-14

A simplified, improved and feature-requestified version of this answer -

What if SO required new questions to achieve a net score of +1 before they can be answered in order to improve quality control?

  • New questions would get a banner to the effect of

    Your question is now under peer review. Once it has a net score of +1 votes, it can receive answers.

  • Once a net +1 has been reached, the question is "unlocked" forever. It doesn't go back into "unanswerable" state if it gets downvoted after unlocking.

  • Higher-rep users would be able to "rescue" a misunderstood and downvoted question by investing a 50 bounty, which they could do without having to wait 24 hours. Putting a bounty on the question would immediately unlock it and make it answerable.

  • Questions that are not answerable cannot be searched.

  • Questions that do not manage to become answerable after a certain time will be automatically deleted.

  • The threshold could be limited to high-volume tags that are ridden with bad questions, for example , , , and .

This is fairly radical, and would make SO feel much less welcoming towards non-perfect questions than it currently still is. But maybe that's a good thing? The only way to really grab an asker by the balls and force them to improve their question is to not provide any answers until it's done.

Side effects:

  • Voting would get an even more crucial role than it currently has. Voting would have to be encouraged even more.

  • Voting behaviour would become even more slanted towards upvotes, because an upvote is required to make a question live at all. To me, this is not a good side effect, but the expected advantages still outweigh it.

  • It would become impossible to answer downvoted questions, something that sometimes makes sense - but in my eyes, so rarely that it's justifiable to lose the possibility. Also, once a question has been unlocked, you can still -1 it after answering it, which is perfectly okay (because the question doesn't go back into "unanswerable" state).

What do you think? I'm not 100% sure whether this would be an improvement, or too much - but as I said in the original contribution, I find this an intriguing idea and definitely worth thinking about.

16
  • 2
    Definitely neither necessary nor desirable for tex.sx.
    – doncherry
    Oct 9, 2011 at 11:23
  • 1
    @doncherry as said, this would be for Stack Overflow only, which is receiving 4k new questions a day.
    – Pekka
    Oct 9, 2011 at 11:24
  • 2
    It would also be interesting to see some data: How many questions below a 0 score have good answers? ("Good" being defined as "+1 score or more")
    – Pekka
    Oct 9, 2011 at 11:24
  • 1
    You might have a problem with people up voting a 0-score question and then removing it a (very) short time later (for random malicious reasons). You would have to wait for a period of time or until the vote is "locked in".
    – flight
    Oct 9, 2011 at 11:42
  • 13
    I am a bit torn. Users still don't vote that much (from my experience on sites where I sometimes post questions), and you certainly can't get them to vote on questions they don't care about. If I post a perfectly good question under a few obscure tags, why should I have to wait until somebody comes along and takes me out of the 0-score limbo? How would you train (or even force) users to upvote good questions?
    – slhck
    Oct 9, 2011 at 11:55
  • @shlck yeah, that is the biggest challenge in this, for sure. You would have to communicate that your vote is what gives a question a chance at all. I think that is possible, because you would see so many questions you can't answer (when you try, you would of course see a popup explaining the situation.) I can see the argument that this would make things more difficult for obscure tags, but I'm fairly sure even the most "niche" question gets more than 1 view from a user who can upvote.
    – Pekka
    Oct 9, 2011 at 11:58
  • Inspired by @Mat and slhck, I refined the suggestion to include only high-volume tags.
    – Pekka
    Oct 9, 2011 at 12:05
  • @Pekka: I disagree less now :) How do you decide which tags get this thing? High closed/question ratio? High negative/positive overall score ratio? Some fixed list? (What if say some of the facebook tags get picked up by this, any problem with the "partnership", or future partnerships of this type?)
    – Mat
    Oct 9, 2011 at 12:18
  • @Mat I'd say a fixed list would work, but a stats-based approach might also work well (= tags with the most downvoted and closevoted questions). Re how this would work in partnerships - good question, but that would have to be solved individually I guess
    – Pekka
    Oct 9, 2011 at 12:26
  • Now it definitely makes more sense, at least for the high volume tags. I'm not too much on SO to be able to judge if that's the best method of dealing with this though, so I won't down- or upvote yet :) Seems like a good idea in principle, but it's a very radical change.
    – slhck
    Oct 9, 2011 at 12:52
  • I don't mind the downvotes at all - it is a terribly radical suggestion - but what I don't entirely get is that the exact same suggestion, when posed as an answer, got 16 upvotes and no downvotes?
    – Pekka
    Oct 9, 2011 at 13:32
  • Maybe people liked the phrase "sandbox" more? (Plus, downvoting answers still costs rep, so …)
    – slhck
    Oct 9, 2011 at 13:44
  • @slhck yeah, I thought the same thing! I might try asking the same thing in a couple of weeks and call it "sandbox" again. We'll see. :)
    – Pekka
    Oct 9, 2011 at 13:48
  • 2
    @Pekka: if you repost, I hope you get a special Your question is now under peer review. banner just for you (:
    – Mat
    Oct 9, 2011 at 14:02
  • @Pekka웃 re:"How many questions below a 0 score have good answers?" I can think of two good examples. First, something that is a great question but asked poorly. It might also have a great answer and therefore value. I'll bet that is pretty common here?
    – DHorse
    Jul 21, 2013 at 10:51

4 Answers 4

10

I disagree. The only 'problem' this solves is the FGITW problem.

Quite simply, this means a major change in the way I would use SO, and it is a change I won't make. When I am at work, I browse SO looking for questions I can answer while I wait for a build to complete so I can debug something. If you implement this idea, I would not be able to browse questions anymore. Think about it this way:

  • I finish a bit of code, and press F5 to start it building
  • I switch to my browser and hit F5 on the SO page
  • I start reading down the list of questions until either:
    • my code has built and I have to switch my attention back to my real job
    • or I see a question that either interests me or I can answer it
  • now I have to wait till the question has been upvoted by someone else before I can answer it

The problems with this flow are:

  • I have limited time to browse questions - once that question is gone from my radar I am not going to go back and find it. This means that person potentially loses the answer they were looking for.
  • I have enough rep, and a hell of a lot of experience in this game - why should I need to wait for someone else to upvote the question before I answer it?
  • questions have to be 'double handled', IOW they need to be upvoted by one person then answered by another, this is just naff. If I really want to come back to a question I need to bookmark it.
  • people may no longer get fast answers, this can be quite crucial in a job situation
  • questions that do not necessarily deserve an upvote will get them
  • just because a question doesn't get upvoted is no reason to not answer it

We already have a system for handling bad questions - it's called a down vote, and it's FREE. The problem is, I see a lot of questions flagged with the NARQ or VLQ flags, but not downvoted - and I ask WHY?? A person with enough rep to flag will usually also have enough rep to downvote, so why don't they? Why are people not effectively using the system we already have, why do we need to come up with a new solution when the current one is simple and should work?

8
  • Why do we need to come up with a new solution when the current one is simple and should work?- because it doesn't work :) But apart from that, am I correct in understanding that the problem you have with this is that you can't answer a question you feel is good? That's an interesting view. Maybe the "your own vote doesn't count" restriction needs to go, or be imposed only to users up to 500 rep or so. Hmmm.
    – Pekka
    Oct 9, 2011 at 12:09
  • I can see your point about not wanting to wait to be able to answer. I removed the "your own vote doesn't count" limit. That allows for different ways of abuse but they may be easier to deal with than this
    – Pekka
    Oct 9, 2011 at 12:13
  • 5
    Definitely true about people flagging rather than using the other tools they have available (down-votes and votes to close).
    – ChrisF Mod
    Oct 9, 2011 at 12:17
  • @ChrisF yeah. As to why - flagging makes an arbitrary number go up. Downvoting and closevoting don't. I think that's the long and short of it :)
    – Pekka
    Oct 9, 2011 at 12:25
  • @Pekka, just removing that self-vote restriction still doesn't quite cut it - just because I (or anyone) answers a question doesn't mean it deserves an upvote. Only good questions should be upvoted.
    – slugster
    Oct 9, 2011 at 12:50
  • 1
    @slugster you would be able to upvote a question to unlock it, then remove the upvote after answering or even -1.
    – Pekka
    Oct 9, 2011 at 14:25
  • @Pekka If answering takes longer than five minutes, wouldn't the upvote be locked in?
    – John
    Oct 9, 2011 at 20:59
  • upvote and then answer? If it's worth answering, it's worth an upvote?
    – Rosinante
    Oct 10, 2011 at 0:01
5

I don't think that will work for the low-volume tags.

Say an expert in a niche tag sees a good question, but the question is neither obviously good nor obviously bad to the "general public", so no one else votes on it1. The expert casts his vote, and you have to wait for:

  • said expert to post a bounty
  • another expert on that tag to answer
  • someone less knowledgeable to posts an answer

The last case isn't bad at all, but it means that if said expert doesn't actively monitor that question to check the answer, or post his own, we might be missing great posts. That's just not a good workflow IMO, too much chances of missing out on good answers.
Low volume tags are easier to follow than the top ones, but still I don't like this.

Worse, what about experts that don't have the rep yet to either up-vote or post a bounty? How do they enter the community on low-volume, "specialist" tags?

I'm all for finding a way somehow to have low-quality questions vetted one way or another, but I don't think this proposal works for all the tags. I think we risk losing out on great, timely answers.

1Or worse, people down-vote it because it doesn't appear to make sense for "the uninitiated" in whatever technology we're talking about


Counter proposal: have sub-options for the Not A Real Question close vote (like there is for the Off Topic one).

You could have say:

  • Really not a real question => normal close semantics
  • Missing code sample to make it a Real Question
  • Missing error message
  • Missing appropriate language tags
  • ...

For the new NARQ cases, display a nice pre-formatted message listing all the reasons why the question is "not accepted", with links to the various resources already available to make the question better.

Questions closed this way could have a lighter re-open requirement, and/or a lower number of close votes requirement.

Maybe get a special tag or queue somewhere so people interested in mentoring can jump on them and help out. People who get irritated with this type of thing could just ignore that tag/not follow that queue or something.

The fate of questions stays regulated by community members that have shown to be trust-worthy by gaining sufficient reputation to vote to close.

I'm not certain there aren't fatal flaws to this, but I find it less "dangerous" than a solution based only on votes.

3
  • Good points about the low-volume tags; this could be limited to high-volume tags, I'll edit that in, thanks. Re experts not having enough points to vote - the vote threshold is 125 points, that is easy to reach for an expert in any tag.
    – Pekka
    Oct 9, 2011 at 12:03
  • @Pekka: yes, 125 rep isn't all that hard, but you have to be able to answer questions to get it. If the questions are "not open for business"... hard to get the rep.
    – Mat
    Oct 9, 2011 at 12:10
  • I changed the suggestion to cover only high-volume tags with massive quality problems. It's no problem to get enough upvotes for a good question there
    – Pekka
    Oct 9, 2011 at 12:11
3

Firstly, I sincerely welcome such ideas. I frequent PHP tag and often see a question has a negative score when it receives an answer.

However, this is not totally dissimilar (like a reverse) to a previous idea of mine - Place questions with X net downvotes on hold automatically.

The issue with your proposal - sorry :(

A major problem with my proposal is unfortunately also true of yours:
It changes the entire use of the voting mechanism.

Instead of voting being based solely on question quality - good/bad, interesting, etc - instead:

  1. Users downvote to close a question
  2. Users upvote to keep a question open
  3. Users don't vote because of 1 and 2 - they don't want to "vote to close or open"
  4. Users who answer poor questions would upvote to keep it open so they can answer
  5. Users with 125 rep can downvote, which means users with 125 rep can now "vote to close"
  6. More/etc

These things pollute one of the major aspects of the site, and why the sites work well - good answers and questions rise to the top, bad ones are pushed out of view.
Because votes are no longer just based on "good/bad".

Because of 1, 2, 3, and 4, the voting becomes a battle between closing a question and keeping it open.
This, if nothing else, means the poor OP get's all kinds of additional or negative rep and not based on their question, but other users believing it should be open or closed.
This is not why someone should receive or lose rep.

2 main issues with bad questions

1)
Bad questions actually receive answers.

If bad questions did not get answers I believe we would most certainly thin the issue down a great deal.

A bad question having an answer:

  • Gives the impression Stack and it's users are happy for this to happen
  • Hinders good users trying to educate users with a bad question "how to improve" for the greater good of the site. As an upvoted answer gives the impression it's allowed, and it's just a few of us who do not welcome such questions. One user stated to me "I have 3 answers, so this question is actually welcomed on this site".
  • Is actually allowed! By giving bad questions an answer or two, we do allow it, it happens, we are not stopping it!

2)
Bad questions are simply not closed quickly enough

By the time close votes get reviewed and actioned, a bad question (usually) has already gained an answer or two.

Also, people answer quickly to avoid not being able to answer by the time the question is closed, and then edit their answer as needed.
Their knowing to post something quickly before the question is closed means they too know this question shouldn't have an answer.

So in addition of quickly locking down bad questions and stopping them from getting answers, we also need to address people who do answer bad questions.
Not penalise them, as they might just need some friendly advice on why not to answer such questions.

Idea to resolve

I have an idea/ideas on how to possible fix the above 2 issues, and work around the problems using the current voting mechanism brings.

I should post it really, as a feature-request.

We really do need to get poor questions locked down really quickly to avoid answers coming through.

1
  • 1
    These are good points. Do link to your idea if you turn it into a feature request!
    – Pekka
    Apr 16, 2015 at 20:35
1

Answer from the future! 3 and a half years later...

It's still a "no" from me. The thing is that some of these questions are really just staying at 0 and will not go down or up. Clearly, the developers have lessened their immense iron grip but this proposal will just give them a major iron grip. A lot of questions are just really easy to answer right away according to a lot of users here. The thing is, a question from a highly-ranked user mght just downvote yet the community says so otherwise.

Disadvantages:

  • Community doesn't really do anything - The quality of the question in question should be determined by the community, not just a single user. Two or three of them will still not work. We have lots of review queues already. Questions that are bad will be sent to the Low Quality Queue for experts and users of all types can determine whether it stays or go.

  • Flame wars - Lots of people are going to have a lot of arguments about the decision made. They would think that the users aren't letting them to the one thing they want to do on these sites: Solve their problems. We already got enough arguments, and more shall erupt. More mods will needed, and chaos will rule SE. Not a good thing.

  • All questions start at 0. Saying that sites like Stack Overflow will have no problem with the entire squad of questions it receives each day is a major overstatement. We should put a threshold for users under x rep to put this request on. Many users know what to do and how to write a proper question anyways. Users will spend forever reviewing each single question while users wait for a valid on to pop up. Traffic in the queues - 1000% or more. Traffic on the site as in questions and answers - -50% or less.

There are many more disadvantages but these are the major disadvantages I should address. So for me:

1
  • @Pëkka, how about now? Apr 13, 2015 at 23:30

You must log in to answer this question.

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