How basic can you get? Well, how about the concept of hyperlinks? The first news that I saw appeared on Tuesday, June 20, in Financial Times (http://www.ft.com): British Telecom (http://www.bt.com) has a patent due to expire in 2006 that covers hyperlink technology, and the company has now decided to commercialize the technology and start claiming royalties. (Every time someone clicks on an Internet hyperlink?!) The first targets are selected Internet service providers in the U.S.

The next bit of news, which came out the same day, confirmed the 2006 expiration date and added that this is a U.S. patent, filed in 1976.

Your intrepid patent searcher decides to try to find this patent. Neither article mentioned a patent number or an inventor name, so the obvious first approach is to look in the IFI/CLAIMS database for a U.S. patent assigned to British Telecom filed in 1976 and issued in 1989. This strategy produces nothing, and indeed British Telecom has no patents filed in 1976. Maybe the Patent Office made an error in the company name? A search of patents to any company filed in 1976 and issued in 1989 produces a number of patents, but none of them looks remotely like hyperlink technology.

Could the filing date be a typo? Patents normally don't sit in the U.S. Patent Office (USPTO) for 13 years, after all. So the searcher tries for patents to British Telecom filed in 1986. This produces a reasonable number of patents still in force, but none of them a likely candidate. The next step is to ignore the hypothetical filing date and look for other non-lapsed British Telecom patents issued in 1989 (17 years before the 2006 expiration date). Again, none of the patents retrieved looks right.

In the meantime, The Wall Street Journal picks up the story and confirms the 1976 filing date by stating that British Telecom "owned similar patents in Europe, but those expired in 1996." However, the patent searcher now realizes that 1976 is probably the patent's priority (first) filing date, in the U.K., in which case the U.S. filing date would have been 1977—one year later, as provided for by the Paris Convention. But British Telecom has no U.S. patents filed in 1977, either.

Now the patent searcher starts to get suspicious and takes a look at British Telecom's U.S. patent filings by year. British Telecom has no U.S. patents filed before 1979 and very few until 1982. Could it be that this patent was assigned to a different company, perhaps a predecessor of British Telecom? Time for some serious detective work on British Telecom's history. An article from Inside Telecom (vol. 1, no. 4, January 15, 1994), recovered through Gale Group's Newsletter database, provides "A history of significant events at BT," which shows the company's connection with the British Post Office. In fact, British Telecom began in 1980 as part of the Post Office. So British Telecom didn't exist in 1976, when this patent was first filed. It wasn't spun off from the British Post Office until 1980.

The patent searcher goes back for a final search, looking for U.S. patents filed in 1977 and issued in 1989, with "Post Office" as part of the patent assignee's name. This attempt hits paydirt: US4873662, issued October 10, 1989, and assigned to The Post Office (GB). The patent title: "Information handling system and apparatus therefor." And, indeed, a June 21 article in the Manchester Guardian confirms this patent number.

In other words, the various news articles got the company name wrong, at least the company to which the patent was first assigned. Could the patent have been reassigned, at least? A look in the CLAIMS Current Patent Legal Status (CRXX) shows that it has not been reassigned.

Well then, if the British Post Office spun off British Telecom, did the British Post Office reassign any of its patents to British Telecom? Another look in CRXX shows 175 patents reassigned from the British Post Office to British Telecom—all of them on May 31, 1988; all citing the British Telecommunications Act of 1984. Very interesting! Back in the IFI/CLAIMS database, the searcher finds that, at the time of this mass reassignment, the British Post Office had 210 active patents; that is, patents issued June 1971 through May 1988. Another look back in CRXX shows that 181 of them had some sort of post-granting legal action. Of these, 175 were the reassignments to British Telecom, and the other six were lapses for non-payment of maintenance fees.

In other words, the British Post Office reassigned most of its active patents to British Telecom on May 31, 1988. But our patent of interest was still pending on that date. So our patent searcher checks another source of legal status information, Inpadoc, and finds the following legal status action record for US4873662: "Other assignments—British Telecommunications * Post Office: 19871028." The same record cites the British Telecommunications Act 1981. This certainly implies a reassignment. A call to IFI technical staff confirms that the information didn't show up in the CRXX file because this file doesn't include pre-issue reassignments.

In any case, whether or not a patent has been reassigned, the bibliographic records list only the original assignee, not the reassigned assignee. So despite the information that has shown up in all the news stories, the patent searcher will never find this patent by looking for British Telecom as the assignee.

Lambert's Rules of Reference #1: Never believe what the client tells you! Lambert's Rules of Reference #2: Never check just one database if you can check more than one.

Incidentally, this patent qualifies to some extent as a "submarine patent," given how long it took to issue. A "submarine patent" is kept submerged (pending) in the USPTO for many years—sometimes deliberately, if the patentee thinks the technology will be hot—by a long string of continuations, continuations-in-part, extensions, and appeals. By the time it surfaces (issues), other companies may have invested megabucks in the same technology. This gives the owner of the "submarine patent" a golden opportunity to torpedo the newly infringing companies; that is, demand royalties. (Note that U.S. patents filed after June 1995 are in force for 20 years from date of first filing rather than 17 years from date of issue; which should blow future "submarines" out of the water.)

One major difference between this patent and other "submarines": Even when it issued 12 years after U.S. filing, the full-blown technology didn't exist yet. Hyperlinking, along with other Internet use, has grown enormously in recent years; and indeed it was only 3 years ago that British Telecom decided to try to commercialize this patent.

Why did US4873662 take so long to issue? Those concerned could order the file wrapper (patent prosecution history) from the USPTO or an outside company—at 75 cents to 95 cents per page, and a 12-year prosecution history is sure to be massive. Less detailed information is available free from the USPTO Patent Application Information Retrieval site (http://pair.uspto.gov). Simply plug the patent number into the appropriate box; when the information screen on that patent comes up, click on "File contents history." In this case, you see several requests for time extensions and an appeal (which took 4 years to complete) after the examiner's final rejection of the patent.
 

Found at Last
In any case, we now have the patent. So what is all the fuss about? The patent's first claim states:

"1. A digital information storage, retrieval and display system comprising: a central computer means in which plural blocks of information are stored at respectively corresponding locations, each of which locations is designated by a predetermined address therein by means of which a block can be selected, each of said blocks comprising a first portion containing information for display and a second portion containing information not for display but including the complete address for each of plural other blocks of information...."

The language is certainly obfuscatory, as is true of most patents. But Derwent, a leading international patent information service, is known for writing clear technical abstracts; so the patent searcher decides to see what Derwent has to say. First interesting surprise: Derwent never picked up US4873662, even though it issued long after Derwent started including all technologies. Well, then, did Derwent pick up the rest of the family? A search for the 1976 British priority number and date, information available from the U.S. patent record, does indeed yield a patent family assigned to U.K. Post Office. This family includes laid-open and/or granted patents in 13 countries—but no U.S. patent.

Here is the Derwent abstract, which (at least for the electronically challenged) doesn't throw much more light on the subject:

"The data processing system has a telephone link (2) for transmitting information to a peripheral terminal from a processor (1) in the form of packets of information. Each of the latter is represented visually at the peripheral terminal as a written page with parallel lines of alpha-numeric characters, and/or graphically. The peripheral terminal has an input keyboard (12), for transmitting signals via the telephone link (2) to the processor (1) indicating the pages of information required.

"The information is stored by the processor (1) as information blocks, each comprising a packet of information and a second part representing details relating to the corresponding page of information. The second part of each block is stored during transmission of the information packet to the peripheral terminal."

Does this patent indeed cover all possible aspects of hyperlinking? (Keep in mind that the Internet didn't exist in the mid-1970s, when the British Post Office was doing the research for this invention.) Is it indeed valid? I would not venture an opinion on either of these questions, but others have certainly done so. An excellent article in Salon, "Stop the Web! We own those links!" by Katharine Mieszkowski (http://www.salon.com/tech/log/2000/06/21/hyperlink/index.html), mentions early innovators in Web technology such as Douglas Engelbart, Andries van Dam, and Tim Berners-Lee, and related art going back to the 1950s. She cites van Dam as being "incredulous that anyone … could claim a patent to the technology behind hyperlinking….' I can't begin to understand what they could patent, given the prior art … and our various late-'60s and early-'70s publications on implementation.'" Mieszkowski also mentions a history of Web technological developments, available on the World Wide Web Consortium (W3C) site (http://www.w3.org/History.html).

British Telecom has already sent letters to 17 major U.S. Internet service providers demanding that they pay licensing fees on hyperlinks. The courts may eventually have to decide on the validity and scope of US4873662. If they hold it to be valid and enforceable, the results could be staggering. Edlyn Simmons, patent information guru and registered patent agent, said to me in an e-mail, "The implications of that patent are awesome—possibly the end of free hyperlinking, free Web sites, free searching, and probably anonymous Web surfing."
 
Certainly if this patent is held valid, other companies will dust off patents covering Internet technologies and start trying to pan for gold. They may have to pan quickly, though. The Trilateral Offices (USPTO, European Patent Office, Japanese Patent Office) just completed and issued a report on the comparative study carried out under Trilateral Project B3b. The consensus summary on "Confirmed current practices on business-related inventions" states, "A technical aspect is necessary for a computer-implemented business method to be eligible for patenting. To merely automate a known human transaction process using well known automation techniques is not patentable." The full report is available at http://www.uspto.gov/web/tws/b3b_start_page.htm. Interesting!