Intel vs. NEC--High-Tech Battle With Big Stakes
At first glance, it appears to be just another skirmish in the battle for technological supremacy between Japan and the United States.
On one side is NEC--formerly known as Nippon Electric Co. A scrappy maverick among the consensus-seeking Japanese, the company has transcended its roots as a supplier of telephone gear and is now the world’s largest semiconductor manufacturer. NEC began life in 1899 as a joint venture of AT&T;'s Western Electric manufacturing arm and a pair of Japanese businessmen.
On the other side is Intel, a quintessential Silicon Valley success story and arguably the world’s most innovative major computer chip house. Intel’s founders include the inventor of the integrated circuit, Robert Noyce. The company is best known for inventing the microprocessor, the computer on a chip that makes desktop computing possible.
At issue is whether NEC infringed Intel’s copyrights on the 8086 and 8088 microprocessors, the “brains” in first-generation IBM personal computers, and whether those copyrights are even valid.
But there potentially is a lot more riding on the outcome, industry experts say. The trial, whose main phase will begin June 14, has all the earmarks of a landmark intellectual property case. Its outcome may help determine the pace of technological breakthroughs and how quickly they are brought to market.
“It’s an extremely significant case,” said Ronald S. Laurie, a San Francisco attorney who specializes in technology matters for the firm of Townsend & Townsend. “This is the first case to really present the defense of reverse engineering. Until now, most of the important cases in this area have been against pirates who flat-out copied.”
“This is also one of the first cases in this area where money is no object in terms of legal fees,” added Laurie, who is not involved in the case. “Both sides have the finest lawyers money can buy. . . . This case will presumably decide, or at least say something intelligent about, a whole range of issues revolving around the subject of cloning.”
Reverse engineering is a process that involves closely analyzing a competitor’s product and then attempting to design a functionally equivalent product without making an exact copy of the original. The process is legal so long as valid copyrights or patents are not infringed. Many of the so-called IBM personal computer “clones,” for example, rely on on reverse engineering.
Japanese companies, in particular, have become experts at reverse engineering. Indeed, the Japanese are believed to be ahead in the race to reverse engineer Intel’s new top-of-the-line microprocessor, the $500-a-copy 80386 chip. An Intel victory in the NEC case could throw a roadblock in the path of would-be cloners; on the other hand, a win by NEC would allow the Japanese to strengthen their toehold in the U.S.-dominated microprocessor market and perhaps other industries.
Each company argues that if the other wins, innovation would be chilled. Intel said companies would be loath to invest millions in new chip development if their work could be copied easily. NEC insists that an adverse ruling would prevent companies from introducing the proverbial better mousetrap.
Intel’s legal team is headed by Jack E. Brown of the Phoenix and Palo Alto, Calif., firm of Brown & Bain; Brown is also spearheading Apple Computer’s attack on Microsoft and Hewlett-Packard in a suit that contends the companies have infringed on copyrights covering the Apple Macintosh display.
For NEC, Alan MacPherson and Robert B. Morrill are leading the charge. They are partners in the Santa Clara, Calif., firm of Skjerven, Morrill, MacPherson, Franklin & Friel. MacPherson and Brown used to be on the same side when MacPherson was general counsel for Fairchild Semiconductor and Brown was the firm’s outside counsel.
The case also opens a window on the shifting alliances and ruthless competitive strategies of the major players in Silicon Valley, where hundreds of millions of dollars can ride on a successful chip design. Until the dispute erupted in the early 1980s, Intel and NEC were among the closest of corporate friends; in 1976, they even entered into a broad cross-licensing agreement that allowed each company to use the other’s key patents.
First Judge Disqualified
The importance of the case is not lost upon the judge who must wrestle with the arcane technological matters the lawsuit presents.
“I understand that the entire computer industry is watching,” said U.S. District Court Judge William P. Gray, the 76-year-old jurist imported from Los Angeles to oversee the case. Gray is actually the second judge who will have to grapple with the case, which is being tried without a jury in a courthouse as futuristic as the products that emerge from factories in the surrounding area.
Two years ago, after completing most of the trial but before ruling on the key question of whether NEC had infringed upon Intel’s copyrights, Judge William A. Ingram was disqualified after NEC learned that he owned $80 of Intel stock through an investment club.
Gray has impressed lawyers on both sides as a quick study. But until he went through a two-day tutorial on semiconductors and microprocessors sponsored by NEC and Intel last month, Gray said, “I didn’t know the difference between a semiconductor and third base.”
“It has been hard,” said the 22-year veteran of the federal bench. “It still is. But I listen very carefully. And I am going to know enough by the end of the trial to determine whether microcode is subject to the copyright law and whether NEC infringed.”
Microcode, which is at the heart of the legal dispute, consists of tiny instructions that are embedded in and help run a microprocessor; as such, it “stands at the very intersection between software and hardware,” said Laurie, the San Francisco lawyer. Indeed, some experts have coined the term “firmware” to describe microcode’s half-fish, half-fowl status.
Intel argues that microcode is software and thus is copyrightable. Judge Ingram embraced this view in 1986 when he wrote in a key finding that “the methodology employed in the creation of microcode is indistinguishable from that employed in the creation of any computer program.”
NEC, on the other hand, contends that the Intel microcode does not meet the definition of a computer program. The microcode, NEC says, is a part of the computer rather than something that must be used “in a computer,” an important distinction in U.S. copyright law.
“Microcode is essentially inseparable from the microprocessor itself,” NEC has argued. If the court agrees with NEC and finds that microcode is indeed part of the hardware, NEC can then point to its patent license from Intel as all the authority NEC needs to market its offending chips, the V20 and the V30.
‘Pattern of Copying’
Beyond that, NEC is arguing that, even if Intel’s microcode was copyrightable, Intel forfeited its copyrights by failing to maintain an aggressive program of requiring copyright notices on the parts made by Intel’s licensees until the dispute erupted.
And if those arguments fail, NEC’s fallback position is that it did not copy Intel’s microcode. Instead, NEC argues, any similarities between the instructions on its chips and Intel’s were required due to constraints on programming imposed by the hardware.
But Intel points to similar errors in both versions of the microcode, NEC’s unusual speed in creating its microcode and what Intel calls inadequate documentation by NEC’s developers as evidence that NEC copied Intel’s work.
Intel also cites NEC’s “demonstrated pattern of copying.” In 1982, Intel accused NEC of making “bit for bit” copies of the 8086 and 8088 chips. Without litigation being filed, NEC agreed to pay for a license on the chips in February, 1983.
NEC now says it was this brush with Intel that gave NEC the impetus to develop its own line of proprietary microprocessors, of which the V20 and V30 were the first. Independent observers say the NEC machines entail significant improvements over the Intel devices. “They consume less power and operate more quickly,” said Michael Boss, a semiconductor analyst at Dataquest, the San Jose market research firm.
When NEC introduced the parts in 1984, Intel’s senior management held a special “NEC day” to plot counter-strategy, according to evidence in the case. In December, 1984, with rumors of impending copyright enforcement action by Intel sweeping the industry, NEC struck first, filing suit seeking a declaratory judgment in its favor. Intel countersued a few months later.
To Intel attorney Brown, this is a case about preserving the U.S. technological edge in a key industry and of protecting corporate investments that lead to innovation.
“The Japanese put (American companies) out of business in memory products,” he said. “When they start copying or emulating, they have got tremendous resources. They’re very much to be respected. In terms of manufacturing prowess, these people have got it.”
NEC sees things differently. “This is not a case about copyright infringement or even about the copyrightability of microcode,” said Linnet Harlan, an NEC lawyer. “This is a case about a company trying to stop a competitor from competing in the marketplace with superior products.”
The outcome is to be determined by a no-nonsense judge with an abiding faith in the legal system, a man who unabashedly calls his job “the best in the world.”
“A judge can learn enough about anything to decide the legal issues involved,” said Judge Gray. “Unlike most technical people, a judge can bring in a bit of common sense and practicality.”