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 might be a meta-meta question, but here is the idea:

  • With 10-10k users on a site, no reputation is needed - everyone knows everyone else.
  • With 10k-1m users on a site, reputation is the best way i've heard of to track how "useful" different people are to the site.
  • With 1m-100m users on a site, does it still work as well?

As an example, on Wikipedia, which admittedly uses a very different model, once the site was viewed as canonical, it started being viewed as a target, instead of just a resource. If Stackexchange starts to get there, what changes would be needed? Similarly, over time, is there any idea about how to prevent older and less useful content from accumulating and poisoning the well? It is too easy to poison, and too hard to police.

  • The number of past questions increases linearly with time, but the cost to organize them is very non-linear. (Just assuming constant numbers of users, over time the site could become less useful as a source because of decay.)
  • The potential for spam increases, and the reward for well disguised spam increases. The cost of eliminating spam is constant, but the volume of spam can increase as a function of the number of users, and therefore the value to the spammers. (Because down-voting costs rep, and marking as spam is an action requiring moderators.)
  • The value of becoming a trusted user increase with size, and the cost is constant - So we should expect to see more users trying to get to high reputation in order to abuse the system. (The cost of getting to a given reputation level, in terms of time invested is fixed, the effective reward for doing so, in terms of power and influence, scales with the site's popularity and visibility.)

These are just some initial thoughts, but I wanted to know if there was discussion about how to avoid becoming victims of the success of the site. Is this something that has already been discussed, and I was unable to find the discussion?

Edit: To reiterate: Are there things that can be done to make the site less vulnerable to these concerns as it continues to grow. Many of these items are not currently problems, but the problems can grow exponentially, and the solutions grow linearly with the size of the user base.

share|improve this question
    
What do you mean by "The probability of users getting to high reputation in order to abuse the system increases dramatically."? –  Bart Jun 15 '12 at 13:26
    
Edited question to clarify - The idea was that, if I want to abuse the system, and that has a value of $10 to me, I might spend time to get reputation legitimately, just to be able to burn it later. If my value is a function of the number of users, say, $1/10k users, then I am interested in investing the time to become a trusted user iff the value is above my cost. –  David Manheim Jun 15 '12 at 13:33
    
And what makes you think that they won't be caught and dealt with just as rapidly as non-high-rep abusers of the system? In all honesty, I don't see any of the issues you point out. Though I might be missing the point. –  Bart Jun 15 '12 at 13:36
    
First, the question is prospective, not retrospective. And do you really not see any accumulation of low value Q&A sitting on the site for years ago that will continue to accumulate and make finding things harder? –  David Manheim Jun 15 '12 at 13:39
    
Second, the system exists for dealing with abuse - it just requires moderators. Once it's worthwhile for abusers, the number of people doing so won't increase linearly, it will increase exponentially. That's why I see it as a scale issue. –  David Manheim Jun 15 '12 at 13:41
1  
Not really no. Bad content is actually dealt with quite swiftly in my opinion. There is no fear to close or delete those questions. Even retroactively questions are closed which have (over time) become off-topic. With the very active community present, we do a fairly decent job of keeping the site clean. Nothing is perfect, but I have not experienced an increased difficulty in finding good quality contributions in among supposed crap. And I don't easily see it happening. Though your experience might differ. I don't know if it depends on certain tags. –  Bart Jun 15 '12 at 13:42
3  
"once the site was viewed as canonical, it started being viewed as a target, instead of just a resource. If Stackexchange starts to get there [...]" Er...people have been trying to turn Stack Overflow into their own personal advertising space and/or blog almost since day one. They have had little success because users at almost every level and detect and act on these attempts: flagging, closing and deleting such content in a hurry. –  dmckee Jun 15 '12 at 13:53
1  
Unclear how any of this deals with the scaling issue. If the site doubles in size, do the number of people interested in spamming only increase linearly? If you are looking at the current situation, you aren't discussing the same thing as the question. –  David Manheim Jun 15 '12 at 14:11
1  
If the spam rate increases the mean time between users spotting it decreases, so unless we have saturated the per-user spam handling capability of the current system it is auto-correcting. Possibly changes will be needed in the future but, that is only a possibility and you are only guessing at what part of the system will show the strain first. Now is not the time to worry about these things: we worry when we have evidence of some particular failing. –  dmckee Jun 15 '12 at 14:18
    
@GardenGnobobby If I was only responding to you, I could have used them. I'm unclear if this is intended to be insulting, or if it was just posted in the wrong place. –  David Manheim Jun 15 '12 at 14:22
    
If the response is - "Don't worry about or discuss the future until you know what will happen", then that can be an answer. I think it's a unsatisfactory one, but everyone commenting here seems to disagree. –  David Manheim Jun 15 '12 at 14:23

2 Answers 2

If the response is - "Don't worry about or discuss the future until you know what will happen", then that can be an answer. I think it's a unsatisfactory one, but everyone commenting here seems to disagree.

There's no point in worrying about something that might happen, unless you have evidence that it is a legitimate possibility in the future. Wikipedia seems to be doing fine, despite being a "target". Sure, spammers get in, but the hugeness of the place, along with a team of dedicated users, ensures that the vast majority of spam/trolling is short-lived.

At present, I don't see how this is any different for SO. You claim that the ratio of users to people who stop spam will be steadily increasing, thus leaving the site more open to attacks. But there's no evidence for this; it's just speculation.

So we should expect to see more users trying to get to high reputation in order to abuse the system.

That would require an exceptionally dedicated troll. Spammers look for low-hanging fruit; getting 10k rep is not low-hanging, so clearly you're talking about a troll.

Getting 10k rep requires 50 days of hitting the rep-cap, or doing a lot of bounty hunting. Even the 2000 rep needed to edit posts freely requires 10 days of hitting the rep-cap/bounty hunting.

A person who is sufficiently capable of getting 20 people to upvote his answers every day for 10 days is obviously skilled at some aspect of programming. In all likelihood, this person would see the value in Stack Overflow. Therefore, it is highly unlikely that he would suddenly manifest as a troll. And if he does, fine; people will see his trolling, he'll be reported and sanctioned by the staff.

Total damage caused? Maybe 50 questions/answers are defaced before he's expunged. Once he's identified, his entire history of activity can be removed via moderator action.

Most trolls, like spammers, aim for the low-hanging fruit. Some are dedicated enough to do this, but honestly, how many such people not only would want to do it, but are actually capable of it? I don't think the intersection of these is very large.

The absolute worst-case you're going to get is a person who is a productive member of the community, yet frequently works some product of his into various questions where it would be on-topic. I think that's something we can live with.

The number of past questions increases linearly with time, but the cost to organize them is very non-linear.

The only organization we have are tags and titles. If you want to find something, you search for it. So what "cost to organize" are you talking about?

The greatest cost is when something needs to be re-tagged. And we have people who are willing to undertake that.

share|improve this answer

So you are worried about spammers participating in a useful manner in order to rack up reputation to aid their less-constructive endeavors?

I can only think of one answer: Mission accomplished!

That said, it takes on average at least few hours for a spammer to gain enough reputation to make a problem of themselves, while also evading the various IP blocks and other administrative measures. It takes a handful of users less than a minute to flag a spammer, inevitably leading to their take down.

Bottom line? I wouldn't worry just yet...

share|improve this answer
    
That is a fair point. No, I'm more worried about users showing up in droves and reputation farming, which could get them to the point where they can cause damage. –  David Manheim Jun 15 '12 at 14:31
1  
@DavidManheim Ah. There are--of course--already several mechanism to inhibit reputation farming, but big voting rings could evade them. Not sure how the team would respond to that, but I note that clique-finding has been an active area of research in the past decade with much progress, so perhaps the tools will be available... –  dmckee Jun 15 '12 at 14:59
1  
@dmckee: there is also the relatively recent change where deleting a post removes any associated reputation. So spammers cannot simply post garbage and harvest points from their buddies - the points will be gone as soon as the garbage is... –  thkala Jun 15 '12 at 15:05

You must log in to answer this question.

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