LW wiki spam filtering

by gwern1 min read30th Mar 201332 comments

40

Site Meta
Personal Blog

Long-time readers may have noticed that spam on the wiki has been a very persistent problem for the past 2 years or so; I've been dealing with it so far by hand, but I recently reached a breaking point and asked Trike to resolve it or find a new wiki administrator. (Speaking of which, is anyone interested?)

So Trike has enabled a MediaWiki extension called the edit filter: a small functional programming language which lets you define predicates applied to edits which trigger one of a set of actions, like banning a user, deleting an edit/page, or stopping an edit from going through. I have so far defined one rule: page creation is forbidden for users younger than 24 hours. This so far seems to have worked well; spam pages have fallen from 5-10/day to ~5 over the past 2 weeks. This is much more manageable, and I am hopeful that this new anti-spam measure will be effective longer than the previous additions did (but if it doesn't, I'll look into adding more rules dealing with images and external links, and perhaps also ban users whose names end in a numeric digit as almost all the spam accounts do).

If you've run into this edit filter before by making a page and seeing the submission rejected with an error message, fret not: merely wait 24 hours. (If your account is more than a day old and you're still getting errors, please contact me or Trike.)

32 comments, sorted by Highlighting new comments since Today at 11:00 PM
New Comment

I implemented Akismet to protect the SIAI blog about 3 years ago. It was very successful.

Why not require positive karma to create or edit wiki pages? There's no relevant downside to it.

Alternatively, why not use a stronger open-source CAPTCHA?

Also, LessWrong and LessWrongWiki accounts are separate. Unfortunately.

Why not require positive karma to create or edit wiki pages? There's no relevant downside to it.

Who's going to program that in?

Alternatively, why not use a stronger open-source CAPTCHA?

We already use reCAPTCHA, IIRC.

Who's going to program that in?

The same one who defined a predicate checking against the age of the user's account, maybe the karma score is similarly available? Also, there's been some fiddling with karma representation recently, so it may be an easy task for the person who implemented that.

The same one who defined a predicate checking against the age of the user's account, maybe the karma score is similarly available?

You're not a programmer, are you?

MediaWiki and Reddit are entirely different programs, so no, using one of the canned rules in a commonly used extension designed for non-programmers is very different from figuring out how to interface a mess written in PHP to a custom Python codebase to extract such information and expose it to the rules in the latter.

Writing the rule based on the extension docs took me maybe 5 minutes. I would be a little shocked if Trike could make the karma available in the live site, with testing that it works correctly, in anything less than 2 orders of magnitude more time (8 hours).

I see.

It seemed like a reasonable guess, I haven't checked out the specific architecture and am not familiar with the MediaWiki API or the extension being designed for non-programmers.

If web programming was clean and nifty, the wiki and the forum system would both reference the same user database on some MySQL backend. If it's an "organically grown" mess, nevermind. But the API should still be well defined. Now, it was only an idle suggestion in case someone missed some low hanging fruit, I didn't expect it to be some assured panacea. I've only ever done one php/mySQL website around 9 years ago, I don't much care for web programming.

Could also just remove the whole "recent wiki edits" potion of the sidebar. It doesn't provide much and takes up screen space on all the important pages.

Could also just remove the whole "recent wiki edits" potion of the sidebar. It doesn't provide much and takes up screen space on all the important pages.

It'd be much improved if it simply filtered out the registration of new user accounts.

Don't we already have code to extract karma points values for the minimum for creating a new discussion post?

Discussion posts are made within the Reddit system which is Lesswrong.com; which is entirely separate from the MediaWiki instance which is wiki.lesswrong.com.

Why not require positive karma to create or edit wiki pages? There's no relevant downside to it.

There may be people who love the wiki, but not LW. Requiring positive karma might make editing the wiki more trouble than it's worth for them.

Just saying "hi" in the Welcome thread or posting basically any random sentence whatsoever in the Quotes thread should take care of that, one single karma would suffice.

To be fair, it is possible to actually be downvoted in the Quotes thread.

Considering that the great majority of wiki edits are spam and spam deletion, some more inconvenience may be warranted.

I have so far defined one rule: page creation is forbidden for users younger than X hours

Never, ever publicly post your constants. If it was a site-specific spammer, he can now create accounts X hours before posting, aka good old cookie-aging.

Overkill security professional solution (if you don't mind Ajax and some coding though) : have the site or at least crucial part of it self-decrypt with one-time-pad. Doubles the size (if whole-site) but robots extremely rarely run scripts so both chunks parse as garbage. And even if they did understand JavaScript you could make the problem "AI-hard" in principle (yes... I do realize there's no such formal class).

Also, for fun: http://xkcd.com/810/

Never, ever publicly post your constants. If it was a site-specific spammer, he can now create accounts X hours before posting, aka good old cookie-aging.

I'm not worried. As you can see from the sidebar, spammers have been prolifically creating accounts for countless months and almost all accounts wind up never being used. My inference is that most of them are being stymied by other anti-spam features. None of the spam seems to be done by hand, and certainly they aren't looking on an obscure post on a different domain for a value that they would stumble upon accidentally ('hm, my spam account didn't work yesterday, let's take a look to see if the error went away - oh, it did, there must be a 24-hr timeout').

I totally agree there's a very low probability of leak there. There's still a (meta) reason to do stuff like this though.

If you have a high value target (like building an AI), you need insane paranoid security. Means you need to train the mindset of patching everything you possibly can not because you think it's unsafe but, you know, just because. E.g. you could tell the difference between good and bad sysadmin just by looking at his own PC. A good sysadmin will have every single drive truecrypted, no matter the inconvenience. This is an opportunity to train the mind into the job.

Thankyou for doing this.

It's pretty silly that the site about instrumental rationality isn't the greatest site on the internet yet. I don't have any concrete suggestions to fix it besides "If that bothers you enough, start looking at source code relevant to these problems".

In a saner world micropayments would take care of this as everyone would expect to have to pay a penny to evidence that they are not a spambot. But as we don't have a good micropayments infrastructure this solution is likely unavailable to you.

Which can of course still be misapplied; the Bitcoin wiki at some point began requiring a payment of some bitcoin, I logged in to make some more edits, discovered it applied to me too, and mentally told them to go screw themselves: I wasn't about to pay for the privilege of working for them, when I had already proven myself with previous good edits.

Requiring people to fork over any amount of money at all usually results in many fewer participants. This doesn't seem better overall as long as other effective solutions to spam exist.

Probably because it's currently very hard to pay 0.1c hassle-free.

Artificial shortages are artificial. Why should people bother giving you money to work for you?

To prove, at an extremely low cost to you and zero cost to society, that you are not going to spam them.

Hence my starting with "In a saner world"

I see. The solution does not work, but it is nevertheless correct, because it should work. It is the people who are wrong.

In the real world, a solution to a problem is something that actually solves the problem. An engineer does not get to blame metal for obstinately rusting despite his wonderful anti-rusting invention. Instead, he recognises that his invention does not work. Neither does a would-be social engineer get to blame the people for not behaving as he thinks they should.

"It is the people who are wrong."

Hence the need for a website called "LessWrong"

As the person whose experience Gerard is citing, I would point out that my beef is largely with how the micropayment is being done: known-good users (like myself) were not being grandfathered in, as would be sane.

Having dealt with so much spam on the LW wiki, I wouldn't be averse to a micropayments thing if it were even slightly reasonable to expect random LWers to possess bitcoins and there were some non-payment way of editing with moderation (eg. having the first n edits by users who choose to not pay go into some moderation queue before going live).