30

I've noticed that recently Stackoverflow has been overwhelmed with help vampire questions, to a level I have never seen before.

I am aware of previous questions, but none that propose the solutions at the end of this post.

The root cause is almost certainly that it works: trivial questions get explanations, requests to post code are almost always satisfied.

I suggest that the problem has three principal efficient causes:

  1. There is no highly visible explanation of what makes a great question, and nothing that would force someone to read and internalise it;
  2. The flagging/vote to close system has no entry for help-vampire behaviour; and
  3. Right now there is no process or policy that denies rep to help vampire-feeders.

In relation to cause 2, I would suggest that things like "exact duplicate" are not apt to deal with failures to do any research (if for no other reason than it may be difficult to identify an exact duplicate), and basic questions do not really fit into "does not fit into our Q&A format"; frequently they are not really "too localised"; and frequently they are on-topic and well defined, but show a complete lack of any attempt to do anything beyond post on SO.

I propose the following changes:

  1. A cap on questions new users can ask: something like no more than 4 questions to be asked until a user has a rep of 50, and at least 4 upvotes on questions;
  2. A link to be added to the question page to a guide on asking great questions;
  3. Adding a "Help Vampire question" flag - with a description like "This question is trivial or can be simply answered from standard documentation or tutorials; or this question amounts to request that someone write the questioner's code for them."; and
  4. A policy that denies rep to feeders - either that vampire questions are deleted, or if they have some merit (non-trivial question, but does have demands to be shown teh codez), that they be changed to a community wiki (it is my understanding that no-one gets rep for community wikis).

Is there any appetite for these sorts of changes?

In relation to change 1, this is different from a daily or other general quantitative cap, in that it forces new users to reach a substantive goal, which is assisted by learning how to ask good questions. It is like needing to earn a badge before your cap is lifted.

In relation to 4, we don't really have anything that discourages feeders - as long as they answer while the question is open, they can get their points. Downvoting is not a solution, because it needs five downvotes to cancel out one upvote. I am proposing that rep be denied to people who answer vampire questions, so the only thing they lose from answering the question is their time (i.e. they are no worse off in terms of points once the process is completed than just before they answered the question). This will gently force answerers to evaluate the question.

To be clear, if moderators were encouraged to both close, and wikify, I would think that would be an adequate solution, as long as the practice would be generally adopted.


Some examples:

  1. How to print objects in python: https://stackoverflow.com/questions/8954322/printing-out-objects-in-python (Basic tutorial question)
  2. How to write a for loop that counts towards zero: Can you have a for loop with a -1 step in Python, like you can in VBA? (Basic reference/tutorial question;rewarded with three working code samples)
  3. Implementing banner ads in django: Show different banner on every user click on django site (On its face a fairly reasonable question, but followed with persistent requests to post the code to do it, until someone else actually did provide the code)
  4. Python-Haskell communication libs: what are the ipc libs to communicate between a python and an haskell process? (Total lack of prior research to asking this question. Eventually closed as an exact duplicate, after locating an alike question on stackoverflow).

In each of these examples, even if the questions are closed (but not deleted or wikified), the vampires are getting fed AND their feeders are rewarded with rep. That hardly amounts to anything that discourages either end of this behaviour. If moderators are reluctant to delete questions, then another option that denies rep, but leaves the question in place, will help break the cycle.

To be clear, if moderators were encouraged to both close, and wikify, I would think that would be an adequate solution for 4, as long as the practice would be generally adopted.

40
  • 2
    You need to provide at least one example link; the help vampires I am thinking about may not be the same ones you encountered.
    – user102937
    Commented Jan 21, 2012 at 16:11
  • 2
    There are safeguards in place for some of your concerns (probably for all, really - examples: questions ban, links to "how to ask" page). You should do a little bit more research before asking, don't be a help vampire :P
    – yannis
    Commented Jan 21, 2012 at 16:19
  • @RobertHarvey: Acknowledged - give me a second.
    – Marcin
    Commented Jan 21, 2012 at 16:19
  • 1
    @YannisRizos: There being things which are similar to what I'm talking about is not the same as the same things being in place - if they were (I propose), we would have fewer vampires.
    – Marcin
    Commented Jan 21, 2012 at 16:20
  • 4
    You can't when a founder is one of them.
    – Pollyanna
    Commented Jan 21, 2012 at 16:40
  • 1
    "recently Stackoverflow has been overwhelmed with help vampire questions" You'd have a stronger case if you showed how this is currently harming stack overflow. There are numerous ways the system combats help vampires right now, so you might need to convince us that current levels are harming us.
    – Pollyanna
    Commented Jan 21, 2012 at 16:42
  • @AdamDavis: Page not found.
    – Marcin
    Commented Jan 21, 2012 at 16:43
  • 4
    I think especially change 4 (or anything directed towards the answerers) can make the difference. New help vampires will keep coming; it's a big world and not all of the vampires have found these sites yet. But I've seen many long term users answering their questions.
    – Arjan
    Commented Jan 21, 2012 at 16:45
  • @AdamDavis: As to coming up with evidence of harm, I don't have access to the stats about what is happening with stackoverflow. It is my impression that the quantity is greater, and I find that more of the questions I visit are vampire questions.
    – Marcin
    Commented Jan 21, 2012 at 16:45
  • 2
    Regarding your examples: 1 and 2 were closed as too localized, and the lack of prior research on 4 is disputed. Question 3 is easily handled if people simply stop responding to the OP's comments asking for more help.
    – user102937
    Commented Jan 21, 2012 at 16:49
  • 3
    @AdamDavis: as I replied to Robert, I don't think it is handling them correctly. The vampires are getting fed, and the feeders are being rewarded.
    – Marcin
    Commented Jan 21, 2012 at 16:57
  • 4
    TL;DR on all these comments. But if someone asks a simple question (with an answer that is easy to find, or just looking for code) and someone is willing to give a relevant, complete and useful answer to it, who cares? This just gives SO another landing page. What problems does this cause? Commented Jan 21, 2012 at 20:18
  • 2
    @ThinkingStiff (a) it turns the balance of pages from useful to trivial (b) it encourages more people to come and ask trivial questions (c) that drowns out more substantial, and takes away their attention. So the cycle continues (slowly, but slowly accelarating).
    – Marcin
    Commented Jan 21, 2012 at 21:26
  • 2
    I disagree completely on your fourth example: it does not feel like a help vampire so much as a guy who might not have known what to search for -- he could be a helpful contributing member to our community with a touch of guidance. Vampires won't be. Furthermore, it was closed as duplicate of the wrong question. He's got two processes: front-end Python and back-end Haskell. The one it "duplicates" is about calling routines in the same process. Those are different things, though there are some problems you could solve with either approach.
    – sarnold
    Commented Jan 22, 2012 at 1:41
  • 2
    @Marcin - Why would killing these question be a "step forward"? Let me get this straight. There are people out there getting help who don't deserve it? The horror.
    – Wayne
    Commented Jan 30, 2012 at 19:43

3 Answers 3

19

Most of the changes you proposed are already in place;

  1. There's a cap on the number of questions any one person can ask in a short time period, and question bans that occur for people who post too many downvoted questions.

  2. There's the https://stackoverflow.com/questions/how-to-ask page which is prominently linked for new users (people don't read, and you can't force them to).

  3. The "General Reference" close reason is no longer a possibility. The "Not a Real Question" or "Too Localized" close reasons can be used instead.

  4. Help vampires don't care about rep, they just want to be spoon-fed answers to their crappy questions.

  5. There's always suspensions for the most troublesome offenders.

The more troublesome problem is that a small number of help vampires are now using workarounds to subvert the system. This tiny group of people has a disproportionately negative impact on the site; I don't have a good solution for that.

15
  • Is the question-cap working on Betas too? If the answer is yes, I've seen the evidence of the opposite then (unless I misunderstood how it works).
    – Alenanno
    Commented Jan 21, 2012 at 16:33
  • 7
    In relation to your 4 - I'm not talking about the rep for vampires, but rather for their feeders. None of the features you identify deal with that issue.
    – Marcin
    Commented Jan 21, 2012 at 16:34
  • 1
    @Marcin There's where voting comes in. Downvote the questions and answers to oblivion (if you really think they deserve your downvote) and get them one step closer to deletion. When that happens, feeders lose their rep.
    – yannis
    Commented Jan 21, 2012 at 16:36
  • 1
    @YannisRizos: Downvotes only cost 2 rep, upvotes earn 10: you need 5 downvotes to cancel out one upvote. The numbers work out in favour of feeding.
    – Marcin
    Commented Jan 21, 2012 at 16:38
  • 10
    @Marcin: You want to penalize the feeders? That's not fair; I've seen egregious vampires whose questions, taken individually, look just fine. The problem is, they ask a string of questions about the same dozen lines of code. It's not fair to penalize people who are just trying to be helpful.
    – user102937
    Commented Jan 21, 2012 at 16:38
  • The point of my first change is actually to force people to read the guidance on great questions.
    – Marcin
    Commented Jan 21, 2012 at 16:38
  • 1
    @RobertHarvey: I want to penalise people who answer obvious vampire questions, the questions so taken individually. If vampire questions not deleted were turned to community wikis (or otherwise treated specially) so that when detected, answers earned no rep, that would make users pause before answering incredibly basic, and "post teh code" questions.
    – Marcin
    Commented Jan 21, 2012 at 16:40
  • @Marcin I never said that downvotes alone would do anything. If the community agrees with your assessment and quite a few other people downvote, then the question will be deleted (at some point in the future), and the feeder will lose any rep gained. The important thing here being that your assessment may not be correct.
    – yannis
    Commented Jan 21, 2012 at 16:48
  • 1
    @Marcin Funny thing is, I was awarded with a gold Reversal badge for feeding an obvious help vampire (and 250 rep). I sympathize and understand exactly where you're coming from, but don't see any point to an extra feature.
    – yannis
    Commented Jan 21, 2012 at 16:50
  • 2
    @YannisRizos: For several reasons, including that the questions as such do not fall into the various flag categories, the questions are not being deleted. An option for moderators to wikify (or whatever) to deny rep without taking the step of deleting, might result in more such questions being dealt with.
    – Marcin
    Commented Jan 21, 2012 at 16:52
  • @YannisRizos: You don't think it makes sense to discourage feeding? Or you don't think a new feature would be effective?
    – Marcin
    Commented Jan 21, 2012 at 16:52
  • @Marcin Oh, I only disagree with the feature, not denying that the behavior exists. I've posted at least one meta question on P.SE asking people to stop providing answers when vampirism is obvious. CW may be a good solution to that end. I don't know.
    – yannis
    Commented Jan 21, 2012 at 16:56
  • 2
    To note, I think we can count "general reference" out for SO. It would be useful, but probably applied inappropriately in many cases. It's something I initially fought for, but while digging up examples, it just became blatantly obvious that it would be used by more experienced users to close anything they consider 'too basic'. See this question and the comments / answers for more.
    – user50049
    Commented Jan 21, 2012 at 18:54
  • @TimPost: Good point. I clarified my answer.
    – user102937
    Commented Jan 21, 2012 at 19:06
  • @robert I really wanted "general reference" to work, but the more we looked at it, the more untenable it seemed in practice Commented Jan 25, 2012 at 9:22
15

In addition to what the others said (and it's all correct), if you see particularly egregious examples of Help Vampires, flag one of their posts for moderator attention and explain why.

We always follow up on flags, and we are always happy to investigate accounts if you honestly believe that these accounts are damaging the ecosystem of the site.

Help us help you!

2
  • No doubt this is a good idea. Do you think there is any merit in having a specific policy to discourage helping help vampires such as wikification?
    – Marcin
    Commented Jan 25, 2012 at 10:49
  • Yup, same thing I learned back in the days after getting to understand how the system has been implemented to discourage this, that it's OK to just downvote and flag help vampires. Some form of wikification can probably be achieved, but that really depends on the type of question asked... Commented Jan 26, 2012 at 23:01
14

You have to differentiate between help vampires and poor questions.

Help vampires ask many poor quality questions, whereas even a great user can occasionally ask a poor quality question.

The poor quality question ban is the primary method to stop help vampires - people who consistently ask questions that are downvoted, closed, or flagged.

And it works.

So the problem of help vampires is taken care of in the system already.

The problem of poor quality questions are handled largely by users who downvote, close, flag, and delete questions as they come in.

However, if you still feel that this is a problem needing an urgent solution, I suggest you do the following so you can convince skeptics like myself:

  1. Develop an objective way to determine if a question is poor or a user is a help vampire. Describe it as a process that users, or ideally the system, can follow to detect these problem questions and users. This is triply important to your proposal because you suggest users should be dinged for helping vampires - but without a good process to detect vampires your solution is worthless.

  2. Show the problem with hard numbers - statistics, examples, etc - that show that the above process works in detecting these problems and that the problem is substantial in scope, affecting most users negatively.

  3. Show how your solution would be applied and would change those specific numbers and statistics, and thus solve the problem.

The issues I have with your proposal are

  • I'm not convinced it's a significant problem (the examples you give show the system to be working, and you've provided no numbers such as "10% of all questions being asked and answered and left open are bad questions, so the existing system is failing, here is my sample I'm using, and the process I'm using to determine whether a question or user is bad")

  • Even if it there are a lot of questions that slip through the existing mechanisms, I don't see how it's damaging stack overflow. I know how help vampires hurt systems, but in this case you're making the very specific claim that the current load of vampires is dangerous - yet you don't specify the magnitude of the load, and in what ways it's actually proving harmful. There are hundreds of ways vampires are hurtful, but what symptoms show that stack overflow is suffering from this assumed problem?

  • You've pointed to bad questions as an example, but not to bad users. Are you fighting against help vampires, or against bad questions? They share similar symptoms, but the solutions for one don't always apply to the other, and it looks like you're confusing the issues. A little more clarity and focus to your message would go a long way to solidifying your position.

6
  • You have already demanded numbers, and you know there is no way I can come up with them, because I don't have access to those numbers.
    – Marcin
    Commented Jan 22, 2012 at 8:04
  • @Marcin "I've noticed that recently Stackoverflow has been overwhelmed with help vampire questions, to a level I have never seen before." If you can see it and feel it, then you should be able to measure it. If you can't measure it, then perhaps it's simply your perception of things, rather than reality. I don't know. But if you can't help anyone else to see it, then I don't see how it's a valid issue - it's not affecting those who can't see it.
    – Pollyanna
    Commented Jan 23, 2012 at 17:25
  • "If you can see it and feel it, then you should be able to measure it." That is not a valid inference. You keep harping on this idea that I should be able to provide numbers, knowing full well that I can't because I don't have access to statistics about stackoverflow, because they aren't public. This suggests that you are not acting in good faith, but simply positing an impossible hurdle because you know it is impossible.
    – Marcin
    Commented Jan 23, 2012 at 17:29
  • @Marcin When I feel there's a problem, I do a quick survey to see the magnitude of the problem. It's tedious, sure, but it's always possible - if you can see it, you should be able to measure it. All you need to do is sample the most recent 100 questions on stackoverflow, then figure out the percentage you feel are bad. Then go back a day later and see which ones were closed. Besides, why do you insist that stack overflow has these statistics? How would they know something that you can only "feel" but can't even provide a dozen examples of?
    – Pollyanna
    Commented Jan 23, 2012 at 18:42
  • 2
    @Marcin I'm not opposing you for the sake of opposing you, if you mean to imply that I'm acting in bad faith. I'm telling you 1) that I don't see it, and therefore don't believe it's a problem, and 2) the steps you need to take to convince me. It's ok if you choose not to convince me, I'm not in any position of power around here, just another user of the system. If it's truly a problem, I would like to see it resolved. But I can't support what I can't see, and I suspect I'm not the only one that is left scratching their head at your feeling.
    – Pollyanna
    Commented Jan 23, 2012 at 18:44
  • If I'm the only person that feels this way, it doesn't matter what the numbers show. Feel free to oppose me if you simply feel this isn't a problem. The activity here suggests I'm not the only one who does feel this is problem, and that you are not the only one who doesn't see the problem.
    – Marcin
    Commented Jan 23, 2012 at 18:53

You must log in to answer this question.

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