What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

This post was inspired after reading The Help Vampire problem.

Some types of questions are just bad for this site. They decrease the signal to noise ratio, they waste my and other people's time. They are 100% certain to be useless as an online resource for Google Searching.

I'm talking about blatant Debug My Code questions, like

Why is my code not working, here's the entire wall of code, please tell me what's wrong ASAP

Do we really want those questions to get answered at all? Because they encourage those who ask them to ask again, and they gain this site a reputation that these questions will be tolerated. May be those who asked them have seen such questions get answered on this site before.

If the answer is no, then we should think about how to stop those questions. I believe those who ask them don't really care about the points system, they are not involved in this site, they , so down-voting the question itself won't help.

I have noticed that very often answers of these questions get a lot of reputation for being clever enough to spot the bug, and I think the people who answer them are doing it for the points, not because they have a desire to help those people - they probably sense that those people shouldn't get answered, but the gamification gives them incentive to answer anyway.

The answer would be to make it cost them points, not gain them, when they answer blatant Debug-My-Code questions. This cannot be done by the users alone, unfortunately, because offsetting even a single up-vote requires 5 down-votes, and it would never happen in practice.

This is why I've marked this post as a feature-request. I have no idea if it will be implemented, but I propose that a new reason for closing a question is added - "Debug My Code" or some other name, and if the question is closed as such, all answers' votes should automatically become fixed to -1. The users who answered would lose only 2 points, but it would be enough to eventually make the community stop answering those questions.

I'm proposing this only for Debug My Code questions, because they are singularly useless in every respect, they don't even help the ones asking because it just postpones them having to learn how to debug their own code.

share|improve this question
13  
Don't penalize answerers. Just make their efforts pointless by more aggressively moderating the questions. If your rep is gone because such questions are ultimately deleted, you take away the whole point of answering them. –  Bart May 15 '13 at 14:28
2  
FYI those asking such questions are penalized by downvotes in the sense that they're likely to be question banned if the behavior continues. –  Servy May 15 '13 at 14:29
    
related: Feeders, not help vampires, are the problem –  Josh Caswell May 15 '13 at 15:51
    
It should be noted that some answerers do their thing not because of rep, but exactly because they want to "help", and feel good about that. The result for the site is the same, but the only way to deter answers with that motivation is simply to prevent them -- close early, close often. –  Josh Caswell May 15 '13 at 16:04
    
@Bart meanwhile we incentivize answerers to do exactly this by way of the reversal badge, which you get for answering a terrible question and somehow amassing upvotes before it gets deleted. –  djechlin May 15 '13 at 19:02
    
@djechlin If those posts end up being deleted, they won't get the badge the next time they would have earned it. –  Bart May 15 '13 at 19:04
    
@Bart it's a badge you can earn multiple times though? Or does the sweep check for the count. In any case reversals are really rare so no one is strategizing this way. I'm thinking about proposing we get rid of this badge or something because it's inconsistent with the rest of our stance on how you shouldn't answer crappy questions. –  djechlin May 15 '13 at 19:06
    
@djechlin Make that a request then, if you think it's worth it. Not sure if it's a dupe though, so check first. –  Bart May 15 '13 at 19:08

1 Answer 1

up vote 8 down vote accepted
  1. Penalize by downvotes and closing the question. This results in a question ban (details of algo are kept secret but these contribute), which means the user can't keep doing it.
  2. Answer informatively and constructively anyway. I've recently argued that it is possible to give constructive debugging advice by way of answer.

In general penalizing answerers is bad. (There are exceptions in the case of wrong, spammy, etc. answers.) It was still volunteer effort on their part for whatever reason, and ultimately the line between a good question and a bad question is more of a function of supply and demand from which our moralizations follow, not the other way around. Let me explain that a bit.

If we lived in a world where any really bad question could get a good answer, then SO would not be worse - it would be even more amazing. It would be contradictorily begging the question to say "SO lets users not learn things for themselves" if use of SO is a reliable, full-proof way to get answers. In other words, SO doesn't make it easier to not help oneself, it moves the goalpost of what helping oneself means entirely.

However, we do have bad questions and big city problems. Bad questions include:

  • ones that can't really be answered, e.g. "debug my code" -> "where's the code?"
  • ones that are tedious and could be expedited greatly by more research effort on the part of the asker, not because these are terribly wrong questions, but because we'd rather other questions be ahead in the answer queue (supply and demand again).

But if an answerer feels generous for whatever reason, it is not our position to moralize that effort. It is our job to come up with reasonable rules and standards and morals to optimize answers, and then let answerers do whatever they please, limited by their own abilities and generosity. A bit of a "free markets FTW" approach, but yes, this is how you make a big city economy like SO successful.

share|improve this answer
1  
My problem is that they decrease the signal to noise ratio by inviting more people to post those kinds of questions. Besides, the same thing could be said for off-topic questions too - if someone wants to answer them, why not? –  sashoalm May 15 '13 at 14:56
    
@satuon if those questions were effectively answered all the time, then those questions wouldn't be noise, they would be signal. SO's domain would extend to include "free impeccable debugging." Even as a relatively experienced Java developer, I would be ecstatic to have such a service available. But we're not quite at that ideal, so downvote and closing is the right thing to do, and a given user can't pull it off more than a couple times until they are question banned. –  djechlin May 15 '13 at 14:59
    
@satuon re. off topic... my argument still goes through. We've had the ability to answer sys admin questions, PC power user questions, "should I quit my job if my boss doesn't let me use my favorite technology?" questions, and rather than say "no go away," we've created serverfault.SE, poweruser.SE, programmers.SE and dozens others. We respected the economy and just tended it well. (I'm not entirely anti-gov't regulation, ya know.) –  djechlin May 15 '13 at 15:01
    
@satuon note that I and probably your hoard of downvoters are agreeing with your point that these questions are kind of like weeds... we're just also saying DV/close manages it well and it's not out of control. We're disagreeing with you that it's a problem worth further measures than what we have in place. –  djechlin May 15 '13 at 15:04
    
OK. My post was inspired after reading meta.stackexchange.com/questions/19665/the-help-vampire-problem, I have been noticing a deluge of bad quality questions in the site of late, which is a bit off-putting to me. I would like to answer a question, but I don't want to wade through countless bad ones until I find a good question. –  sashoalm May 15 '13 at 15:10
    
@satuon more than usual? And are you sure it doesn't just have to do with semester-ly cycles, you know, finals and stuff. –  djechlin May 15 '13 at 15:11
    
@satuon try searching for questions with upvotes and no answers. If it's more than a day old there's a good chance it's in a specific technology or a hard question, and those are the ones we most need attention from experienced and/or assiduous users for. Hypothetically if I were going to employ you from your SO answers, I would look more for things that might earn a necromancer badge than the +5 votes you'd get for being a FGITW (fastest gun in the west) on an easy Java question. –  djechlin May 15 '13 at 15:13
1  
@satuon looking at your profile it looks like you have some expertise in QT. There seem to be open, good questions in QT that might be easy for you. Once you get as specific as QT you've already avoided the deluge of "help me with my C/C++/Java homework" questions, so try toying with your "favorite" flags to fish these out better. –  djechlin May 15 '13 at 15:20
1  
Thanks, my question was more like a rant, really, I just wanted to stand on the soapbox. I was upset that very basic questions would get 5-6 answers, and those answers would get a lot of up-votes, while more difficult questions would have answers with much fewer up-votes. That's because the people will up-vote the questions and answers that are most accessible to their level, and the basic questions are the easiest to understand. –  sashoalm May 15 '13 at 15:26
1  
@satuon if you can come up with a proposal to incentivize answering harder questions beyond the silver necromancer badge, without diluting the "ask questions get really fast answers" principle of SO, I would be interested in that. –  djechlin May 15 '13 at 15:32
    
I agree. Having someone tell me "you shouldn't have helped that guy" in a comment really ticks me off. I'll answer questions as I see fit, TYVM. –  John May 15 '13 at 22:39

You must log in to answer this question.

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