I know you can't comment +1 any more on stack overflow but a common phase is to complain "it's not working" in the question's body, when this doesn't say anything. The OP always has to be asked; What's not working?, what did they try? and what did they expect to happen? If the phrase is not banned, they could be prompted to rephrase the statement with something more useful.

Also, can we insist that questions contains a ? When there is no question mark, often it's because it's not a question which cannot be given a definitive answer and it will quickly be down voted and finally closed.

It doesn't have to have a ? in the title, but if there is no ? in the body, it probably hasn't be phrased as a question anywhere. If it has been phrased as a question it should be easy to add a ?

  • 3
    "It's not working. When I click X, Y shows." — would you ban that? – slhck Jun 3 '15 at 6:49
  • possible duplicate of Why are question marks not mandatory in titles? – slhck Jun 3 '15 at 6:50
  • @slhck The question would be; what do you need it do? I would have it suggest "When I click X it Y shows. How do I get it to Z?" – Peter Lawrey Jun 3 '15 at 7:30
  • @slhck I agree that ? in the title is perhaps unneeded but if there is no ? in the body, it has a high chance of being closed. – Peter Lawrey Jun 3 '15 at 7:31
  • But if the question already mentions what the OP wants to accomplish, then there'd be no need to add it in a comment again. It's really complex to come up with a system that automatically detects a comment that doesn't include the required information. – slhck Jun 3 '15 at 7:33
  • As the duplicate link states "The question body should contain the actual question." – Peter Lawrey Jun 3 '15 at 7:34
  • W.r.t. "?" in the body, you could still phrase a question that ends with "I would like to know why the foo doesn't bar." Again, false positives. You'd do more harm than good trying to ban this. If the question is really not a question, vote to close it. – slhck Jun 3 '15 at 7:34
  • @slhck again, that's a bad question, often without an definitive answer. The reply is still, what did you expect to happen? Often this sort of "Question" is usually; "I expected Java to do something which doesn't make sense and wouldn't be useful, why doesn't it do that?" – Peter Lawrey Jun 3 '15 at 7:35
  • And again, the OP could have stated that before. I was merely giving an example. "I'm trying to add a DOM element using this JavaScript code, but it doesn't appear. Given this code, I'd like to know why it doesn't work." Perfectly valid question. Without a question mark. Perhaps you can show some examples of comments/questions that should've been banned? I really think this is a non-issue. – slhck Jun 3 '15 at 7:36
  • @slhck In that case you should change "Answer" to "Reply" and not block questions which don't have a definitive answer. In any case, "I'd like to know why it doesn't work." is replied with "Because your code is broken and based on false assumptions" It's still a non-question. – Peter Lawrey Jun 3 '15 at 7:39
  • @slhck you can't provide a definitive answer, unless you make assumptions about what the question really is. Posters should get into the habit of asking actual questions because too often when you make such assumptions you end up answering what you think is the question when that is not what the poster assumed. – Peter Lawrey Jun 3 '15 at 7:41
  • The implied / expected answer in such a case would not say "Because your code is broken", but also explain what in the code is broken and what can be done to fix it. Anything else would be extremely unhelpful—almost rude—and if the majority of people were really that literal in interpreting questions, I wouldn't want to be part of this community anymore. If a question is unclear, vote to close as unclear. Simple as that. – slhck Jun 3 '15 at 7:44
  • @slhck computers are literal and often people write code based on their intent rather than what will get the computer to do what they want but you find again and again that if there is no clear question, there is no clear answer so as you say the best option may be to close it, though I find that rude and discouraging to new posters and they might not come back and not learnt anything from it. – Peter Lawrey Jun 4 '15 at 5:54

It's perfectly fine to construct a question that goes like this:

Title: Achieving X with the Foo language

I am trying to do X.

I've already tried Y and Z.

I'd like to know why it's not working and what can be done to achieve X.

(of course, with much more detail)

There's no need for a question on Stack Exchange to contain a literal question mark. In fact, on some other sites, such as Code Golf or Software Recommendations, we don't even expect there to be an actual question (in the grammatical sense) in the first place.

Such a filter, enforcing at least one "?" in the question body, would likely result in many false positives. And even if you blocked the posts containing no "?", the OP would simply add a "?" at the bottom of the post or somewhere in between. It wouldn't improve the quality of the actual question in any way.

If you believe a question cannot be answered for the lack of information, the best option would be to vote to close it.

| improve this answer | |
  • While valid, I have never seen a question structured this way, but I have only answered ~11K. ;) You appear to be allowing an edge case when the vast majority of poor questions are not phrased as questions and thus have no definitive answer. – Peter Lawrey Jun 3 '15 at 7:45
  • Your last suggest would leave me voting far more answers as closed, which is a bit late for many users. Better to encourage a coherent structure as they write it. What might be an alternative is for me to add what I think the question should be. +1 – Peter Lawrey Jun 3 '15 at 7:47
  • 2
    This seems to be a real problem for you, so could you please provide examples of such poor questions that do not include a literal "?"? I'd be curious to see them. I'm pretty sure that 99% of questions I close as "unclear" actually include a question or at least an implied one. It's more for the lack of details that we have to close them. And that you cannot teach with a filter. – slhck Jun 3 '15 at 7:48
  • I mean, take this one. If the OP was forced to change "how do I solve it." to "how do I solve it?" it wouldn't make the question any better. – slhck Jun 3 '15 at 7:50
  • Good point. Perhaps there is a better way of handling them. (Apart from ignoring or closing them) let me get back to you when I have some more time, off to work. – Peter Lawrey Jun 3 '15 at 7:50
  • true, but it should be trivial, if annoying for them to do. The next time they might not even think about it. – Peter Lawrey Jun 3 '15 at 7:51
  • But I agree, the real issue is false negatives (Valid requests without an ?), which would take more time to check. – Peter Lawrey Jun 3 '15 at 7:56
  • 2
    Hm. Those are indeed quite bad questions. But again, if we blocked them and told the OP to add a question, it would most probably be "help plz=??!!" or something similar. We've had similar p0rblems when we banned the word "problem" in question titles. – slhck Jun 3 '15 at 7:57

The current checks aren't very complex. That is by design, since it prevents false positives. The +1 is a good sample. It does only check on the start of the comment. That prevents matching to I tried 1+1.

In your case, the It doesn't work, should probably replace with X doesn't work, where X is the name of a product, technology, etc. The problem is, when you try to match that, you need complex rules that often give false positives. The last thing we want is people getting discouraged asking questions because of the title matching algorithm giving false positives.

Therefore, I do get the point, and I think its concept is useful, but very hard to implement.

| improve this answer | |
  • The last thing you want is a significant percentage of questions which can be definitively answered because they are not structured as questions and thus have no answer. yet this is what happens. – Peter Lawrey Jun 3 '15 at 7:49
  • But this feature request isn't the solution to that. – Patrick Hofman Jun 3 '15 at 7:50

You only have to go as far as "Problem" to see bans on general words or phrases do not work.

And this is just the variations I felt like trying. There are problem hundreds more different variations caused by the same porblem problem. Plus, who knows how many hundreds (or thousands) of these workarounds have already been edited out by other users.

This shows that rather than actually improving anything when they encounter the blacklist, people will just work around it. And you end up with a lot of ugly workarounds that are probably as bad or worse than the original rule you tried to enforce.

Then you have to deal with the false positives. Can there be a questions where it is entirely appropriate to not to have a question mark? Absolutely! And I am virtually certain I can come up with an excellent question that contains either "it's not working" or "it doesn't work".

And let's not forget that there are still 46,802 questions with problem still in the title. These questions cannot be edited unless someone changes the title to remove problem. Adding new banned phrases or keywords would make another obstacle to people who want to edit the post as they now will need to figure out how to rephrase the question.

The point is banning those words or enforcing a question mark may seem like it will help with question quality, all it will do is annoy users who plan on using those phrases appropriately while at the same time most of the people you are targeting won't actually fit it, they will just work around it.

(The counts above don't quite add up as some of them above are duplicated in a couple of different searches, but I still think it gets the point across).

| improve this answer | |
  • 1
    The ban on "problem" worked beautifully. You found less than 500 examples where people misspelled the word to get around the ban. Prior to that there were over 46,000 titles that contained it. Hundreds of pages of meaningless titles. – Bill the Lizard Jun 3 '15 at 15:56
  • 1
    @BilltheLizard Yes, but how many of the work arounds were edited out? My point is it doesn't change user behavior, just causes them to work around? – psubsee2003 Jun 3 '15 at 16:19
  • Do you think it was edited out over 45,000 times? There's very little evidence to support the idea that people just work around the restriction, and 100 times as much evidence that they change their behavior. – Bill the Lizard Jun 3 '15 at 16:53

Can you add "it doesn't work" and "it's not working" as banned phrases?

Where do you stop?...

it doesn't work
it's not working
It is not working
It just won't work
I can't get it working
Working it is not
It won't work
It does not work
Why won't this work
Why is this not working
I've tried but it does nothing
It just outputs nothing
I get nothing
It does nothing
There is no output
My code doesn't work
This code doesn't work
Dammit Jim I'm a Front End man, not a Programmer
My code gives blank screen
This following code does not work
Why won't the following work
Captin, we dee nee haff tha poower
What is wrong with my code
It's broken
How do I fix this

And so on...

A basic/rough idea of the main requirements to add something to a blacklist are:

  1. It's a very frequent issue
  2. It's worth the resources (scripts to check every question, management, monitoring, fixing false positives, etc)
  3. It doesn't create issues/false positives
  4. There is no easy way around it making the "fix" obsolete/pointless

"it doesn't work" and "it's not working" fail on 1 (it's not that frequent).
They fail on 2 as stopping them likely doesn't make the OP suddenly provide loads of info, they just use another generic phrase, or leave it out and Submit.
They fail on 3, as there are genuine cases to write "it doesn't work" - i.e. after a full description of the issue and code stating "it just doesn't work".
They fail on 4 as per the start of my answer.

I can see you are trying to make improvements which is great, but I just don't think this particular avenue is worth bothering with.
Even putting all resources required aside, if it took 10 seconds in total to do, it's still not a good thing to do.

| improve this answer | |

You must log in to answer this question.

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