Science Progress | Where science, technology, and progressive policy meet

Better Patents Through Crowdsourcing

The Peer-to-Patent Program at Work

Peer To Patent logo SOURCE: Peer To Patent Project Want to clean up the patent mess? Start by admitting government can’t know everything. Then put the public on the task.

It’s the incident that still makes lips quiver on even the most hardened Hill staffers. In the winter of 2006, it looked likely that the Blackberry would soon be silenced; the indispensable personal digital assistant was poised to fall victim to a multi-year lawsuit that pit its maker, Research in Motion Ltd., against a Virginia patent-holding firm. At issue: technology patents that cover the wireless sending of emails, many of which should never have been issued in the first place.

Research in Motion finally saved its BlackBerry by cutting a settlement check for $612 million dollars, but the same sort of patent problems that threatened that device abound in the U.S. today. Software patents’ fuzzy boundaries are widely reviled for stifling scientific innovation. But change is afoot.

With nearly a million patents backed up at PTO’s door, those examiners are being swamped by the deluge.

Over the past year, the U.S. Patent Office worked with academics and the IT industry to test out a plan that will be instantly familiar to anyone who has ever been to Slashdot, the online nerd community. Inspired by the decentralized Internet, PTO’s Peer-to-Patent project is raising hopes that one solution for tangled bureaucracy may well be putting the public at the levers of government.

Can it possibly work? Peer-to-Patent’s trial run was admittedly small-scale. From June 2007 to April 2008, just 40 applications were community vetted. But it worked well enough to earn it a second year of life and an expansion to include the related field of business methods, such as Amazon 1-Click, an online shortcut that makes buying books and other goods on the commerce site as simple as a single touch of the mouse. And as it grows, Peer-to-Patent raises the possibility that as the information economy gets ever more complex, the U.S. government won’t insist upon making sense of it alone.

The Examiner: Patent’s Single Point of Failure

Overlapping patents issued by PTO create what’s known as a patent thicket, a deep and dark underbrush of conflicting claims that is enormously expensive to cut through. Research in Motion’s out-of-court settlement check was cut even as the patents at issue were failing in the courts. Why has it been so difficult to avoid bad patents? The system has long been closed to outside help, and, internally at PTO, it’s in crisis. Seventy percent of patent examiners have considered leaving because of the unreasonable pace at which complex patents have to be considered—just 20 hours or so on each application, according to a landmark report in 2003 by Federal Trade Commission on the economic cost of our patent woes. It’s a pace set in the 1970s, when software was simpler and Windows was still a glimmer in Bill Gates’s eye.

No matter how dedicated each of those examiners might be, they are the single point of failure at PTO, according to IBM Corp’s associate general counsel Manny Schecter, one of the Peer-to-Patent project’s industry leads, “We the people,” says Schecter, “entrust the government in the form of the patent examiner to, in theory, know all the prior art that exists,” referring to the record of past innovations that go to show whether an invention is novel. But with nearly a million patents backed up at PTO’s door, those examiners are being swamped by the deluge. And they’ve been suffering alone.

Calling in the Cavalry: Government Inspired by Social Networking

Until now. Peer-to-Patent grew out of a blog post written by New York Law School professor Beth Noveck that was picked up by, which in turn inspired a series of workshops. PTO, ever eager to find a way out of the software patent mess, jumped aboard. The problem to be solved was simple: information deficiency. Noveck frames it this way: “If you want to patent a battery-powered golf club slash weed wacker,” (a genuine product marketed as the Big Daddy Driver: U.S. patent # 6,988,954) “then you can find the state-of-the-art in golf technology just by looking at the patent database.” But software—only really patented since the late 1980s, lacks the terms of art of other scientific fields, and is often undocumented and unpatented to begin with.

The upshot: software hasn’t produced same robust patent record to draw on. The answer then, is to use what the last ten years of the interactive Internet has taught us: Let’s crowdsource it. Applicants with software to patent opt-in via a simple form, after which their proposed patent is posted to The program is structured attract to “citizen-experts,” such as scientists, Ph.D.s, and programmers, all of whom can vet the application’s claims by submitting prior art that is either supportive or debunking.

Reviewers use Digg-like tools—which allow online communities to boost news stories through individual ratings—to vote on the best prior art, which goes to a PTO examiner for final determination. Along the way, these Peer-to-Patent reviewers tag the application’s claims with keywords. The hope is to use the social web’s ideas about folksonomy, or the collaborative creation of taxonomies, to evolve an accepted lexicon for the field. It’s really, says Noveck, “about opening the conversation about science.”

If all goes well, a patent examiner uses his or her limited time judging the merit of the application rather than hunting for prior art. And the successful patentee gets the confidence that comes with having withstood peer review.

That anyone can participate, either by name or anonymously, naturally raises fears that it’s an excellent chance to sink a rival innovator. But inventors, so far, are unworried. Nanomaterials researcher Blaise Mouttet, a Peer-to-Patent applicant, finds assurance in the idea a patent yea or nay is “a fact-based determination.” He wants, he says, “the best possible prior art so that I get the best possible patent.” Indeed, bad patents have a way of revealing their flaws eventually, and it’s better for a weak patent to die an early death.

But some, of course, aren’t sold on what Peer-to-Patent can accomplish. While generally a fan of the project, Michael Meurer, author of Patent Failure, worries that there is only so much that better prior art solves. Software patents, he argues, are “intrinsically vaguer and more problematic” than those in other fields. But Mark Webbink, former intellectual property officer at the open-source company Red Hat Inc., who will be heading to New York Law School to oversee Peer-to-Patent’s expansion, shakes off the criticism. “Minimal reform legislation is tough,” he explains. “Major reform legislation is damn near impossible. You can hold your breath until software patents go away, but you’re going to expire before they do.”

U.S. Commissioner for Patents John Doll argues that when his patent examiners get good prior art in front of them, they make the right decisions, and so he offers a bit of a challenge. When it comes to the state of software patents, he says, “if the public has criticisms, this is the opportunity for them to step up.”

From Better Patents to Modern Government

Peer-to-Patent is, to be sure, an exercise aimed at fixing the patent process. But it’s just as much an experiment in using the tools of the social web to create models for participatory government. “We could substitute almost any area of policy that depends on good information to make a decision,” says Noveck, whether it’s a patent examiner facing an enormous stack of software patent applications or a Capitol Hill staffer staring at a 300-page energy reform bill. Could Peer-to-Patent be pointing the way towards how we can engage the public in government?

When we talk about citizen participation in government, we’re so often talking about tapping into people’s feelings and judgments, says Noveck. Peer-to-Patent works by focusing on fact-based expertise. “The idea of a free-for-all, of putting something up on a wiki,” Noveck says, “is terrible.” But when you architect a system with structured roles and group checks on individuals, it seems like wider civic engagement online is possible.

At its core, Peer-to-Patent starts with an admission that one particular bureaucracy needs a helping hand. But why stop there? “It has to be okay for a patent examiner to say, ‘I don’t know,’” says Noveck. “And that needs to be true across government.” The first step towards a government that can cope with the complexities of the modern world might well be acceptance of the fact that it can’t do it alone.

Nancy Scola is a freelance writer in Brooklyn, NY.

Tags: ,

Comments on this article

By clicking and submitting a comment I acknowledge the Science Progress Privacy Policy and agree to the Science Progress Terms of Use. I understand that my comments are also being governed by Facebook's Terms of Use and Privacy Policy.