My little sister, a sweet thing who could tear your heart right out of your chest if she felt like it, once worked as the assistant to the chief patent attorney for Esmark Inc., a Chicago-based conglomerate that made STP oil treatment and Swift hams, though not in the same factory. Esmark was itself later bought by Beatrice Foods, which was then bought and dismembered by KKR, the 1980s buyout kings, leaving Esmark just a memory. But when my sister was there, the patent department at Esmark was bubbling with a great idea: The company spent $200,000 trying to trademark the word "soup." I am not making this up. It didn't work, of course, but how grand for Esmark if it had! And that's the attitude we have to keep in mind while revisiting last week's topic of the hated SBC Communications web patents.

For those who just joined us, SBC is a big American telephone company that acquired these patents when it bought the Ameritech phone company — patents that on the surface would seem to cover menus, frames, and any sort of link in a computerized document or data representation where one end is static and the other end changes based on context. SBC would like royalties for use of this patent, which is to say they want royalties from almost every web site in America, including that of the United States Patent and Trademark Office.

The Patent Office, which was too smart to think that Esmark had coined the term "soup," was not too smart to think that Ameritech had actually invented this concept — a concept that can be traced back to at least 1962! So last week, I called for readers to look for prior art — examples of these techniques being used prior to the Ameritech patent application that could be used to overturn the patent. And, as always, you readers came through by the hundreds. I received so many submissions, in fact, that I have to admit that I can't reply individually to each of you, but you have my thanks.

Having culled the duplicates and a few well-meaning but clueless submissions, I have in my hand, as Joe McCarthy said, 43 pages of prior art encompassing several hundred specific examples. Slam dunk, right? Alas, it is not that easy.

There are two problems. One is my doing and the other can be blamed on Congress and the Patent Office. I made a mistake saying that prior art was defined by the patent application date of May 17, 1996, when in fact the real date is when the inventors first wrote down the idea and documented that they had done so on a specific date. We don't know when that was, but it was probably no more than a year earlier, so we should probably consider the cut-off date to actually be May 17, 1995.

The second problem stems from U.S. patent law and how challenges are handled. Patents can be challenged under several bases, but the two that most interest us are whether what is patented is obvious and whether it was done before, which is to say there was what is called "prior art." Under Rule 56 of the patent law, applicants are themselves supposed to find the prior art and disclose it to the patent examiner. Obviously, this is a bit like asking Saddam Hussein to provide a list of his weapons, but that's the way the system works. If third parties have prior art, just as I now do in these 43 pages, we can submit it to the Patent Office for consideration. But here's where it gets sticky. The prior art claim that I might submit is included in the patent file, the patent holder is allowed to comment — to give a response to my submission that is also a part of the file — but once I have made the submission, the law says that any evidence I have provided can never be used in a subsequent court case.

Why is that? Lawyers, damned lawyers. Say I am that guy who sued Sears for stealing his idea for a ratchet wrench with one-hand control. It took years, but that fellow was eventually awarded millions. But if he had simply contested the Sears patent, the patent might have been invalidated, but the guy would have received no reward for his invention. Bummer. Worse still, if I submit YOUR prior art, I can take away your chance at receiving damages without your permission.

I'm not sure it really matters in this case, since none of the hundreds who contacted me claimed that they ought to be the patent holders, rather than SBC, but this lousy part of the law deserves exposure.

Now to the prior art, of which there is plenty.

The Wall Street Journal's Money and Investing Update, Wired, Infoseek, Yahoo, CNN, Nando, Cool Site of the Day, Netsurfer Digest, and Mirsky's Worst of the Web all had features covered by the SBC patents at least a year before the filing. So did the California Abortion and Reproductive Rights Action League.

Netscape 2.0 had frames support when it appeared in September, 1995 and that feature had been in development months before.

Some readers saw the SBC claims as covering templating. If you create your HTML by hand, back in 1995 you'd use web server technology, server-side includes (SSI), to reduce the 'grunt' work. Creating pages with SSI meant you didn't have to copy the 'look and feel' to each and every page — you'd edit it once, and use SSI to include the elements on each page.

Little Eolas, which is presently wiping the courtroom floor with Microsoft, introduced in September 1995 its WebRouser (http://www.webhistory.org/www.lists/www-talk.1995q3/0566.html ). Pay particular attention to the descriptions of the LINK tag and the GROUP tag. They did exactly what the SBC patent claims appear to describe. Support for those tags was also present in other products including a browser from the University of Calgary that appeared about the same time: http://ksi.cpsc.ucalgary.ca/archives/HTML-WG/html-wg-94q3.messages/0059.html.

Also in 1995, Sierra On-Line introduced a web based game called Stock Market Challenge — one of the first browser based games ever built. The game involved a user attempting to buy and sell stocks to their best advantage over a period of time (usually about three months). The users built and managed their portfolios by navigating through the web site using a set of graphic buttons that appeared on the top and bottom of every page. The buttons were the main navigation points for Buy, Sell, Status, Help, etc., which are covered in the later SBC patent.

But the SBC patent is not at all limited to the Internet. In fact, it never even specifically mentions the Internet. So Apple's late, lamented eWorld online service qualifies as prior art because its primary interface had unchanging pictures of buildings and places with names on them. When you clicked on them they took you to the place on that they represented. I was a daily contributor to eWorld and can say for sure that was in 1994.

Here is a Mosaic screen shot from 1994: http://www.opus1.com/www/jms/mosaic-home-pages-2-infoart.gif. The menu section on the bottom screenshot has four elements, both graphic and text, that point to sections of changing content. This is covered by the SBC patent.

O'Reilly's Global Network Navigator was the first commercial web site ever. It was also the first web portal. The GNN site had a navigation menu that was always present while the content changed. And according to O'Reilly's timeline (http://www.oreilly.com/25anniversary/timeline.html ) the GNN was live sometime in 1993.

Reading the patent, it sounds like anything that has an image tag in an href anchor would qualify, and that was possible in HTML from the very beginning. Here is a screen shot of Tim Berners-Lee's browser as of 1993: http://www.w3.org/History/1994/WWW/Journals/CACM/screensnap2_24c.gif.

The physics preprint archive predates SBC's patent by several years. The site is currently located at http://www.arxiv.org. The parts of the site relevant to the SBC patent are the links labeled 'new' and 'recent,' which point to lists of preprints that change daily and weekly respectively. The web version of this archive first appeared in December 1993.

Viewtouch is a graphic touchscreen point of sale device for the food & beverage industry that originally ran on an Atari starting in 1988. The right edge of every page had the repeating buttons even then, as it still does.

On British TV's there's a Ceefax system called Teletext. In the 1980s, a further extension called Fastext became available. Fastext was four (some static, some changing) links to other pages color-coded to four keys on the TV remote control. For that matter, wouldn't ANY TV remote control be prior art, since the button label never changes, but the on-screen result it initiates is always different?

The "HyperTIES" hypermedia browser was developed in 1988 at the University of Maryland Human Computer Interaction Lab, under the direction of Ben Shneiderman. Here is a streaming quicktime video showing a demonstration of HyperTIES: http://www.lushcreations.tv/Movies/HyperTIESDemoStreaming.mov. It's not explicitly demonstrated in this video, but the navigation buttons that are visible along the bottom edge of the screen are described by normal "storyboard" files just like the content, so that clicking on the buttons causes the content to change in the other windows, while the navigation buttons remain in place.

For that matter, Bill Atkinson's HyperCard had many of these same features in 1987.

Also in 1987, Paul Medlock's program called farVIEW seems to be prior art. It runs underMS-DOS and VMS and is still used by the Aerospace Corporation for quick access to schematics and documentation at their satellite rocket launching facilities in southern California. FarVIEW uses buttons and an unchanging background object that frames a changing foreground object, which would clearly be covered by the SBC patent had it been filed a decade earlier.

Also in 1987, Jef Raskin's Canon Cat computer, which sold 20,000 units and I wish I owned one, used an interface that clearly violates the SBC patent.

The Corvus Concept computer also used just such a button interface in 1984.

Lotus 1-2-3 had text-based menus in 1983. In this case the menu didn't appear unless you hit the '/' key, but was always in the same place, even as the spreadsheet moved around, and even if you loaded a new spreadsheet.

And finally, Doug Engelbart seems to have predated the SBC patent by 34 years! His NLS (later called Augment) system was probably the first implementation of links (and they were even called that). You could insert a link in a document such that it jumped to another document location, or sucked the content of the other document in so it looked like part of the current document. Inother words, OLE- and WAY before Microsoft, too. NLS also worked like frames, although it wasn't described as such. You could "freeze" the top [n] lines in a window and scroll the rest of the window underneath it...and this was user-controlled through the window interface. A really dramatic example of this is a video made in 1968 that shows Engelbart presenting this system to an auditorium full of people. It's completely weird to see him mousing around with a five button chording mouse, following links and creating documents on the fly 35 years ago: http://sloan.stanford.edu/mousesite/1968Demo.html.

So the patent is bogus and I have hundreds of other examples. Feel free to send a link of this page to the director of the U.S. Patent Office and tell him what you think. That is, of course, if you can find the guy's e-mail address on their confusing — but still infringing — web site.