This was my first exposure to Barr, who was not a bit stuffy. I discovered that he, too, was from Kansas City. Rather short and slightly stocky, partial to elegant suits that might well have been hand-tailored in London, he exuded supreme self-confidence. At fifty-four, he was at the very pinnacle of his career.
I had heard all kinds of stories about Barr, whose defense of IBM in the antitrust cases was legendary. The battle had started in the late 1960s, when IBM customers and competitors, claiming that IBM was a monopoly, sought billions of dollars in damages. In 1969, the U.S. government joined the attack and sought to break it up into several separate companies.
IBM fought back in a fierce legal campaign that lasted some thirteen years. (A top IBM executive quipped that during these years the IBM legal team had no budget, but every year exceeded it.7) Tom Barr was its commanding general and ultimately triumphed,8 in what one commentator dubbed “the most brilliant and sustained legal representation in history.”9
Barr had served in the Marines, he was proud of the Corps, and the experience had shaped his view of litigation as all-out combat. He hated to lose cases and rarely did. He was once quoted as saying, “I never know exactly what people mean by the ‘theory of the case.’ My theory of the case is ‘We’re right and the other guy’s wrong. We’re going to win and the other guy is going to lose.’ ”10 He demanded total commitment from his troops, which were famous for working harder, preparing more thoroughly, and fighting more aggressively than any opponent. In Barr’s office hung a poster depicting a scowling Marine drill sergeant with veins popping out of his neck. The caption read, “We Don’t Promise You a Rose Garden.”
I didn’t see this aspect of Barr in London. He and Raven, who already knew each other on a first-name basis, were very friendly not only to Jack and me but to each other. They shared an interest in finding a third arbitrator they could agree on.
Although they found no one to their liking in London, they ultimately found the man they were looking for in Donald Macdonald, a Toronto lawyer with extraordinary credentials. He had been a member of the Canadian Parliament and had held five cabinet posts in the first government of Pierre Trudeau. When we three arbitrators met for the first time in the fall of 1985, simply to organize ourselves and establish a schedule, I found Macdonald equally impressive in person: urbane and polished, as tall as Jack and as well dressed as Barr. At my suggestion, we made an important decision about our roles as arbitrators—that all three of us serve strictly as neutrals.11 Later I would be very glad we had taken this course.
What exactly were these two computer giants fighting about? Boiled down to its essentials, the fight was about operating system software and other sophisticated software programs, sometimes called “middleware,” which work in close cooperation with the operating system. Some background is necessary to see why this was such a big deal for both companies.
An operating system is an organized collection of software used to manage the flow of work inside the computer and provide services to other programs. Middleware is software that works closely with the operating system, using its services and providing specialized services to application programs. Software applications, such as programs to manage a company’s payroll, inventory, or customer orders, are written to run “on top of” a particular operating system and associated middleware programs. Today, for example, the operating system for PCs is typically some version of Microsoft Windows. If you own a PC and use Word or Excel, those applications are written to run “on top of” your Windows operating system.
In 1969, IBM completely dominated the mainframe market, selling about 70 percent of all computers in the world.12 The overwhelming majority of applications for mainframes were written to run on the IBM operating system.13
In the early 1970s, the Japanese government aimed to get a piece of the action for Japanese computer manufacturers, especially in Japan, which was by then the second-largest computer market in the world. As a matter of industrial policy, the powerful Ministry of International Trade and Industry (“MITI”) decided to develop the Japanese computer industry into a force capable of taking on IBM. It encouraged two companies, Fujitsu and Hitachi, to manufacture and market “IBM compatible” computers.14 These machines would be virtually interchangeable with IBM machines, able to run any application written for an IBM mainframe.
The big question was what to do about the operating system. In retrospect, Fujitsu could have saved itself a world of trouble if it had simply let its customers license an operating system from IBM.15 Instead, Fujitsu made the fateful decision to develop its own operating system, which would be IBM-compatible and contain special features of its own. At the outset, Fujitsu may have believed this would not be overly difficult. Throughout the 1970s, IBM distributed to its customers the source code for the IBM operating system. With the source code, a programmer could figure out just how IBM had programmed its software.16 Copyright constraints weren’t an obvious concern back then. (IBM didn’t claim copyright protection for its software until 1977, and even then no one knew whether copyright law applied.) But as Takuma Yamamoto would confess in his memoir, developing a compatible operating system “posed a challenge of far greater and more severe dimension” than Fujitsu had anticipated17 and would “sow the seeds of the software dispute destined to be fought farther down the road with IBM.”18
In the short run, the compatibility strategy appeared to pay off. In 1976, Fujitsu sold its first IBM-compatible mainframe, the M Series. Six years later, Fujitsu mainframes were outselling IBM’s in Japan—the first time any company had outsold IBM in a major market.
But IBM had long suspected that Fujitsu and Hitachi were copying its technology, and in 1982, it got the goods on each of them.
Hitachi fell first, in a “sting operation” in California that received worldwide attention. Several Hitachi employees were caught on videotape trying to buy IBM hardware design manuals. Criminal indictments followed for the individuals caught in the sting and for the company.19 The scandal was humiliating for both Hitachi and the Japanese nation: a tremendous loss of face. A Japanese cabinet official was quoted as saying that Japan had been “slapped in the face.”20
All this took place against a backdrop of intense economic rivalry between the United States and Japan. American firms were alarmed because Japanese firms were undercutting them in the marketplace, often through industrial espionage. Japanese firms feared that America would adopt discriminatory trade policies toward Japan. Waves of xenophobia swept both countries. American commentators claimed the Japanese were engaging in “sneak attacks.” Shortly after the Hitachi incident, a Japanese newspaper reported that an anonymous Ameri-can IBM engineer had made the following boast: “Guess you’ve been reading about what IBM did to those Japs in California. We sure got them good. It took a little time, but we made sure they knew who was boss. … [I]n a sense we are at war with Japan all over again.”21
Later that year, IBM caught up with Fujitsu as well, after acquiring a Fujitsu mainframe and some of its software. Operating software contains hundreds of programs, but IBM had to analyze only a few of these to confirm its worst fears. There was a vast amount of copying. (This was not a criminal matter, but a potential civil case of copyright infringement.)
The news immediately went to the very top of IBM. The concern was: How many more programs had Fujitsu copied—and placed in the hands of its customers? IBM had reviewed only a tiny sample. There was no telling how far the damage had spread. What was clear was that Fujitsu’s copying had to be stopped.
The first shot was fired in October 1982, when the president of IBM wrote directly to the president of Fujitsu to declare that Fujitsu had infringed IBM’s copyrights. He demanded that Fujitsu pay for these offenses and clean up its software development process.
Fujitsu parried these claims with equal vigor. First, it said, IBM hadn’t even bothered to copyright its operating system until 1978, well after Fujitsu had started selling compatible computers. Second, Japan
ese copyright law did not apply to operating system software.22 Third, IBM so dominated the mainframe market that its operating system had become the de facto standard; therefore, Fujitsu had every right to develop and maintain a compatible operating system.
The legal conflict had begun. Eight months of arduous negotiations ensued.23
On June 29, 1983, the parties signed a pair of documents—the “1983 Agreements”—that attempted to square the circle. But as you already know, these contracts didn’t solve the problem and in some respects made the relationship worse. Because they play such a prominent role in our story, I will explain a bit about them here.
The first document, called the Settlement Agreement, tried to get a grip on the copying problem. The parties appeared to have three simple goals: to identify Fujitsu’s “tainted” programs, to compensate IBM for their use, and to prevent further copying disputes. But because the contract language is quite complex, some paraphrasing may help nonlawyers understand it. IBM had the greater bargaining power, so imagine IBM saying the following to Fujitsu:
“Okay, let’s start by making a list of all the programs that arguably contain copying. You compile the list, because we don’t know how many programs are affected. Include as many programs as you want—anything you’re worried we might sue you for. We will sell you immunity for those programs. We’ll call them ‘Designated Programs’—‘DPs’ for short—a nice neutral term so you don’t have to admit to copyright violation.
“You will pay us a lump sum of $65 million. In exchange, we promise not to sue you—or your customers—for having used these programs in the past.
“As for the future, well, we could make you call up all your customers and pull the programs off the market, or we could sue your customers who are using your tainted software. But that would be terribly inconvenient for them—and you. We understand. So we’ll let you leave the programs in your customers’ hands, but only on one condition: you pay us a license fee twice a year, equal to what we charge our customers for the corresponding IBM program.” (These semiannual fees would soon dwarf the $65 million lump-sum payment.)
That addressed the first two goals of the Settlement Agreement. The third related to Fujitsu’s future programming activities and its new programs. To what extent could Fujitsu use IBM programming ma-terials? And, what use, if any, could Fujitsu make of its DPs in future programs?
Here’s where things got a little vague. Fujitsu promised to “refrain from infringing the copyright and other legally protectable rights [of IBM] under applicable law.” But the scope of these rights was not defined. Fujitsu promised to adopt an internal protective procedure to make sure its programmers didn’t have access to IBM programming materials and wouldn’t do any more copying. But there were important exceptions to this rule that were poorly defined.24 Fujitsu promised to make copies of all its new software available in a special depository so that IBM could inspect them for compliance. But exactly what material it would provide, and when, weren’t spelled out. What was spelled out was the stiff penalty for cheating: Fujitsu would have to pay three times the normal license fee and immediately withdraw the program from its customers.
Moreover, as murky as the Settlement Agreement was, the second contract was even murkier and would later cause IBM untold grief. Blandly titled the “Externals Agreement,” it was Fujitsu’s consolation prize—and gave Fujitsu hope that it might remain compatible as IBM technology changed. At the time both Fujitsu and IBM made available to their own customers certain interface information that allowed the customer to write their own application programs that would work with the operating system.25 The parties called this “external” information. In this second agreement, the two companies promised to provide, upon request, the other with external information in exchange for a license fee.
At the time, IBM may not have thought this Externals Agreement was of much consequence. IBM had no interest in licensing information pertaining to Fujitsu’s operating system. But Fujitsu felt this second agreement, and a license to use IBM interface information, was essential if it was to maintain IBM compatibility. And it was willing to pay for this information.
But the Externals Agreement didn’t define in any detail what counted as external information. Nor did the parties even bother to fill out the exhibits, which should have listed the programs covered by the contract—and the price.
They were on a collision course of their own making.
It wasn’t long before IBM once again found what it considered flagrant evidence of copying, and it showed no mercy. IBM demanded that Fujitsu pay triple fees for all the offending programs and pull them off the market immediately. Fujitsu couldn’t possibly agree to these terms, which would have dismantled much of its software base. There were confrontations, even yelling and screaming.26
Later on, Jack and I would often ask ourselves: How could sophisticated parties represented by counsel be foolish enough to think that the 1983 deal could possibly have resolved this conflict? I am now more sympathetic. With the benefit of hindsight, I think many factors played a role, including different cultural perceptions of the obligations created by contract.27
For an American company, the details of a contract are critical. In IBM’s view, the Settlement Agreement created specific obligations that were legally enforceable. IBM had caught Fujitsu breaking its word, the contract spelled out the remedies, and IBM was entitled to them. Period!
In Japanese culture, the precise words of a contract are far less important than the underlying spirit, which should be interpreted with an eye to preserving relationships. For Fujitsu, the purpose of the 1983 Agreements was to repair a damaged relationship and create a new one in which both parties would behave reasonably. In Japan, litigation over contracts is rare. Fujitsu assumed that if new disputes arose, the parties would sit down and work out their concerns in a reasonable way. Instead, what was IBM doing? Impugning Fujitsu’s honor and issuing ultimatums. Its behavior was insulting and its demands humiliating. To meet them would be economically ruinous.
As the feelings of betrayal mounted, each side viewed the other’s behavior as increasingly outrageous. Fujitsu belatedly offered to add a slew of additional programs to the DP list, even ones that IBM hadn’t complained about. IBM thought this was sneaky and disingenuous. IBM said, in essence, “Too late! Read the Settlement Agreement! You missed the deadline. You owe us triple fees for these programs and you have to pull them off the market.”
IBM’s hard line never softened. In IBM’s view, it was losing both hardware and software profits because Fujitsu had copied. In June 1985, it sent Fujitsu a “Final Report” of its alleged offenses, coupled with a nonnegotiable demand: If Fujitsu didn’t withdraw all offending programs from the market and pay damages amounting to hundreds of millions of dollars, IBM would seek to impose these remedies in arbitration.28
As I’ve since learned, President Yamamoto saw the Final Report as a declaration of war—the final betrayal of the new relationship he had hoped to create with the 1983 Agreements. Those agreements had been terribly one-sided, in his view, but Fujitsu had shown its good faith by paying IBM vast sums of money—more than $150 million in fees since the agreements were signed. Now IBM was coming back with new demands that threatened the very survival of his company. Under a Japanese code of honor, you didn’t negotiate with such an enemy. You said through gritted teeth: “I will fight you; my son will fight your son; and my grandson will fight your grandson.”29
Both parties had chosen war. Was this wise? What would Spock have said? I will return to this question later.
The panel decided to hold the first formal arbitration hearing in Palo Alto, California, conveniently located between Tokyo and New York (it is also the home of Stanford University). The parties booked a conference room in a local hotel and made the necessary arrangements. Meanwhile, I hired Jonathan Greenberg, a former law student of mine, to work half-time with me on the arbitration as my law clerk. A gifted lawyer (and playwright), Jonathan w
ould become an indispensable part of the process and later serve as counsel to the Arbitration Panel.
On the appointed day in February 1986, I met Jack Jones and Don Macdonald in the hotel lobby, and we entered the conference room precisely at 9:00 a.m.
I was stunned by what I saw. There were at least a hundred people in the room, probably more, arranged in two hostile phalanxes like the armies of Athens and Sparta. In the front row on one side sat Tom Barr, accompanied by his lieutenants: Cravath partners and Dan Evangelista, soon to be the top in-house lawyer from IBM.30 Behind them sat their troops: associates and paralegals wielding huge black litigation briefcases. Behind them glowered a platoon of IBM executives and managers, each wearing the uniform—suit, tie, and white shirt—specified by the IBM dress code.
On the other side of the aisle, Raven commanded an army that looked even bigger than Barr’s. I guessed that Raven had told his clients something along these lines: “I know from earlier battles that IBM will spare no expense. They will try to beat you into submission by outspending you. Your only hope is to demonstrate that you will match them man for man—that you have the will and resources to fight all the way. Open your checkbooks. If they put 1,000 men on the battlefield, you should field 1,100.”
Fujitsu had done just that. Among the Japanese troops I recognized Michio Naruto, the gregarious golfer, and his immediate boss, Shoichi Ninomiya, a taciturn chain-smoker whom I had also met during my Tokyo visit. I spotted Yuri Morita, the young man who had served as our translator. Many of the other Japanese men looked familiar to me because they had been forced to sit through my lecture on American law.
Every seat in the room was equipped with a headset that could be set to either Japanese or English. The room teemed with translators. A glass booth at the back of the room had been provided for the simultaneous translators, whose murmurings would provide a constant babble during the hearing. In addition, each side had its own translator, and there was one “neutral.” A court reporter was there to create a transcript in English. Perhaps she consulted with the neutral. Who knows? A table for the panel had been placed at the front of the room. I took a seat on Macdonald’s right; Jack sat on his left.
Bargaining with the Devil Page 18