Separate "useful" from other downvotes

One thing that I have been thinking of for a long time on the SE network is that I find it strange to have “useful” grouped together with “clear” and “shows research effort”. Especially since downvotes can result in a question ban. But also because that downvotes can feel a bit hostile.

On the other hand, I do believe that it has some value that questions that are very unlikely to help someone else disappear. But I don’t think it’s fair to punish people that come with a clear and well researched question just because it seems unlikely that it is “useful”.

I think that it’s reasonable to expect that people create clear questions and do their research. But I do not think it’s reasonable to expect people to think about if their questions are useful to others or not.

In my opinion, downvotes should be reserved for cases where posters really have themselves to blame to some extent.

I’m not sure how to solve it in the best way. What do you say?


We can’t, of course, control how people actually vote, but we can and should write guidance – tooltips, help topics, etc – that reflect what we actually want. And I agree with you that specialization shouldn’t be a reason to downvote. The “not helpful to anyone else” cases I’ve seen usually have other problems too, like not explained clearly enough for anyone to be able to help or major, obvious pilot error.

Perhaps I misremember, but I thought “useful” was only a criterion for answers, not questions.

You misremember. Just hover over the vote buttons on a question on SE and it says “This question does not show any research effort, it is unclear or not useful”

OK, then I guess I just always ignored the “useful” part for questions, since it doesn’t make any sense. The site may provide some guidelines, but in the end it’s my call how I vote. I’m going to downvote crappy questions, upvote good ones, and leave the mediocre ones in the middle alone. Crappy and good are what I deem them to be by the criteria I think matter. I expect most users do the same thing.

This all says that while the guidelines should be sensible, we have to understand few people are going to take them too seriously, or bother reading them at all.

1 Like

I think “not useful” can be a reasonable reason to downvote, and it isn’t misused all that often.

There are a few ways in which someone might consider a question to not be useful:

  • It involves some issue or feature that’s too obscure.

    I don’t think people downvoting such questions is currently a huge issue on Stack Overflow. If anything, people are quite fond of them.

  • It’s extremely simple.

    People downvoting for this reason might be a problem, as 20 (or however many) years as a programmer might affect your judgement about what’s too simple to ask about.

    But there are also some cases where it could be justifiable. Stack Overflow (and I believe Codidact) was never intended to teach people programming from scratch, and isn’t particularly well-suited to doing so. If you’re asking some question about basic syntax or something that should be covered in every introduction to some language or programming in general, a good argument can be made that that doesn’t belong on the site because you should be learning that through a textbook or something.

  • It’s asked in a way that wouldn’t be helpful to others, even (especially) if there is a good question hidden somewhere underneath.

    Asking how to fix a NullPointerException in general can make for a good question.

    Dumping some code and asking why you’re getting a NullPointerException is not going to be all that useful to others, with or without the existence of a canonical post on the subject.

    You could argue “research effort” includes breaking the code down until you have a minimal example, but there are also other ways to do research effort that wouldn’t lead you down this path, this may not apply to all such questions and plenty of good questions don’t show much research effort.

Now there are a few other viewpoints I might agree with:

  • We shouldn’t punish a good question for the asker not doing enough research. Although maybe one could argue a “lacking research effort” implies it’s a bad question because, for example, it’s a very common duplicate. But then it’s probably also “not useful”.
  • (Perhaps) we should reward an asker for asking a clear, well-researched question, even if people don’t think it’s a common or interesting (i.e. “useful”) enough issue to deserve an upvote.

Although these points would need to be discussed further to see whether there’s a viable idea behind them. Maybe these don’t need to be addressed at all and doing so would do little more than make the GUI unnecessarily cluttered, complicated and confusing. Or maybe they can be addressed by simply providing good guidance in terms of how to vote or addressed in a less direct way (like considering whether and when a question gets closed).

1 Like

I have suggested previously that there should be more than 1 way to express your opinion on a given post, in addition to the usual “up and down arrows”. Like reactions with specific meanings. Like explicitly a “Useful” vote or “Clear” vote. I don’t think an emoji with tooltip text is going to look ideal, maybe the text saying “40% of voters found this post [useful/clear/other]” would indicate how well a post is received.

Steam and other entertainment reviewing platforms have “70% found this review useful” which is used to filter out low effort reviews of video games and movies. As you might imagine, lots of young people play video games and lots of amateur users who don’t take reviews seriously post them on various platforms, and it’s useful to have the ability to filter out the completely superficial posts like “lol bad game” or “it’s an okay movie, I guess”. Or even pretty clear but hard to relate to posts, which may have been written in a serious manner, but in the end are hard to understand and don’t explain well to an average review reader what the reviewer meant to say about the movie or a game (did they recommend it or what?), in which case the same text still applies just as well.

Also, it is my understanding that for each community, users will decide for themselves what types of voting options they want to have available for their posts. So maybe people who prefer the old and tested “Up/Down arrows only” system would choose to have that, and people like you and me would experiment with these extra voting options. So I am in support of having the technical support for this extended functionality for those who want to add it.

1 Like