Yesterday brought news that the U.S. intelligence community has a secret wiki, modelled along the lines of the collaboratively created Wikipedia, that it hopes will revolutionize how intelligence is shared among the nation's spooks and analysts.

A "top secret" Intellipedia system, currently available to the 16 agencies that make up the U.S. intelligence community, has grown to more than 28,000 pages and 3,600 registered users since its introduction April 17. Less restrictive versions exist for "secret" and "sensitive but unclassified" material.

The system is also available to the Transportation Security Administration and national laboratories.

Intellipedia is currently being used to assemble a major intelligence report, known as a national intelligence estimate, on Nigeria as well as the State Department's annual country reports on terrorism, officials said.

Full story.

I held off writing about it, because I really wanted to think about what the news meant.

Tellingly Wikipedia really had the scoop, since the first entry for Intellipedia showed up on September 27, about a full month before what I think was the first press mention in a U.S. News and World Report story on October 23rd (current Wikipedia entry).

But what's really interesting about wikis isn't just that people can add to and edit it. It's that wikis start with totally blank pages, and more importantly, totally unwritten processes.

That lesson was made clear to me this summer when I wrote a story about the future of wikis for Wired News and talked with Yoz Grahame, who worked a Wikipedia-like project in England known as H2G2 and who currently works as a developer advocate for a DIY application site called Ning.

Grahame told me:

Although it seems that with wikis that people are just editing text, there's something more important going on, which is the editing of structure. And quite often in the discussion parts, like the talk pages of Wikipedia and the forums going on around these thing, that's where you see process evolving.

Instead of communities changing the logic of the underlying system, they are dynamically reconfiguring their own underlying process logic.

That's how Wikipedia evolved. How do we manage this huge amount of incoming data? They evolved a process. The great thing about wikis is that a wiki is such a blank and restructurable slate, it means we are able to evolve with them.

So the question here isn't whether Intellipedia will make the National Intelligence Estimate more accurate, it's whether wikis will fundamentally alter the bureaucratic rules and processes of the intelligence community.

Would a new process emerge such that there will never be a replay of the CIA's dismissal of the Energy Department's strong dissent over the conclusion that intercepted aluminum tubes were intended for an Iraqi nuclear program?

It's hard to say if wikis can change a culture that much.

But for those interested, my wiki story this summer was turned over to readers prior to publication, so they could edit and add to it. And they did -- and the results were surprising and the process fascinating.

May the intelligence community have as much luck with wikis as I did.

Update: Lots of other have things smart things to say about the project, like Michael Hampton and Dan Farber. Thanks also to JQP and others who pointed me to the story yesterday.

- Ryan Singel