Can You Patent Common Features of the Internet?

Telcom giant SBC is sending cease-and-desist letters  to hundreds of Web site operators, accusing them of infringing on an SBC patent covering “frames,” those stationary menus that innumerable Web sites employ to help users navigate sites. SBC says individual infringers owe it licensing payments ranging from $527 to $16.6 million per year. Does SBC own such a patent? And if so, how was it allowed to patent such a seemingly obvious feature?

The answer to the first question is yes. SBC secured its first frame patent in 1999 and a second in 2002. It’s standard practice for large corporations to quickly patent the work of its engineering corps, even when the exact nature of future applications may be uncertain. It’s a process that only wealthy players like SBC can afford to engage in on a regular basis—legal fees alone can cost tens of thousands of dollars.

How could the government let the company patent a common feature like frames? Keep in mind that software patents are a relatively new phenomenon. Until the early 1980s, the courts generally considered software to be nothing more elaborate than applied mathematics, and thus not patentable. That’s why the geeks behind such pre-’80s computing wonders as the first-ever database, word processor, and spreadsheet missed the boat: An inventor has just one year from the first public use, printed publication, or commercial promotion of their invention to file for a patent.

That anti-software bias changed slowly, as attorneys became more savvy about describing the unique challenges of programming and judges got hip to the complexities of coding. In 1996, the U.S. Patent and Trademark Office issued its first guidelines intended to “facilitate the patenting of computer software.” Two years later, the federal case State Street Bank & Trust Co. v. Signature Financial Group affirmed a relatively low bar for the patenting of software, listing a wide range of criteria under which programs could be considered to provide a “useful, tangible, and concrete result”—a patent prerequisite. Essentially, State Street said that any software with a practical use had a good shot at passing muster. Computer-related patent applications nearly doubled.

Problem is, the USPTO was ill-prepared to handle the flurry. Its personnel was not particularly computer-literate nor accustomed to researching programming history. To receive a patent, an invention must be both non-obvious and new. Critics say USPTO’s staffers didn’t possess the technical expertise to decide the first issue, and they didn’t know enough about computing literature to resolve the latter. When researching a patent application, the USPTO is supposed to seek examples of “prior art,” or similar inventions from the past that might disprove the novelty of the subject in question. Typically, this involves looking through peer-reviewed journals and other traditional scientific sources. But computing knowledge has been disseminated in much more haphazard ways—through mailing lists, at open-source conferences, in university dorms. The USPTO just wasn’t clued in to where it should look for the patent realm’s equivalent of exculpatory evidence.

Thus many thousands of patents may have been granted to software bits that would have failed a proper prior-art test. San Jose Mercury News columnist Dan Gillmor recently asked his readers to dredge up prior instances of intrasite frame navigation in order to lampoon SBC’s claims. Dozens of suggestions have poured in, stretching back to Apple’s Reagan-era HyperCard and a 1994 manual for IBM’s BookManager Library Reader.

Patents are not written in stone, and many are successfully overturned amid legal challenges. In 1994, for example, the USPTO overturned a patent it had given to Compton’s New Media, which covered a broad technique whereby data is retrieved from CD-ROMs.Had it been allowed to stand, the patent might have squelched the development of digital media since potential competitors would have owed Compton’s millions in licensing fees.

And just last August, a federal judge rejected British Telecommunications’ bizarre claim that the entire Internet violated its 1989 patent on the concept of the hyperlink. BT sued the Internet service provider Prodigy, claiming that it owed massive licensing fees for hosting pages that included hyperlinks. U.S. District Judge Colleen McMahon dismissed BT’s case summarily, finding that the patent did not accurately describe the Internet. (The BT patent included language referring to a “central computer,” which the Internet lacks.) SBC, by the way, owns Prodigy.

Next question?

Explainer thanks Michael J. Madison of the University of Pittsburgh School of Law and Jeffrey Kuester of Thomas, Kayden, Horstemeyer & Risley.