Patent review goes Wiki – August 21, 2006:

That’s the basic concept behind a pilot program sponsored by IBM and other companies [including HP and Microsoft], which the U.S. Patent and Trademark Office appears poised to green-light. The project would apply an advisory version of the wiki approach to the patent-approval process.

The issue is that patent applications have tripled in the past two decades, leaving examiners only 20 hours on average to comb through a complex application, research past inventions, and decide whether a patent should be granted.

As a result, critics contend, quality has declined and lucrative patents have been granted for ideas that weren’t actually new.

One solution is to let astute outsiders weigh in during the patent-review process, as online encyclopedia Wikipedia does, vastly increasing the information available to the patent examiner.

This acknowledges there is a quantity and quality issue with people who are approving patents. The applications are complex (certainly when I read them my eyes glaze over). Do the examiners have access to good research tools? Are the examiners good at digesting the research they find? It sounds like all one needs to do is create a patent full of buzzwords the examiner is not likely to understand.

Personally, I think the USPTO should forbid hand drawn figures. Take the Blackboard patent for an “Internet-Based Education Support System and Methods” granted this year and the basis of a lawsuit against Desire2Learn (who posted the complaint and patent). I know the systems pretty well having supported a few. However, I find these hand drawn figures of a browser screen more difficult to understand than the same figure would be of a screenshot or a CAD drawing. I figure the ubiquitousness of drawing software should make this a fairly reasonable request.

Also, I would like to see more in patents about what existing technology the patent is based upon. The major complaint from people about these patents being granted is the amount of prior art. The USPTO is dependent upon the patent applicant and anyone who reviews these applications to find prior art. In knew of students who had papers rejected because there was obviously not enough references. Why not reject a patent for the same reason? Back when the office was founded, I could understand because it was so difficult to find evidence of other’s work. But the USPTO has given patents to Google! Surely with the wealth of information out there they can tell applicants they need to provide more information about prior art? Even make them provide information about items that are similar?

This project has the potential of an RFC (Request For Comment) to the whole world. To go the wiki route and allow people to change the language of the request seems kind of scary?