Wiki Process ============= Talk given at `Eric von Hippel's`__ Innovation Lab Meeting at the `Sloan School of Management`__, June 7, 2007. :Author: `Benjamin Mako Hill`__ :Contact: mako@atdot.cc :Date: Thursday, 7 June 2006 :Affiliation: The `MIT Media Lab's`__ `Computing Culture Group`__ __ http://web.mit.edu/evhippel/www/ __ http://mitsloan.mit.edu/ __ http://mako.cc __ http://www.media.mit.edu __ http://compcult.media.mit.edu Introduction ============= Quick Questions to gauge the audience: * Who has edited a wiki? * Who has a user page on a wiki? * Who has edited a talk page? * Who has no idea what I'm talking about? Today I'm going to talk about something that, for lack of a better term, I am calling wiki process. About Me --------- * Free and open source developer since I was 12: - Programmer for almost 15 years and never "sold" a line of code; - Debian__ and Ubuntu__ projects; * Sometimes social scientist; * Researcher into application of free software principles to non-software areas. __ http://www.debian.org __ http://www.ubuntu.com About Wiki Process -------------------- Quick overview of the key components of what I'm calling wiki process: 1. Lowering the bar to and fostering contributions by a community of users; 2. Facilitating and ensuring high quality products by lowering the cost of response to abuse, disagreements,and mistakes; 3. **Focus:** Extensively documenting process to inform future decision-making and to, ultimately, enrich the final product. Benefits ---------- * All of the benefits that folks here are familiar with: more minds, more ideas, more innovation, etc; * The ability to tackle whole new sets of problems and turn replace a scarcity of human resources with an abundance. e.g., Talking to SJ about conflict resolution code: What we have lots of people who are willing to working something out by hand. What we do haven't is any way to help them solve that problem; * Increased documentation and accountability of process; * Reduction of risk of "opening up" a process to increased community participation. Scope --------- Wiki process happens in a variety of places: * One of those places, and one whose results we're familiar with, is wikis (like Wikipedia); * Another is Free and Open Source Software practices; * We also see wiki process in many well organized structured brainstorming sessions that some companies are employing; * Many instances of user generated content; * Many of the types of behaviors that many in the popular press are calling "crowd sourcing." Wiki Process ============== Fostering Community: Lowering a bar to community participatoin is reasonably easy (reduce logins, open up repositories, ask for advice, hold open meetings, etc). Balancing that with an environment where feel that contributing, one way or another, is worth doing requires a lot more work. Institutional independence, models of volunteers, systems of reward and recognition, and more need to be taken into account in a context specific situation. - e.g., Interview with Trolltech: This is not marketing or advertising and it's not about asking, "how can we get people to work for us for free." Rather it is about creating an environment were users are an empowered part of the process of creation because in those situations everyone benefits. - The cubicle effect. Building High Quality Products: An open process is only valuable -- to the initiator or to the community who is contributing -- if the quality is kept high. Systems must have built in mechanisms that ensure high quality. In traditional systems, this is done by restricting access or contributor status to a small number of people who can be trusted to provide good contributions. In wiki process, it mus instead be about building a system (technologically and socially) that makes responding to and rectifying abuse and quality issues easier. Process Documentation: Wiki process should not only provide tools for ensuring quality and participation but for creating "rich" documentation processes. Wikis, and other projects I'll show, do not only create the most product but information where the process itself is documented by the actions themselves. The result is information that is highly qualified and much richer and more useful than in tradition methods. History and Background ======================== Free Software -------------- - Free licenses (1) - Anonymous checkouts (1) - Internet collaboration (2,3) - Version Control Systems (2,3) Early Wikis ------------ The original wiki, (known as the `c2 wiki`__), was designed to just be a website that anyone can edit: * No version control; * No talk pages; * No history; The idea was that anyone should be able to quickly and easy create web content. So far, that's just 1. __ http://c2.com/cgi/wiki Wikis Today ============= A quick walk through a Wikipedia article that I have participated in that will try to show you how Wikipedia creates something that is much more interesting than just an article. Article__ (0) The extent of what most people see on Wikipedia. Roughly a 3000:1 view to edit ratio. Eighth biggest website in the world. __ http://en.wikipedia.org/wiki/Vegetarianism `Edit Boxes`__ (1) Every page allows editing, including anonymous editing. - Quick description of wiki-markup (what I talked about before); - Links to languages; - This is where early wikis stopped; - Edit summaries (hinting at 3); __ http://en.wikipedia.org/w/index.php?title=Vegetarianism&action=edit History__ (2,3) - We can get back to any version of the article including the original__; - Walk through the `history of semi-vegetarianism`__ over a few edits. __ http://en.wikipedia.org/w/index.php?title=Vegetarianism&action=history __ http://en.wikipedia.org/w/index.php?title=Vegetarianism&oldid=293443 __ http://en.wikipedia.org/w/index.php?title=Vegetarianism&diff=next&oldid=104977976 Identifiable Editors (2) Not everybody logs in but everyone is identifiable. We can look at `the person`__ who created the vegetarianism article or at someone more recent. Here's an example of vandalism__ that I reverted this morning. Also important are User Pages (example__) which are a rough indicator of a user who has stuck around and is a community member. It's important to note that both examples were pseudonymous. What's important is identifiable and *blockable* editors, not necessarily people with known names and expertise. Aaron Swartz's `Who Edits Wikipedia`__. __ http://en.wikipedia.org/wiki/Special:Contributions/Cpt __ http://en.wikipedia.org/wiki/Special:Contributions/124.181.46.146 __ http://en.wikipedia.org/wiki/User:Benjamin_Mako_Hill __ http://www.aaronsw.com/weblog/whowriteswikipedia Revertability (2) That provides a quick segway into revertability which can be seen by looking at the diff for the last edit or by "rolling back" or undoing changes. In all situations, everything is tracked. *Quick examples from the latest edit on the vegetarianism article.* Watchlists__, Wikiprojects__ (2) Most wiki editors are able to respond because there are groups of people watching this. There are feelings of ownership created in individual projects. __ http://en.wikipedia.org/wiki/Special:Watchlist __ http://en.wikipedia.org/wiki/Wikipedia_talk:WikiProject Robot and Machine APIs (2,3) Certain routine operational, and reversion of the worst types of vandalism can be caught automatically. SmackBot__ is an example of a bot. The system can be expanded and interacted with via external editors, programs, etc. This was not built into the system. Rather, users have created frameworks for working with and improving the system. __ http://en.wikipedia.org/wiki/User:SmackBot `Talk Pages`__ (3,-1) - Section on 'What does "herbivore"...' - Section on semi-vegetarianism. __ http://en.wikipedia.org/wiki/Talk:Vegetarianism Social Systems (2,3,-1) * Tags__ * NPOV__, TRR__, `Be BOLD`__ They serve to frame discussion and creating community norms that promote productive collaboration among a diverse group. They also serve to exclude certain groups of people or types of contributions, so must be used carefully. __ http://wiki.laptop.org/go/Autoreinstallation_image __ http://en.wikipedia.org/wiki/WP:NPOV __ http://en.wikipedia.org/wiki/WP:TRR __ http://en.wikipedia.org/wiki/WP:BOLD Visualizations (3) As you might imagine, talk pages and history become difficult to manage on articles with tens of thousands of edits. But openness, and things like APIs and accessible database dumps, provides better ways for understand complex interactions and data. Fernanda Viégas & Martin Wattenberg's `History Flow`__ is one particularly pretty example. __ http://www.research.ibm.com/visual/projects/history_flow/gallery.htm Stable Versions (2) Having multiple version of articles is a good idea, but wiki software today have poor support for it. One example is the `Freedom Defined Wiki`__ with it `unstable page`__. __ http://freedomdefined.org/Definition __ http://freedomdefined.org/Definition/Unstable Wikis Tomorrow ================ Quick Description of TextNet: Research project that allows for *permanent* divergence and collaboration between people who have partially diverged. Do the quick demo__. __ http://epoxy.media.mit.edu:8080/ Conclusions: What *You* Can Learn ================================= You can use wikis! Good examples of organizations I've worked with include: * `Ubuntu Wiki`__: Quick anecdote about the Korean documentation team * `OLPC Wiki`__: Show `User:Php5`__ * ThinkWiki__ __ http://wiki.ubuntu.com/ __ http://wiki.laptop.org/ __ http://wiki.laptop.org/go/User:Php5 __ http://www.thinkwiki.org/wiki/ThinkWiki Just as wikis took, and have continued to take, principles from wiki General principles: Expose information, allow for meaningful contributions, record history, allow for personal expression, provide means of inter-collaborator communication, provide proper attribution.