7

More than infrequently, opening posts on SO are not questions at all, in the very grammatical sense: They're just statements or vague musings (for instance this recent one), and it's not clear what exactly is being asked (since nothing is technically being asked).

I always feel that it shouldn't be our job to construct a question from someone's undirected ramblings, and so this is usually a NARQ candidate for me.

I was wondering if we could add a very simple check that would count the number of literal question marks in a post and refuse to accept it if there are none. This could be accompanied by a small note on what a question should look like.

What do you think? Should we add this feature?

Update: Since people seem to get hung up on the details, I'd like to clarify that an entirely satisfactory alternative would be to pop up an overridable notice with a warning, like "Hello, you appear to be writing a question. Would you like help?", with a "No thanks" button to dismiss it and submit the post regardless.

Update 2: If you disagree that absence of a question mark is a strong signal that the question is poor, could you please provide examples of good, useful and appropriate questions that do not contain question marks? I've provided several additional examples of when the signal works in the comments, and more come up regularly, so I would appreciate to see some evidence of the contrary.

28
  • 4
    I don't think we should refuse such questions (without any question mark), but we could additionally warn user (popup or something) before submitting that this question are likely going to be closed as NARQ.
    – om-nom-nom
    Commented Mar 25, 2012 at 14:36
  • @om-nom-nom: Yeah, sure, a non-blocking warning would also be OK... whatever fits in most nicely with the infrastructure.
    – Kerrek SB
    Commented Mar 25, 2012 at 14:37
  • This could easily be gotten around by dropping in a block of code that uses ? as an operator. Commented Mar 25, 2012 at 14:39
  • 1
    @BoltClock: I will personally answer all (non-)questions that use the ? operator if that's your only objection :-)
    – Kerrek SB
    Commented Mar 25, 2012 at 14:40
  • 1
    @BoltClock count only plain text (no code and no quotations)?
    – om-nom-nom
    Commented Mar 25, 2012 at 14:40
  • 13
    I count 3 question marks in your post and its title. Which of those indicates the feature you're requesting?
    – Arjan
    Commented Mar 25, 2012 at 14:45
  • @Arjan: I'm afraid I don't follow. The question marks in my text are part of the text. The feature I'm asking for is described by the text. Is any part of it unclear?
    – Kerrek SB
    Commented Mar 25, 2012 at 16:18
  • 5
    I was trying to show that looking for a question mark does not solve the problem. 1) Your title does not summarize what you're asking for. (And the title doesn't describe a feature request to me.) 2) The last paragraph does not add anything to the post. 3) In general, we'd like one question in a single post on a Q&A site. So, though of course I understand what you're proposing: I could say that the above 3 question marks are just a way to circumvent the very quality filter you're asking for, and might even make things worse when folks add unrelated sentences with question marks.
    – Arjan
    Commented Mar 25, 2012 at 16:54
  • 2
    @Arjan: But Meta isn't the same as SO. On Meta I can reasonably say "We need feature X." On SO there's no equivalent. And also, while you could of course go on finding edge cases where my suggestion doesn't help until the cows come home, my point is that for a large majority of cases the proposed warning would help the potential poster to create a more successful question. The question title describes the underlying cause; the question body suggests a way to deal with it.
    – Kerrek SB
    Commented Mar 25, 2012 at 16:56
  • Here's the third candidate in the space of a few hours, just in C-or-C++...
    – Kerrek SB
    Commented Mar 25, 2012 at 18:47
  • 1
    I don't know if it helps anything but I do enjoy asking people to amend their question to include an actual question, complete with a question mark, to indicate specifically what they would like answered. It's a time when I wish we had delayed-close votes that evaporate if the question is edited to save me the effort of returning.
    – sarnold
    Commented Mar 25, 2012 at 23:10
  • 1
    You really think, that amount of question signs is a sign? I can easy write perfectly correct question without any question mark Commented Mar 26, 2012 at 7:44
  • @LazyBadger: Simple observation has lead me to believe that the vast majority of questions that lack a question mark are also lacking in many other departments. I know that there are always some cases where the heuristic doesn't apply, but I think it would overall be a very useful feature.
    – Kerrek SB
    Commented Mar 26, 2012 at 7:57
  • 1
    @KerrekSB - I, personally, don't think so and doesn't support idea "think instead of user". But local looosers disagree with me Commented Mar 26, 2012 at 8:39
  • ...and here is another candidate.
    – Kerrek SB
    Commented Mar 26, 2012 at 20:28

3 Answers 3

6

No, I don't think we should be doing it.

It's possible to describe a problem and even a question with no question marks. "How do I do that?" Becomes "I'm currently looking for a way to do that".

The problem is usually with beginners (identified by low reputation or recent join date) not realizing that this isn't a forum, but a Q&A site, and that the problem should be very clear.

Perhaps a notice before posting for low reputation users, but I don't feel we should completely reject posts without question marks like that (Because then I'd just add a bunch without changing the content and be done with it).


P.S. @LazyBadger, before you comment, yes, this is essentially the same suggestion as yours, but your reasoning and mine are different, I disagree with your reasoning. Users are expected to meet our quality standard, not the other way around.

4
  • Mm. Well, I don't think I've ever seen a question-that-wasn't-one that was good or valuable, and I've seen many posts-that-aren't-questions that were completely worthless and needed serious work. Everything is telling me that this is an easy way to automate quality control...
    – Kerrek SB
    Commented Feb 22, 2014 at 17:53
  • @KerrekSB: Don't forget that the moderation approach Stack Exchange took was the after-fact approach. You can try to prevent low quality content before it happens, or do it after it happens. Stack Exchange chose the latter. Not everything needs to be automated, and I have seen questionless questions which were good and on-topic. I don't deny that there are many many out there which are not, which is why I suggested a notice instead of absolute rejection. Commented Feb 22, 2014 at 18:34
  • Sure, it doesn't need to be absolute - but something noticeable that pops up and says "please think about this" might go a long way!
    – Kerrek SB
    Commented Feb 22, 2014 at 18:38
  • @KerrekSB: I agree there. Commented Feb 22, 2014 at 18:43
-2

StackOverflow isn't Jeopardy. Questions don't necessarily need to contain question marks to be understood and answered. I also think that questions shouldn't need to contain question marks to be postable.

If we turn them away completely by saying: "You didn't include a question mark so you can't post here", we're essentially saying "We don't want your kind here".

If instead, we allow new users who do not understand the system to post their "questions" and have community members point these users in the right direction, the system is more inviting, and gives the impression of "Thanks for giving it a go, we're all here to help you".

Certainly some users will never learn from their mistakes, and keep posting poor quality questions, but I would hazard a guess that they're in the minority. Questions that are good enough will be upvoted, and if there are improvements to be made that are worth making, members of the community are generally pretty good at stepping in and lending a hand. It's why SO is so useful. People actually help.


Addressing Update 2:

If you disagree that absence of a question mark is a strong signal that the question is poor, could you please provide examples of good, useful and appropriate questions that do not contain question marks?

I hold my judgement on whether there is any correlation between the absence of a question mark and the quality of the post. The onus is on you to prove that there is a correlation via evidence. In the absence of any evidence, I cannot say one way or another.

I've provided several additional examples of when the signal works in the comments, and more come up regularly, so I would appreciate to see some evidence of the contrary.

You have provided anecdotal evidence, which neither strengthens nor weakens your argument. Asking for more anecdotal evidence is not an appropriate way to proceed in bolstering any argument.

If it's simply evidence that you want, it would be appropriate to query the SO archives for the data:

All questions should be broken down by two categories into four groups based on their first draft (remember, this whole issue is to address the initial posting of a question, not the subsequent revising).

The categories are:

  • Questions with first drafts that contained question mark characters vs those that did not
  • Open questions vs Closed questions

The groups are:

  • Questions with first drafts that contained question mark characters (?) and are still open
  • Questions with first drafts that did not contain question mark characters (?) and are still open
  • Questions with first drafts that contained question mark characters (?) and are now closed
  • Questions with first drafts that did not contain question mark characters (?) and are now closed

If you can provide evidence* that the lack of a question mark character in a first draft strongly correlates to the question being closed, then I will be more than happy to delete (or modify) this answer and create a new answer that supports your proposed change.

* Please also be sure to provide the steps you used when gathering the evidence so that we may reproduce it

11
  • 4
    Out of curiosity, could you point out some good questions that aren't questions?
    – Kerrek SB
    Commented Feb 22, 2014 at 20:55
  • 1
    @KerrekSB, actually you're right, I should rephrase my answer. I did explicitly state that "well-written" questions don't need to contain question marks, but really my intention was to say that questions shouldn't need to contain question marks to be postable. As it is there's a lot of friction for new users to get used to the system (because it's different from forums and facebook and anything else new users might be familiar with). Reducing the friction for new users helps us expand and train our new user base. It certainly makes more work for the community, but it's a sign of growth.
    – zzzzBov
    Commented Feb 22, 2014 at 22:24
  • 1
    @KerrekSB, also, a "question" could have all the relevant information listed as a statement. Statements like "I've been trying to use [concept] in the provided code. I do not understand [concept]. Please explain how [concept] works." are perfectly valid ways of asking for help.
    – zzzzBov
    Commented Feb 22, 2014 at 22:31
  • OK, sure - I'm looking forward to seeing an example of a postable post that contains all the relevant information! Will you share one? And reducing friction is my goal, too -- I simply believe that we have here an very reliable signal for when an automatic system could provide early and useful user education and guidance.
    – Kerrek SB
    Commented Feb 22, 2014 at 22:43
  • @KerrekSB, the community liked this one, hold on while I look for more
    – zzzzBov
    Commented Feb 22, 2014 at 22:50
  • That's just a typo - the OP says "Can someone explain", which is certainly a "proper question" but forgets the question mark :-S A gentle hint would probably have helped make that one even better. (The question also got closed as being inappropriate for this site.)
    – Kerrek SB
    Commented Feb 22, 2014 at 23:05
  • @Kerrek FYI, this is proper English: "I would like to know how to fix this problem." Another example: "Please, help me find a problem in this code." The rules of English dictate no question marks in these sentences. Commented Feb 23, 2014 at 14:55
  • @AndreiVolgin: I realize that, and the user should certainly be allowed to post if they insist. What I say is that I feel that absence of a question mark is a very strong signal that the question is poor, and I invite you to show some examples of (not closed!) questions that don't have one and that you think don't need improvement.
    – Kerrek SB
    Commented Feb 23, 2014 at 15:20
  • Your proposal will force people who ask good questions to start using question marks just to avoid an annoying popup, but it will not do much to reduce the number of bad questions, most of which already include a question mark. Commented Feb 23, 2014 at 15:30
  • @AndreiVolgin: Please provide some evidence of the damage you think this would do, i.e. "good questions" that don't currently have a question mark. I think the negative impact is much less that you might think. And it would be overridable anyway, just one extra click.
    – Kerrek SB
    Commented Feb 23, 2014 at 17:27
  • "I invite you to show some examples of (not closed!) questions that don't have one and that you think don't need improvement" you've just changed your requirements, and it sounds like you've misunderstood my answer. I'm not saying that questions without questionmarks don't need improvement. I'm saying that it should be acceptable to post a question that needs improvement because StackOverflow uses a wiki format that encourages changing, adjusting, correcting, editing, and revising. Users should be allowed to post a crappy first-draft, and then encouraged to update.
    – zzzzBov
    Commented Feb 23, 2014 at 19:01
-13

It can be notice, but must not be refusal. And, after all, "don't dictate me, how I write my texts!!!". Strong and strict management needed only for small kids, which isn't our case, isn't it?!

Count my voice "contra" idea in your current form.

8
  • 2
    don't dictate me, how I write my texts! We do not dictate a good style (that includes grammar and spelling)...we imply, need and partly demand it. Commented Mar 26, 2012 at 8:17
  • 1
    @Bobby - refusal isn't good style. And downvoters-assholes want refusal?! Or they just understand me? Give them brains! Commented Mar 26, 2012 at 8:31
  • Small kids or new users who are not yet aware of the high standards kept by the current community...
    – Lix
    Commented Mar 26, 2012 at 8:43
  • @LazyBadger: No, it isn't. But sometimes it's all we can do. Commented Mar 26, 2012 at 9:17
  • @Bobby - just don't do it?! If you want to be nice even for idiots, you'll be nice only for idiots Commented Mar 26, 2012 at 9:21
  • 1
    @LazyBadger: Okay, now you lost me...what were we talking about? Commented Mar 26, 2012 at 9:23
  • 1
    @LazyBadger: Actually, OPs proposal seems to be the exact opposite. If there's no question mark, refuse it. That would keep some bad content from entering the system. Though, there's not much hope that it gets any better if it finally enters the system...but it's an idea. Commented Mar 26, 2012 at 9:42
  • @Bobby OP proposed to serve every stupid idiot, which not able write good content (-1 from me 1-st), use refuse instead of (rating-dependent?) warnings (-1 from me 2-nd) and use bad, stupid algo (-1 from me 3-rd) Commented Mar 26, 2012 at 9:47

You must log in to answer this question.

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