answer to the question. I could come up with plenty of examples where this would be acceptable: Imagine asking how you can write data to a CD-ROM, or how you can make a low quality compressed movie high quality. It's just not feasible.

4

I asked a similar question on Meta.SO, just over a year ago.

What should we do about potentially unanswerable questions?

In the comments to the question's single answer, Yawus states:

[Q]uestions with content issues can generally be closed as "Not constructive" or with a custom mod flag.

This was the best answer that I could get at that time, but after a bit of thought on the issue, my general take-away has been that not all "impossible" questions are impossible.

For instance, consider a question that I posted in August 2011 regarding a feature of an old C compiler.

LCC— How to stop debugger from breaking at the start of the application?

When I initially asked this question I assumed there was a setting buried somewhere that could cut this behavior off. However, after a bit of comment discussion I learned that this was impossible from a practical sense. Eventually, though, a less than conventional solution was found!

User ruslan provided a not-too-simple answer that involved decompiling and patching the binary! I had assumed my question would never be answered with a real response but I was wrong.

The bottom line is, impossible doesn't always mean "impossible". Sometimes legitimate hacks and tricks can provide a solution to seemingly unanswerable problems. Also, impossible questions regarding software can become answerable, in the future, if the software is updated by the developers to provide the discussed functionality. This can be especially true if you are working with pseudo-beta software that some companies like to release into production (cough anythingmadebygoogle cough.)

If an update to the software occurs that addresses the questioner's issue, it's completely legitimate to post an answer that mentions the software update that solved the problem. If a software patch is ever a likely scenario, I personally think it's worth letting the question "dangle" as unanswered until a solution is found.

I've had this happen in many cases. I've been surprised with some of the legitimate, albeit unconventional, answers that I've received.

3

Saying anything is impossible is neither a valid answer or a valid comment :-) Because nothing is impossible , it just isn't easy, or has not been done yet.

Putting "it is impossible" at the top of an answer, and then explaining a few ways that it can be done via another method or another program, qualifies as a votable answer. Providing your experience on how you found that it was impossible, is votable too.

Just dumping in a negative is so little help, it would be better to leave the question yet unanswered.

The future , you already covered well. I have seen "locked" and "Closed" threads on other forums where the "answer" was "impossible". It WAS possible, because I had since done the impossible, and was unable to post it into thier forums. Thats just wrong, very valuable information was lost because the powers that be deemed it impossible.

So get out there and do the impossible, humans can't live in caves forever. Ugg Mano says cave only way.

  • 1
    Some things are certainly impossible. I need to accelerate my R18 Honda Civic to C without converting it energy <-- Impossible. I need to legally install Mac OSX on my IBM PC, currently it is running the factory installed version of OS/2... Also not possible... – Austin T French Sep 17 '13 at 17:24
  • 2
    It might be useful to say why "You really do not want to be messing with that fiiiine museum piece, nor invoke the ghost of jobs". Impossible is boring. Teasing apart why it is impossible is an interesting mental exercise, and makes for an entertaining answer. – Journeyman Geek Sep 18 '13 at 11:19
  • 1
    @AthomSfere: And those are votable answers. Psyco makes a good point these questions shouldn't be closed. A determination of impossibility requires subject matter expertise, and therefore is not suitable for moderation. – Ben Voigt Sep 28 '13 at 15:23
  • @BenVoigt I don't think they should be closed, just that impossible as an answer is sometimes correct, given sufficient details. – Austin T French Sep 28 '13 at 15:28
  • @Athom: Yeah I'm agreeing with you just providing clarification. Even for example legally installing Mac OSX may not impossible if some jurisdiction doesn't allow tying a license to particular hardware. – Ben Voigt Sep 28 '13 at 16:41
0

My preference is something like "Currently it's not possible given the constraints you specified, (i.e. without jailbreaking your iPad)

However, some alternative solutions might be (this app which doesn't do it exactly how you want it but pretty close)

Otherwise the only current way for the iPad to froob without frozzling is to use the defrozzler tool on Cydia. Apple does not provide any official api or library access for developers to that particular function

0

I definately agree with @Journeyman Geek, that is what should happen. You should get told why it doesn't work, for example, it might not be implemented yet.

Saying something is impossible, straight out, is just a pain. When answering, "It's impossible" should sound more like "It's not possible in your situation - it has not been implemented/too damaged/no program that can do that/will mess up other aspects of your item/file.

It's just generally better for the asker, who wants to learn as much as he/she can on the subject, as they learn why it isn't possible, and they might find a loophole in some circumstances.

You must log in to answer this question.

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