(heading back to Boston from the Midwest) I’ve been thinking about when it’s a good idea to develop knowledge in an open way–by inviting anyone to contribute to a common pool. Wikipedia is a remarkable resource that has been built by volunteers who not only write, but also edit others’ work. Google often leads you quickly to the best information, even though nobody sits at Google’s HQ writing websites or picking the best ones. Google’s search results are driven by choices that millions of other people have made.

Yet the White House’s recent open discussion of “transparency” was quickly dominated by people who wanted Barack Obama’s “real” birth certificate to be released. Their comments couldn’t be deleted as irrelevant, because they did have a concern about transparency. I think their concern was simply embarrassing (to them), but I don’t get to decide what’s valuable or ridiculous in an open forum. The whole discussion was mostly unhelpful, in my opinion, because they dominated.

I recently visited Project Vote Smart, which employs hundreds of college students every year to collect and code candidates’ position papers, speeches, and votes. This is a labor-intensive model that is threatened by automated systems that promise equally good results without human labor. Yet I suspect that the careful work of Project Vote Smart is indispensable. I doubt that we can rely on a wiki or a search engine to provide reliable information about local candidates.

Reflecting on these examples, I would propose three general principles for deciding when to use an open process:

1. It works best when value-conflicts are minor or absent and information is the main issue. That means that an open process works better in science and technology than in politics or religion.

2. It works best when millions of people participate, because they can swamp small groups of cranks. But in American politics, below the presidential level, millions of people do not participate actively. An open forum about a candidate for state legislature, for instance, is likely to draw just a handful of actual contributors.

3. It works best when stakes are relatively low. If I organized a public discussion of transparency on this website, there would be few participants, but their comments would probably be well-intentioned and thoughtful. There would be no motivation to disrupt a discussion on my personal website, because my site has little or no political importance. But if the White House organizes a discussion, all of its political opponents have motives to disrupt it. The White House is powerful, and it has enemies. When power and conflict are involved, many of the old rules of politics reassert themselves–even online.