In many on-line communities the evolved community expectations for interactions between users are less respectful than the standards that one might encounter at a local farmer’s market. Philip and others did a good job of setting an expectation that the person you were interacting with electronically was also a member of the same physical space community and should be given the same level of respect that you’d give to someone in person.

This lead to events in town where wiki editors would gather for face to face interaction as well.

Another factor is the high percentage of editors who use their real names. People tend to be more respectful of other editors humanity when they are standing behind their words with their real name. We’ve had some really positively contributing anonymous editors as well.

The thousands of generated pages on wikipedia became a canvas for contributors perhaps because wikipedia itself had already reached an intense critical mass, both from a community perspective and a google-juice perspective: those generated pages likely showed up in searches all over the place.

But, as Michael hints at, wikipedia has different contributor dynamics than Davis Wiki. We really need as many people in the community to contribute for the project to work. Wikipedia can get by with a very small percentage of folks, mostly because of its intense breadth.

1) Among various successful wikis, what percentage of pages were bot-generated?

2) What’s the user-to-contributor ratio for Wikipedia now, and what have its historic levels been?

Maybe there’s a distinction to be made here between bot-created pages and bot-enhanced pages? I’m a huge fan of the info boxes, categories, and linked phrases that bots have used to bolt various Wikipedia pages together.

But if I found a wiki that had a large proportion of pages consisting entirely of uniform, bot-generated content, I would get that artificial McWiki feeling Evan and Thompson are recoiling from.

Not that bot-generated pages can’t be pretty useful. But as a potential contributor, I think I’d have sort of limited expectations for the page. I wouldn’t feel the frission that made the early, sort of crappy DavisWiki pages interesting and seductive. (Or here’s a random early Wikipedia page. Don’t these pages just make you want to … improve them?)

And converting visitors to contributors (and contributors to devotees) is SUCH an important part of wiki projects.

The more devotion is required, the more important aesthetics become, I think.

I absolutely agree that you can’t create a community with a database dump. But to say that there aren’t massively successful wikis who have used bots to support the work of people is patently untrue. The human element is the most important, but it’s not the only element.

Trying to create a community with a database dump, no matter how well formatted or accessible it is, is a dubious act. Creating one by comparing which Mexican cuisine in town is the most authentic and a roundup of which public bathrooms are clean (and then making photo galleries of the ones with interesting or whimsical decor)… that’ll get people involved. Plus it grows a more interesting wiki. It’s the human element that drives the community, and an initial focus on that seems to be more successful versus trying to splatter around 10,000 template entries, all stamped as stubs (or, as the Davis Wiki calls them, seeds) and then wait for people to spontaneously become interested.

Letting people do the work means they become more and more invested in the wiki. A group of people with large amounts of empathy and passion are much more key in the long run than autoinserting category boxes.

–
Evan “JabberWokky” Edwards
http://daviswiki.org/Users/JabberWokky
jw@timewarp.org -*- 814.889.8845

You riff on Thompson’s point that discounts the role of bot-created wikis as less valuable than that which is hand-built (so to speak). His primary evidence was that the world’s number one wiki, Wikipedia, is a place “whose every page was built, word by work, link by link, on the actions of individual people.”

The only problem is, that’s not true.

Bots play a huge role in creating and curating content in Wikipedia, and have for years. If you graph the article growth of Wikipedia, everyone immediately notices a large jump at one point. That was when a single bot created thousands of pages for municipalities using public data. Those pages became the canvas for people who lived in those towns and cities to work from.

Most of the largest wikis incorporate bots in some fashion. While it’s obvious that dedicated people are the lifeblood of a wiki, some of those people need to be coding bots.

1) For anyone interested in the community-building aspects of wikis, Rich Millington’s FeverBee is essential reading. Philip’s story reminded me of this, for example.

2) The seventh wiki tip I left out to save space: Don’t be scared of the word “wiki.” Lots of projects, such as Omaha Commons, another Omaha-based city-wiki project, figure it’ll scare people off. Neustrom &co. didn’t intend to keep the “DavisWiki” name for very long — it was especially weird in 2004 — but it stuck. Neustrom thinks it made the site feel less generic, which was important.

My hunch: the word “wiki” does scare some people off, but for those who figure it out, it’s sort of a secret they’re proud to know. A shibboleth!)

Here’s a postulate: raising the entry barriers to an online community increase user loyalty once people make it in.

For a wiki, which depends heavily on superusers, this tradeoff is sometimes a net benefit.

Indeed, comscore / alexa are inadequate when it comes to small sample sizes and certain demographics.

I’ve emailed Michael a webserver log of a day’s worth of edits and a python script to compute the number of uniques on daviswiki.org so he can verify.

I appreciate your skepticism, and maybe I should have had more. But based on my rough understanding of how those services work, his answer seemed plausible.

If you or anyone else would like to weigh in on this, I’d invite it.

a quick search with Comscore, Quantcast, Compete, and the dreadful Alexa puts the traffic nowhere near the 10,000 uniques a day you cite.

This stunning accomplishment in strong community websitery and hard work by the founders shouldn’t be associated with such misleading and overheated stats.