(Feel free to edit)
Crucible of creativity
Dirk Riehle introduces Ward. (Audience cheers)
Thank you for being here today. Hey, this really makes my day. How many people here have written a wiki? How many people ha
So I don't need to explain wiki. It's one of those things that's very simple, but it's something that you can't understand without having done it.
That wouldn't have happened without you. Give yourselves a hand. I'd really like to meet all of you and know what your aspirations are.
I'd like wikis to become a symbol of the new world of work. That's rapidly evolving, largely driven by the dropping cost of communication. We have a lot of big problems. I hope the wiki idea can make some contribution there.
I'd like to elaborate on creativity and talk about what I did in wiki that facilitates it. I think what I'll do is to state my contribution, but I'm not sure I can defend it. I'll spend the 2nd half of my talk to showing the tools I use to watch wiki work. I don't generally make these tools available, though I've made them available to a small group.
"A wiki is a work sustained by a community."
"One's words are a gift to the community."
"To collaborate on a work, one must trust." It's deep in our nature to do things together. In fact I think it's important to make a distinction, between commercial transactions and the situations where I give of myself to this collaboration, thinking only that some good will come out of it, and that's why I do it. The overwhelming message of the success is that good things happen when you trust people more than you have reason to.
"Refactoring makes the work supple". We get the ability of other people to take our ideas and mix them around. Refactoring was adopted in agile methodology, comes from mathematical parlance, means you'll rearrange things without changing the meaning. I'm imagining clay that holds still when you want it to, but accepts your expression. Computer programs and documents get very brittle very quickly. What wiki's done is say, let's accept that things are evolving, . The cool thing about hypertext is that it lets a document start small and grow to get the right size. When a page gets two long, a person says, there are really two ideas. We give names to tne ideas, we're using ideas with precision when we let people link to them.
"Anonymity relieves refactoring friction". I like to discourage people from signing. I don't want to give up on anonymity too quickly.
"Can one trust the anonymous?" If you define trust as the belief that someone will behave the way they've
"Web has been an experiment in anonymity". It was designed in the low-level protocols that you wouldn't be exposed. There's been all kinds of things built on top of it so that it can become a kind of shopping mall, thought the machinery isn't all that great for that.
"result: people can and do trust works produced by people they don't know". The real world is wondering how it is that you can trust strangers. Wikipedia is a great result, lots of people are confused as to how it can exist.
"result: clubby days of friendly internet are over"
"opportunit: reputation systems for the creative (non-transactional)" If we keep a lot more track of who you are as a participant in online communities and make that visible in controlled ways to other people (eBay being a prime example of how you can create a new space that didn't exist)… I'm thinking about something that is not connected to commerce. People can cooperate on eBay to get a transaction done, but collaboration requires, I think, a different kind of reputation. We need collaboration between the scientific community, which we represent, and the practitioners who are building those communities.
"opportunity: organizational forms supporting creative work". I think corporations, since the railroads, have existed to coordinate and deploy capital to get things done. We had lots of people speaking different languages, communication was expensive, but now it's different. I hope that wiki can be a flagship in the industry to support this kind of work that is different.
Questions?
EEK: How do you reconcile anonymity as a design goal and the reputation requirement?
WC: Let's do it extremely carefully by observing what effect it has on. I don't think it has to be complicated, but has to be careful in terms of what it reveals. Example, you get into a shopping mall and start acting up, and security guards come and ask for your ID,
RPG: Reputation can be attached to a lot of different things. It can be attached to a person, or to the words (Seb: pseudonymity, basically.)
WC: Great idea. Take notes!
Don: What about spam? In a public wiki, spam has been a problem.
WC: Spam is mostly for Google. My experiment, the Portland Pattern Repository, has been an experiment in preserving anonymity and managing spam. Less than a year ago, my site which had been doubling in size every 9-12 months, got to the point where . We'd always had a dedicated set of volunteers who would do the housekeeping. The volunteers were getting worn out. There was argument about whether some pages should remain. Because they were programmers, people would write programs to handle that. Automated edit wars.
We had one guy (or maybe two or three depending ), whose intent was clearly to take attention from the community and not giving anything back. I started getting emails from volunteers asking me In the midst of this I'm trying to have a career too… I had two weeks of travel. I created a community within the community to help me take care of this problem. I put the smallest amount of login. Basically a box saying "enter the password here", with no indication whatsoever of how to get it. You had to know me.
To manage this, I asked people for home addresses. I judged people on whether they were opening up. I was putting my server bills, my networking guys, saw who was prepared to chip in.
I set up this wiki for the community of stewards. Let me show you around. "We're a small community of givers devoted to restoring and preserving the traditions of WardCunningham's first and most interesting wiki."
When I talk to people who were there, they said that my wiki kinda plateaued, because it had largely fulfilled its purpose for programmers. Now it has a broader purpose which is to be the original wiki, but I wouldn't be disappointed if it stopped growing.
Here's a chart showing that Wikipedia is growing. I asked them, what do you want to be once you've plateaued?
New Steward Orientation. "Welcome. We have stewards becaues Ward wasn't keeping up with his own wiki by himself." "I was inspired by ClayShirky's A group is its own worst enemy." It says that there's an in-group and an out-group, and that if you're designing social software, you might as well accept that. I'm sharing it with you today, which puts you as part of the incrowd.
"Look at each of the PowerTools. Be careful, you're controlling the live wiki."
"Even when closed, the site is writable to those with the off-line password that you can share with PeopleYouTrustAndPeopleTheyTrust."
What I've learned about visualization is that as soon as you get stuff on the screen that you can understand, it raises more questions. There needs to be a certain amount of visibility as to what the people are doing.
The #1 request was mass revert, so I did that.
There's a page where you say, restrict. You can cut out 1/256th of the net at a time. I'm proud to say that people eventually emptied the restrict list. The wiki is unsupervised right now.
You can visualize RecentChanges and see the number of chars that were added and removed, and click to inspect. I show other pages that were changed by the same host in the last few days.
CountPostsByHour shows the overall distribution of activity, in a table with rows=days, col=hour. There's more posting on weekdays than weekends. Right now it's in the tens per hour - it used to be 50/hr with peaks to 300/hr. I can also visualize the activity profile for a specific page.
I can overlook edits by a specific IP addres, and truncate the octets in the IP address, going from 203.145.29.12 to, say, 203.145.*.*
So, slice & dice the logs is basically what we're doing.
I have an electrocardiogram-style display of edits, with one line per IP and time running from left to right. This one looks like a machine. Oh, it's me! (I actually tripped myself up on that one when we changed internal IPs, and banned my own script.)
This other guy
Q. Can you make that code available on CPAN
WC. I'd love it, if I can get someone more familiar with CPAN. I'd be glad to share the formatting code. You'd be surprised how small it is.
Omar Ismail, ProductWiki: Have you thought about the
WC: I've thought about the stewards changing. INside the site it says what stewards are doing with each other. They have to sign on with their real names. I wouldn't see my thinking as very sophisticated. In the long term I'd see much shallower distinction between the stewards.
Q. How many steward-hours per day would you say are necessary?
WC: For 9 out of the 10 years I was the only steward and it was perhaps 15 minutes a day. In early days. If you're not on once a week
Pete Kaminski: How did you recruit people?
WC: I started with the people who took the initiative to send me email. There are a few people who I respected, and then friends of friends. I had 100% acceptance on the first round, than 50% in the second round, mostly I think because people had either a financial or time commitment problem.
Pete: The people that they trust is an interesting heuristic.
WC: I thought it would spread that way. I don't have checks for removing the key from people . But I want the key to travel. I don't want it to be a clique.
Q: regarding the visualization, it looks like you drew inspiration from Tufte's sparklines, did you have other sources of inspiration?
WC: I don't know about sparklines, but I'm inspired by Tufte's idea of being able to view both specific data and overall trend. The screen's resolution is limited as compared to paper, but I routinely dump megabytes into the browser and use the scrollbar.
Thank you. I hope you feel like an insider.
EtceteraIs Ward's presentation available anywhere? Thanks!