Here's my position paper (below) for the upcoming Social Computing Symposium that the fine folks in Microsoft Research will be conducting later this month. All problem, no solution? Not if I can help it.
The vast majority of existing research into visualization tools and techniques for managing unstructured, collaboratively-developed Webs, namely wikis, is focused on one problem: identifying and preventing “bad behavior” such as spamming and content deletion. Unfortunately, such research misses the mark. For a variety of reasons, malicious outsiders rarely succeed and almost never persist in attacks upon socially-active wikis. Instead, the greatest threats to wikis come from within; in the form of duplication of effort, perpetually diminishing discoverability, inconsistent authoring and editing, and neglect. Rather than focusing all research on the identification and mitigation of deleterious usage patterns, we need to think more broadly. We need to enable the members and administrators of collaborative content development communities to identify both healthy and aberrant patterns of usage and utilize their knowledge of such patterns to improve the usefulness of their wikis in real time.
Pattern following is a durable pattern. The human mind is acutely and often unconsciously sensitive to the presence or absence of patterns. for example, i am a fan of danah Boyd. You see! We can’t help but notice pattern deviance. If given a choice, social animals generally follow established norms, conventions, laws, and precedents. We follow existing patterns because doing so is less costly—socially, politically, economically, or in time and effort—than bucking convention or creating a new PATTERN.
On the Web, layout patterns are of variable importance to the discoverability of on page items. If a designer places a single “Log In” button at the top of a Web page, most users will find and use it whether it is on the left- or right-hand side of the page. In a database however, patterns are invariably important. If a data entry operator places a customer’s home address in the Income field of a table called Company Information, the data is much less discoverable than a Log In button placed at an unconventional location on a Web page. To realize the full potential of database storage—fast access to reliable information—users must abide by strict data input patterns. Consequently, database administrators routinely create input validation rules, which prevent users from diverging from set patterns and doing things like adding an address to a field intended for dollars and cents.
The problem with wikis is that they are both Web pages and sub-relational databases. Almost all, including the canonical Wikipedia, are used to create, manage, and edit documentation of some sort. In the absence of discoverable and established editorial patterns and sans concrete validation mechanisms for administrators, wiki users all too often develop their own patterns. These include but are not limited to: topic naming conventions, information attribution schemas, deciding when and how to link to related topics or when to edit, delete, or create a new or existing topic. In the absence of easily-identifiable editorial conventions, existing wikis have a tendency to grow and evolve in an unpredictable and organic way.
By creating a flexible visualization framework and toolset that helps users discover otherwise imperceptible editorial patterns, we can harness the innate human aversion to disorder to ensure that wikis of the future grow and evolve in a much more predictable and crystalline way than they do today. More importantly, we can do so without imposing upon the unlimited editorial freedom that makes WikiWiki such a unique and powerful medium for collaborative content development.