Patent review by the public
What this entry documents is not a solution.
This practice may be ineffective or useless in the long term.
ESP's position is that abolition of software patents is the only solution.
Community patent review is a proposal from Microsoft,[1] IBM, and other large companies to increase the standard of pre-grant evaluation by patent offices. People who are worried about patents could review patent applications themselves, inform the patent office about prior-art, and request reviews of patents that seem to fail the "newness" criteria for patentability. However, raising examination standards wouldn't fix much.
Contents
Limits to effectiveness
- If the patent office has a policy of accepting software patents, then there will often be no grounds for an external reviewer to complain about any particular software patent application.
- The patent holder will be given the opportunity to rewrite their patent to avoid the complaint, so unless the complaint completely invalidates the application, the result could be the granting of a stronger patent.
- Participating in the review costs time. We thus give our resources to solve a problem that shouldn't exist, and the resources we give will never be acknowledged. This could be resolved by allowing the finder of prior art to receive some of the patent registration fee, thus creating an incentive for people to locate prior art. Furthermore, a disincentive to file patent claims where prior art exists could be created, by levying an additional fee on a patent filer where prior art is found. This way the incentives of the patent filer, and prior art searchers can be aligned with the public interest, meanwhile minimising work for typically under-resourced patent offices.
- This can backfire since enhanced review can also be used to invalidate defensive patents.
Example: peer-to-patent
(This is a modified and updated version of an explanation by Ben Klemens[2])
According to Bessen and Hunt (page 47), about 70 software patents are granted by the USPTO per day.
Meanwhile, the Peer-to-patent project reports that as of May 11th 2008, they have gotten 183 items of prior art submitted on 56 applications.[1] The project opened in June 2007, so in eleven months they have been able to advise about six hours worth of output from the USPTO.
Successes of peer-to-patent
Peer-to-patent's website was used to gather prior art which lead to Red Hat and Novell's victory against a patent troll in the case Acacia v. Red Hat and Novell (2010, USA). However, the prior art used seems to all have been mentioned by the Groklaw community,[3] so the contribution of the peer-to-patent website is unclear.
Related pages on ESP Wiki
External links
- Christopher Wong describes the peer-to-patent project
- Wikipedia: Peer-to-patent
- Japanese patent office trials peer-to-patent
- Also: Japanese original: Links to original, official document (English translations[?]: Google, bing translator)
- Peer-to-patent's introductory page
- USPTO likely to adopt 'peer-to-patent', February 2010
- Another example: http://www.innovationcommons.org/
References