Song lyrics are some of the most searched-for topics on the Internet. This has lead to fierce competition among lyrics sites. If you scroll to the bottom of one of these websites, you’ll see the claim: “Song discussions is protected by U.S. Patent No. 9,401,941.” We are honoring this “song discussions” patent as January’s Stupid Patent of the Month.
The patent (we’ll call it the ’941 Patent) is owned by CBS Interactive and discloses a “computer-implemented system” for “processing interactions with song lyrics.” It explains that other websites display lyrics in a “static form” and suggests there is a “lack of mechanisms for increasing the engagement of users with song lyrics.” The patent suggests allowing users to interact with lyrics by allowing them to “select a segment,” displaying a “menu of options,” and allowing the user to enter an “interpretation of the selected line.”
The patent dates back to an application filed in February 2011. Although it is 23 columns long, in our view the patent does not describe any software or Internet technology that was remotely new or innovative at that time. Rather, it describes common and mundane features, such as a “menu of options,” “user-inputted text” and a “user interaction database,” and applies these features to a lyrics website. That should not be enough to get a patent.
In fairness, the ’941 Patent’s claims were significantly narrowed during prosecution. While the Patent Office often does a poor job searching for prior art, the examiner in this case did at least find the Rap Genius website (now known as Genius). As a result, CBS narrowed its claims to require that the website suggest possible comments to users based on what others have commented in the past. This means most lyrics websites likely won’t infringe the patent.
But the ’941 Patent should not have been granted even in its narrowed form. Online annotations were certainly not new. And there was nothing revolutionary in 2011 about suggesting entries to a user based on previous user data. (To take just one example, autocomplete for Google search debuted in 2004.) Simply applying these techniques to a song lyric website should not have been patentable. Indeed, the patent itself notes that its methods could be applied to any form of text.
Ultimately, patents like this one reflect the near total failure by the Patent Office to police obviousness in software and Internet patents. Any website will involve multiple design decisions. It requires choices regarding user accounts, passwords, encryption, cookies, software languages, advertising, user interface/user experience, database structure, APIs, server architecture, etc. Given the number of choices, major websites usually reflect a unique combination of those decisions.
But giving someone a patent merely for having a unique combination of features is absurd. The patent system is supposed to reward innovation that we wouldn’t have without the incentive of a patent. It should not reward routine web development with a 20 year monopoly.
This month’s patent is similar to the patent on filming a yoga class or Amazon’s infamous patent on white-background photography. In both of those cases, the examiner found some prior art but the applicant persisted by adding mundane features to the claims until the examiner could not find documentary evidence of those exact features. An applicant can effectively game the system by adding elements so obvious no one would ever write them down in a reference. Together with Public Knowledge, EFF recently filed an amicus brief [PDF] asking the Supreme Court to consider the obviousness standard in patent law and to reaffirm that examiners can reject common sense combinations of known elements.
Leaving aside obviousness, the ’941 Patent should also have been rejected under Alice v. CLS Bank. Routine web development decisions should be considered “generic” computer processes that are insufficient to elevate an abstract idea to patent eligibility. A patent application like this one, with rote recitations of basic computer functions and a bunch of boxes and lines in a flow chart, should at least draw a searching analysis under Alice. Yet the Patent Office never even raised Alice and subject matter eligibility during prosecution. We have submitted multiple rounds of comments (1, 2, 3, and 4) to the Patent Office asking it to be more diligent in applying Alice.
Fortunately, the ’941 Patent has never been asserted in litigation. But software patents like it are the raw materials behind the rise of patent trolling. Ultimately, we need broad patent reform so such patents are not issued in the first place.