Talking to Ward Cunningham about wiki

Last week, The Guardian published an article I wrote about wiki. The story started as a quite different idea that I pitched to the editors there, but metamorphosed over a period of weeks into a much more generalised overview of the state of wiki.

The centrepiece of the article was a series of comments by Ward Cunningham, the guy who created the original wiki and started the ball rolling for a whole series of web sites using spin-off technologies.

As usual, Ward said a lot more than I was able to fit into the article. We chatted on the phone for about an hour. Below are my transcribed notes of his main points - I've checked that he doesn't mind me posting them here, and his response was: "Those notes are faithful to my meaning, though not always the words I would use. I would feel comfortable having them published as is only if the page included your explanation that they are your best recollection of things said quickly on the phone."

Well, they are my best transcription of hurried shorthand notes, made by someone whose shorthand has always been shoddy. But I think they're an accurate summary of what was said.

Roughly what Ward said

For years I would turn down offers to be cited because I feel that wiki is extremely fragile.

I couldn't withstand disinterested visitors. As my wiki has got larger and larger it has a certain mass, and becomes more difficult to abuse.

A lot of people will come to mine having seen other wikis, and its plainness is an advantage. People look at it and say, "hmm, this looks boring", and go away. The quality is deep, not at the surface.

A cadre of devoted housekeepers were there to keep it healthy.

Every wiki develops a set of norms. Every member of the community sets themselves against those norms. If you have people who post stuff that is waaaay beyond those norms, such as posting pornographic images in pages, then you find that kind of thing gets dealt with very quickly. It just gets removed.

But since last Fall we have had an individual who has been posted only slightly outside those norms, so close to what's acceptable that others have been unable to agree on whether or not his contributions should remain.

As a whole, the body of text is safe. Even with months of free time, this person touched no more than 100 pages out of 26,000. What he did do was wear out some of the housekeepers, which made some of them leave and not come back, and that was sad to see. But, there is a steady supply of housekeepers.

Thing is, abuse has happened before. Some people have arrived at my wiki and started out abusive, then slowly become part of the community.

I called Sunir and said: "We should write a paper about this guy. He's testing us."

People said "ban him" but I'm not really sure I'd be able to effectively do that even if I wanted to. I'd be getting into an arms race that I could never win.

Sunir understands what he calls "soft security". I was using code against behaviour but I didn't feel that I was in a very strong position.

The problem was that the abuser had too much time. He was too active and could get too worked up about things, so much that he had to fight.

So I put a post-limiter in place. People can only post so many times during a set time period. And it worked, almost straight away. We haven't banned the abuser, merely limited his ability to post so that what he does post is more within the norms we can expect and deal with.

I keep a short-term history of changes. Even active pages would only have 5 to 10 changes a day. Only the SandPit would have hundreds. But our abuser, when he was at his peak, would cause hundreds of changes to several pages every day.

I've limited his participation to a degree that the community is willing to accept.

Wikipedia gets more of this that I do. They have a stricter writing policy. But they also have a strong community. I don't keep track of how often they have problems, but I know they do.

"Wiki is a process for organising and explaining experience"

I think wiki is a miniature version of science. Science is a fascinating process where you make observations and hold them up to peer review.

Science is a process for organizing and explaining nature. Wiki is a process for organising and explaining experience. I ask people to tell me their stories, and people like to tell stories. It's a natural, social thing. Wiki provides a machinery for weaving together those stories.

TBL saw all this when he imagined that people would press the "post" button as much as the "get" button. But people weren't ready for it then, so the web developed in a different way. I just took some of the pieces and created a system that very nearly doesn't work ... but it works.

Civilisation is a construction of people that exists longer than anyone has a right to expect it to.

I like walking in city parks, but they are very easy to destroy. And sometimes someone will come along in their SUV and destroy one, but we still build city parks. Just driving along, you can see how many buildings, how much of civilisation remains in one piece and working when it could so easily be broken by a very few people.

Filed under: computers
(6th April 2004)