 We're next going to turn to Harlan Yu, who is to my left. Harlan's going to talk to us about FedThread.org, a project that he launched when he was at Princeton. And you're now part of the US government. So you're here in your official capacity, right? For the time being, yes. OK, so to briefly describe our site, FedThread.org. It's an online interface to view the federal register, similar to what the GovPulse guys did. But the idea is a lot simpler. It allows users to annotate the federal register in its documents at the paragraph level. So users can click on any paragraph of the register and leave a comment or note on any specific section, with the idea being that citizens will come to our site to discuss the merits of federal rulemaking proceedings. The site also helps users find federal register documents using our full-text search engine. And then they can track the issues that they care about based on those search terms, using customizable RSS feeds or email notifications. And so the site was built by me and three fellow computer science graduate students at Princeton University last October, over the course of eight days from conception to release. And looking back, I'd say that there were three primary raw inputs that made this site possible. Lots of coffee, lots of Diet Coke, and the XML version of the federal register. Everyone knows about Moore's Law, where computational power is increasing and its price decreasing at exponential rates. But nearly as important as the GovPulse guys just alluded to, and at least in our realm and for our work, has been the trend and the availability of powerful open-source tools, web frameworks like Django and JavaScript libraries like jQuery, that make large-scale web development significantly easier, faster, cheaper, and more accessible to more people than ever before. We built Fedthread.org using these tools and others, and then propped it up on a scalable web hosting infrastructure the day before we released. There was no large upfront investment in capital for either software or hardware. In fact, most of our limited resources were spent on our web designer, who took our bare site with the logic already built in and styled it beautifully to present our site in a way that was usable. These technological trends, the increased capabilities to present and visualize large amounts of data and put it up online, has changed the underlying assumption about how government should prioritize its IT investments as it tries to inform the public about what it does. And the digital federal register, in my opinion, is really a model for how modern open government is done right. They released their machine-partsible XML first. And only after they did did they start building their Federal Register 2.0 project, which will be released next month. They also released their XML early, knowing full well that this data set wasn't in perfect shape. But it was good enough for us to figure out all the pieces. And by putting the data set through real use, we were able to provide them with substantive feedback about which improvements to the data set would make our work and other developers work easier. And now after these iterative improvements, they're building out the official Federal Register 2.0 project on the same XML that we're still using today. So rather than declaring that their own project is the one and only way to read the Federal Register online, they first created a level playing field where we could build FedThread.org with the bells and whistles that we thought were useful, simultaneous with the GovPulse team and the work of the GPO and NARA doing their work. And with features that government might not want to risk implementing or are prohibited from implementing themselves. Civic innovation, and indeed all innovation, is always about trial and error. It's more often than not about errors and then iterating to improve and fix these errors. And as it turns out, eight months after releasing FedThread, our little site that allows ordinary citizens to discuss government regulatory documents in their free time has shockingly not been the big hit that we all imagined. But that's OK. From government's point of view, this experiment cost them nothing other than putting out the XML version of the Federal Register early in their work cycle, which they're now using anyway for their own purposes. And it's provided a real world lesson about how most users may or may not want to interface with the Federal Register. Now as a tax-paying citizen, I would much rather have government enable others to experiment than spending tens of thousands of dollars and a good part of a year on an RFP to build out an idea that might be the right answer, but perhaps to the wrong questions. And this is why the Law.gov movement is so appealing to me. The one thing that the private sector can't do and that only our government is able to do is to publish the raw data that underlies our democracy. Publishing raw data is a low-risk proposition and is an extremely efficient way of spending taxpayer dollars. And while publishing data the correct way in a useful manner in a way that's easy for developers is admittedly a lot of hard work. But we know how to do that right more or less in a way that can maximize and harness the creativity of civic innovators in order to bring government and its citizens closer together. Thanks. Fantastic. Thanks so much, Harlan. If you did that in eight days, I can only imagine what you can do in another eight days, given the time and resources.