This afternoon in Manhattan, the Markle Foundation hosted Beth Simone Noveck for a coming out party for her new book "Wiki Government," a work subtitled "How Technology Can Make Government Better, Democracy Stronger, and Citizens More Powerful." Noveck, you might know, is now the deputy U.S. CTO for open government. But she was once a law professor, innovating from the outside. Noveck has spent years seeing how what we know about technology can fix what we know is wrong with government. In particularly, Noveck is known for the innovative Peer-to-Patent project. In that experiment, citizen-experts go online to vet applications for software patents. The hope is that doing so helps the notoriously overworked and overwhelmed patent examiners in the U.S. Patent and Trademark Office can make smarter, faster decisions. The difference is that now Noveck gets to innovate from inside the White House. And from inside a White House that has shown in the last six months that it isn't afraid to try something out online even when they don't know how it will turn out.

It's the new White House's new media experiments that get a lot of attention. Whether it's YouTube, Facebook, or citizen-participation projects like "Open for Questions," the flashy everyone-can-participate! projects tend to get the most public notice. But Noveck has her sights set on something slightly different, something more targeted. Something more nuanced. It hasn't always been the case. Way back in the dark ages of the web --2003, in fact -- Noveck created a software platform called UnChat. What the software aimed to do was to provide the architecture for, as she wrote in the compendium Democracy Online, "deliberative practive in cyberspace." That software is no longer in use.

With the ideal of digital democracy in constant evolution, Noveck is eyeing a different end than in her UnChat days. "It's not about deliberation" anymore, she said today. And it's not about necessarily making sure that every American's voice is heard in every context. Rather, "informed by data and characterized by collaboration," Noveck's vision of the future of participatory government is one that "takes advantage of the expertise and know-how of people aren't at the center of an institution, but who are at its edges." It's about, in other words, building systems and putting into place policies that let into government those most prepared to serve it.

Take the recent Declassification Policy Forum. The National Archives' Public Interest Declassification Board asked for public feedback on how the White House should revise Executive Order 12958, the presidential directive on how sensative government information should be treated. Noveck asked the team behind the project how, in their view, it went. They were thrilled, she said. Not because the call for feedback got thousands and thousands of comments. It got a few hundred. But that represented a crowd larger than the usual suspects who participate in DC discussions over secrecy. At the same time the pool of responses was entirely manageable, substantive, and useful. That reveals a different motivation than the White House's new media team, said Noveck, where the aim is to attract as many millions of eyeballs to the White House's Flickr stream or Twitter profile as possible. The goal of her experiments, said Noveck, is to put people "to their highest and best use." The method is to create small teams of people willing to work on the problems of government. And, said Noveck, "teams work well when they are small and bounded."

(Noveck got laughs from the 50-plus person crowd by joking that, after successes with Peer-to-Patent and the Declassification Policy Forum, she plans to specialize in "ironic" experiments in wiki government. That is, ones that empower citizens to improve government processes that -- like the patent system and national security classifications -- are at their core about locking down information.)