Would You Buy A Patent License From This Man?
Ian Mount | Excerpt from eCompany Now Magazine
You may not have a choice.
In all likelihood, you have never heard of U.S. Patent 5,253,341. But if you have a website, you may be in violation of it. And someday soon, a man named Tony Brown might ask you to pay for this transgression.
Patent 5,253,341 was granted in 1993 to two inventors, Anthony Rozmanith and Neil Berinson, and it covers a process whereby a computer user at a terminal asks a remote server for information, receives it, and then decompresses it on his computer. This, as it happens, describes what takes place pretty much anytime someone looks at a picture or listens to an audio file on the Internet. When they submitted the patent application in 1991, Rozmanith and Berinson couldn't have known how common this would become, but they happened to secure the license on something that happens millions of times each day.
That license is now in the hands of Brown, the president and founder of TechSearch, a tiny Chicago-area firm with an elegant business model: It buys up patents on commonly used technologies, then approaches corporations, informs them that they are violating patents, and demands licensing fees. If they refuse to pay, TechSearch takes them to court. Brown has sold several dozen licenses for the Rozmanith patent, at $30,000 to $80,000 a pop, to clients ranging from Sara Lee to Walgreen's to Harrah's Entertainment. On every sale, TechSearch gives Rozmanith a cut, pays its lawyers, and keeps the rest for itself. It's a nice little business, and if you listen to Brown, it's practically a philanthropic enterprise. "We perform a valuable service that permits the patent system to work the way the drafters of the Constitution intended," he proclaims.
There is only one small hitch: The Rozmanith patent may well be worthless. According to a number of computer experts, it covers technology that was around long before the patent was granted. Indeed, if you ask TechSearch's detractors, the company is little more than a band of thugs, brandishing junk patents in order to extort payments from large corporations. "It is a form of legal blackmail," says Gabriel P. Katona, a New York patent lawyer who has worked for one TechSearch adversary. "A shakedown, if you will."
That may be a bit harsh. It is hardly Brown's fault that he happens to own the right to display images on websites; the U.S. Patent and Trademark Office (PTO) issued that patent fair and square. Of course, if you talk to those who follow the PTO closely, they'll tell you that the TechSearches of the world are symptomatic of a larger problem: an overworked, underfunded PTO issuing hundreds -- if not thousands -- of ill-advised patents. Indeed, while the companies that have had to pay tens of thousands of dollars to TechSearch have little love for the firm, their rage might be somewhat misdirected. After all, who's more to blame: The people who decide to squeeze you for tens of thousands of dollars? Or the government agency that blithely gave them permission to do so?
TechSearch was founded in September 1997 by Brown, a hard-faced but friendly former corporate lawyer from the Chicago firm Jenner & Block, and five other partners. None was a technologist. Indeed, somewhat remarkably for a firm whose product is technological innovation, it has no lab and invents nothing. Instead, the firm was founded on Brown's realization that there were plenty of people who held valuable patents but who didn't have the resources or the know-how to demand payment. So he began striking deals with inventors: He would buy or license their patents from them, and then see if he could get companies to pay him.
TechSearch got off to a rough start with its first purchase, in January 1998, when it agreed to pay since-bankrupt Texas chip developer International Meta Systems $50,000 plus 10 percent of future recoveries for a patent on chip emulation -- the technology that lets a processor run a piece of software designed for previous versions. TechSearch then took the patent and tried to work out a licensing deal with Intel. Intel refused to pay, so TechSearch sued Intel. Not one to do things by half-measures, Brown and company asked for damages of roughly $500 million. They lost. Indeed, the judge ordered TechSearch to pay Intel $132,500 in legal fees; TechSearch has appealed. (Meanwhile, TechSearch is suing Intel for libel and slander because an Intel spokesman remarked to the Wall Street Journal that "[TechSearch] exists solely for the purpose of purchasing patents and extorting funds from another company.")
Since then, TechSearch has had a bit more luck. It has expanded its portfolio of patents to about two dozen, and according to Brown, the company has sold between 70 and 100 licenses in a little over three years of existence. Of the company's patents, the jewel in its crown is the Rozmanith database-querying patent, which it bought in 1998, and for which it has sold several dozen licenses, according to Brown. TechSearch has also taken at least 18 companies to court over Rozmanith infringement and settled out of court by signing licensing agreements with at least seven.
To critics of its techniques, TechSearch's business model amounts to nothing more than a low-grade extortion racket. With the average price of a patent infringement lawsuit -- including appeal -- running to $1.5 million per side as of 1998, according to the American Intellectual Property Law Association, an $80,000 payout looks like a bargain. "There are two ways to get rid of these people. One is to pay them off; the other is to battle it out in court," says Robert Merges, a professor at the University of California at Berkeley school of law. "The second is so expensive that, even if the first is reprehensible, [companies will] hold their nose and pay them off."
If you ask Brown, he will tell you this is all misinformation. Brown sees the patent world as a clubby society where big companies trade their patents with other, like-size entities -- and will do anything to lock the little guys out. (And Brown has something here; during the Intel trial, the California chip giant, hiding behind a Cayman Islands shell company, made a bid to IMS's creditors to wrest the chip-emulation patent away from TechSearch.) Brown paints himself as a Robin Hood figure who provides small-town Edisons with the necessary muscle to get the money they deserve. As an example of TechSearch's good citizenship, Brown points to Rozmanith, along with the widow of his co-inventor. "[Ruth Berinson] is a struggling, elderly widow, and this has made a lot of difference to her and it has made a lot of difference to Tony Rozmanith," Brown says. "He's not a wealthy man. He's elderly. And I think it's very rewarding for him to see someone acknowledge and recognize that what he did was important." (For his part, Rozmanith says Brown drove a hard bargain on the patent, but adds that he's pleased with the work TechSearch has done. "They have the resources to pursue this in a vigorous manner, which an individual inventor can hardly do," he says.)
Certainly, helping individual inventors -- not to mention benefiting the elderly and the widowed -- is a noble goal. But there is a nagging question: Isn't downloading and decompressing images a bit of an obvious idea? Rozmanith's lawyer advised him not to speak about the issue, but Brown bristles at the suggestion. "It was a brilliant idea at the time," he says, "and the only reason people say, ‘It's crap,' is because they're looking in hindsight 10 years later."
TechSearch's opponents argue otherwise: Katona, the patent attorney who has worked for a TechSearch target, says, "The [Rozmanith] patent is so broad it would cover almost everyone ... It's a bottomless barrel. They have a license to massacre with that worthless patent." And Greg Aharonian, a 42-year-old patent researcher and owner of the Bustpatents.com website, has used his 4,500-subscriber e-mail newsletter to nip at the Rozmanith patent like an angry terrier, calling it "pathetically crappy," among other choice phrases. "The claim is just so horrible," he says. "It's a truly bad patent in every sense of the word."
TechSearch's response? In a move that doesn't necessarily bespeak great public-relations savvy for a company accused of using its patents as a cudgel, it sued Aharonian in July for violating the Rozmanith patent, since he uses compressed images on his website. "I have a lot of respect for these inventors," Brown says. "And when someone like Aharonian -- who's never had an invention in his life -- sits there and he calls this patent ‘crap' … I mean, that irritates me."
It is, of course, not surprising that Brown thinks the patent is the work of a visionary, or that his targets argue that it's the weapon of a blackmail artist. What is needed here is an independent mediator -- a central clearinghouse of experts that can render authoritative judgment on the worth of a patent. In theory, this role would be filled by the PTO. But the more one looks at the PTO, the clearer it becomes that the office might not be up to the task.
To some extent, this is because the very nature of the PTO makes it ill-equipped to cope with new technologies. In deciding whether to grant a patent, the PTO relies largely on past patent filings to determine whether an invention meets its criteria of being novel, useful, and non-obvious. In most cases, this works just fine -- if you claim to have invented, say, a new gardening tool, a patent inspector will go through all the patents on record for similar gardening tools. But whenever the PTO must face a brand-new technology -- as it did in 1981, when a court decision cleared the way for the granting of software patents -- it has little expertise and no record of prior patents. Which means that, for a while at least, it may be a bit difficult for the patent office to make appropriate judgments.
This would not be a serious problem if the PTO were staffed with a large number of technologists steeped in the history of software development. But according to Bernard Galler, a member of the executive committee of the Software Patent Institute, founding editor of the Annals of the History of Computing, and professor emeritus of computer science at the University of Michigan, the PTO did not start hiring software experts until 1994, "because their reasoning was that patents were about machines and computer scientists didn't know about machines." (The PTO confirms this; according to a spokesperson, until that time, the patent office accepted only computer science applicants who'd taken a certain amount of electrical engineering -- read: hardware -- coursework.)
Making matters worse, according to patent experts like Andy Gibbs, a charter member of the USPTO Patent Public Advisory Committee, an independent panel that reports to Congress and the president, is the fact that the patent office can't really afford a large number of software experts. Back in 1992, the PTO began following a congressional decree requiring that its budget be made up entirely of money collected through patent-filing fees. That was hard enough, but it wouldn't have been so bad if Congress hadn't also begun diverting money from the PTO's revenues to fund other programs. According to Gibbs, the result of this redirection of revenue has been that the PTO doesn't have enough money to hire a sufficient number of patent examiners to stock its high-tech areas (called "art units"), and it can't offer high enough salaries to compete with high-tech companies for new science graduates and to keep the ones it has. (The average salary for a patent examiner is $61,000 a year.)
Meanwhile, the PTO has been deluged with requests for software patents. The number of patents issued on "information processing and computers" jumped from 3,644 in 1990 to 15,898 in 2000, and the average number issued per examiner has risen from 48.6 to 53 per year. Pulling it all together, Gibbs paints a vividly dire picture of the patent office at work: "Underpaid, budgets robbed ... high attrition, and high demand by industry," he says. "[All] in the most demanded art units at the patent office." Is it possible, then, that some technology patents slipped through a quality check that should have caught them? According to Gibbs, "You can count on it."
The question before TechSearch and its opponents, then, is whether the Rozmanith patent is one that shouldn't have been granted. That will soon be decided by a patent examiner in Arlington, Va. Patent lawyer Katona and the California firm Blakely Sokoloff Taylor & Zafman -- both representing unidentified companies that TechSearch approached for licenses -- filed requests for the PTO to reexamine the Rozmanith patent's validity. The PTO accepted the requests in July 2000. That may bode poorly for TechSearch, because the PTO's acceptance of a reexamination usually means there are significant questions. According to PTO data covering the last two decades, only 26 percent of patents have all their claims affirmed on reexamination. Ten percent are thrown out altogether, and 64 percent have their claims changed. The Software Patent Institute's Galler, who has examined the patent, says he wouldn't be surprised if the patent is overturned, since the technology outlined in it is rehashed from decades before. "I was using mainframes with terminals in the 1960s," he says. "Much of the technology was known before."
A ruling on the Rozmanith patent should arrive within a year. In the meantime, Brown and company have been peddling hedge licenses, which work like this: If the PTO upholds the Rozmanith patent, it will be bulletproof -- TechSearch will be able to jack up the price and companies will have no choice but to pay. But if you buy a hedge license today for just $10,000, and TechSearch prevails in the review, you will have locked in a total licensing fee of just $60,000. (If the review goes against TechSearch, of course, that $10,000 will have been a waste of money.) So far there have been no takers, and Brown plans to cancel the offer.
But if TechSearch loses the Rozmanith patent, the company is hardly going to go away. It has plenty of other patents, covering everything from plug-and-play technology to Internet learning. And even if TechSearch were to fold, there are plenty of other companies out there engaged in similar enterprises. Mark Lemley, a professor of law at UC Berkeley, estimates that companies spend about $131 million a year examining the validity of patent infringement claims and pay another $500 million for "nuisance value" (i.e., low-cost) licensing fees.
There is, however, at least one beacon in this fog: The PTO recently added a second level of review for its newest category of patents -- business methods -- and has expanded the size and accessibility of the research materials available to patent examiners. During the last four years, members of the congressional subcommittee that oversees the PTO have also tried to stop the diversion of PTO revenue to other programs, albeit unsuccessfully. And given that the PTO has been issuing software patents for several years, its examiners now have more of a knowledgeable base to prevent them from issuing overbroad ones.
Still, considering the higher stakes inherent in the incoming wave of biotech and genome-related patents, it's natural to wonder if there's a better way to handle patents. Some people argue that the government should vastly increase the PTO's budget. Others say the patent office should be turned into a registration bureau, where the validity of a patent is examined only when someone tries to enforce it. But the most radical argument is simply this: Do nothing. Mark Lemley, one of the Berkeley law professors, points out that the overall economic cost of a few bad patents is far less than a radical increase in the PTO's budget or a redirection of its mission. Simply put, Lemley figures it's just cheaper to let some bad patents slip through.
As a matter of economics, that's a fine argument. But if the Rozmanith patent is ultimately found to be worth less than the paper it's printed on, those who paid to license it will no doubt beg to differ.