From Fruit Shed to Fortune 500: The inside story of ROLM and its continuing influence on Silicon Valley
Decades before Facebook, seven years before Apple, four young men were hard at work in a prune-drying shed designing “the world’s toughest computer.” That was the founding of ROLM Corporation, at a time when the orchards of Santa Clara County were being transformed into what would become Silicon Valley.
By 1984—merely fifteen years later—ROLM was a Fortune 500 company with worldwide offices and a park-like campus. That same year, IBM bought the company in the biggest deal Silicon Valley had ever seen. By then, Silicon Valley was the world’s center of innovation, with a hallmark culture very different from the rest of corporate America. ROLM set the benchmark for that culture by providing significant financial rewards for smart, successful work, and an environment where employees could unwind—swimming laps, playing tennis, or dining brookside. ROLM’s influence extends today, in campuses like those of Google and Cisco, where onsite masseuses and sushi chefs are commonplace.
Starting Up Silicon Valley reveals
• leadership’s challenges, doubts, and convictions, from start-up to buyout and beyond;
• how ROLM’s technological innovations disrupted two industries;
• why ROLM was known as a Great Place to Work (GPW) and how that style can influence today’s workplace;
• the dirty tricks that giant AT&T undertook to smash competition that threatened its domain; and
• the hopes and frustrations of an IBM merger, from both sides of the story.
Humorous anecdotes and the wisdom of some of Silicon Valley’s most respected leaders make Starting Up Silicon Valley an intimate story of one of the Valley’s most important and culturally influential companies.
|Publisher:||Emerald Book Company|
|Product dimensions:||6.10(w) x 9.00(h) x 1.20(d)|
About the Author
Katherine Maxfield distributed phone books one summer during high school in Ohio. That did not lead to her 25 years of work (as an employee or as a consultant) for technology companies in Silicon Valley. An MBA did, coupled with her innate (as in, who knew?!) ability to translate technology into features and capabilities that customers needed and could relate to—marketing.
After retiring from the corporate world, she turned from writing marketing and business plans to writing fiction, with the help of an MFA in creative writing. Her short fiction has appeared in numerous literary journals, and her personal perspective nonfiction has appeared in the San Francisco Chronicle.
She lives in Saratoga, California, with her husband.
Read an Excerpt
Chapter 1The Roads to ROLMRice Sets Roots in Silicon Valley A jet fighter from Moffett Field crashed in downtown Palo Alto just blocks away from where Burt McMurtry, age 22, was interviewing over lunch for a job at Sylvania. The Cold War was at its chilliest that spring of 1957. Almost all technology work in the Bay Area supported the American military. Government grants funded most university research and government purchases drove the revenues of the area’s major firms, like Sylvania, Lockheed, Philco, Varian, Hiller, and Ampex. McMurtry’s interview proceeded undisturbed by the plane crash. In fact, he was surprised to read about it in the newspaper the next morning. In any case, McMurtry was not the sort of young man to be thrown off course. He was about to graduate near the top in his electrical engineering (EE) class from Rice University in Houston, Texas, his hometown. He had grown up working summers as an oil-field laborer. He had worked in electronics only once, his last summer in college, for General Electric. Sylvania’s new Microwave Tube Division was focused on radar applications and the opportunity to work in early microwave research was quite appealing to him. The real clincher, though, was Sylvania’s work–study arrangement with Stanford University. Rice’s engineering degree was a five-year program, granting a BA in four years and a BS the fifth year. McMurtry was well prepared for further study at Stanford. He would be allowed time off to attend classes alongside regular Stanford students working toward a graduate degree and then on to a doctorate—with fully paid tuition while bringing in a full salary. Stanford charged companies double tuition for the privilege. The school’s tuition then was $1,000 per year (equivalent to $7,700 in 2012 dollars.) The surcharge was worth it, enabling companies to attract vital engineering talent like Burt McMurtry, who accepted Sylvania’s offer. He and his wife (Deedee Brown, Rice ’56) packed up and moved to Mountain View. McMurtry would later become chairman of the Stanford Board of Trustees. McMurtry returned to Rice each fall to recruit engineers for Sylvania. “After I talked with the students throughout the day, I’d stop by to chat with a few professors, to get their take on my initial impressions. In what I came to understand as classic Rice style, if my opinion was higher than his, the professor would say nothing, just nod and listen. But if I’d underestimated anyone, he’d say something like, ‘Maybe you could take a second look at so-and-so.’ That was my cue that there was more there than met the eye.” In the fall of 1958, McMurtry talked with Walter Loewenstern (Rice ’59 BSEE), another Houstonian. Loewenstern had been accepted to Rice and MIT for his undergraduate schooling, but Rice was tuition-free. “It would have been a financial burden for my family if I’d gone to MIT,” he said. “I wasn’t the kind of guy who would put them through that.” He lived at home during college and joined Rice’s ROTC (Reserved Officer Training Corps) to cover books and expenses. In exchange he owed the U.S. Navy two years of active duty following graduation. Loewenstern explained that to McMurtry during the interview, and McMurtry told him to call when he was clear of his obligation. In 1961, after two pleasant years with the Navy in Japan, Loewenstern stepped off the ship in San Diego and dropped a coin into a pay phone. McMurtry set him up with interviews in Sunnyvale. He drove north, accepted an offer, then went home to visit his folks for a few days before moving to Sunnyvale to work in Sylvania’s Electronic Defense Lab. In the fall of 1962, at the end of a recruiting day at Rice, McMurtry cross-checked his impressions with Professor Paul Pfeiffer, who asked, “Did you talk to Ken Oshman?” No, he hadn’t. “Too bad,” the professor said. “We’ve never been able to challenge him sufficiently.” Oshman, top in his class, happened by at that moment. Pfeiffer stopped the slight young man who wore black rimmed glasses that seemed to cover half his face. The student was clearly impatient to head out. McMurtry asked Oshman, “What will you do when you graduate?” “Get an MBA at Harvard.” Oshman had already applied and apparently felt confident of acceptance. “What will you do then?” Oshman inched away. “Start a company.” “What kind?” “I don’t know. Might be a shoe factory,” Oshman said, clearly making something up as he backed off, trying to make his get-away. As he left, McMurtry said, “Give some thought to coming to California and Stanford.” After a few months of phone conversations, Oshman was still headed to Harvard. On a flight home from Europe in March 1963, McMurtry penned a six-page letter to Oshman on onionskin paper. He emphasized the young man’s interest in both management and engineering, a combination seldom encountered. He cited some of the exciting things going on in Sylvania’s laser research and development laboratory and spelled out the Stanford work–study program. The opportunity to earn a salary and get a Stanford degree at the same time sold Oshman. He accepted Sylvania’s offer to work in laser R&D. He and his wife (Barbara Daily, University of Texas ’62), both from small towns near Houston, moved to Sunnyvale. Gene Richeson, from a small oil town in northeast Texas, was in the same graduating class as Oshman. He too went to work for Sylvania, in the same military defense division as Walter Loewenstern. He, Oshman, and one-third of Rice’s thirty-two electrical engineering graduates of 1963 went to Sylvania. “Those were pretty typical results for about ten years there,” said McMurtry, thanks to his own annual fall pilgrimages from 1958 to 1968. Rice engineering graduates are rated among the best in the nation. They remain naturally attracted to the spirit of entrepreneurship and technical excellence that thrives in Silicon Valley, and to the Stanford Graduate School of Business. The ‘Rice Mafia’ in Silicon Valley now numbers about 2,000 graduates, the largest collection of Rice alums outside of Texas. Bob Maxfield, from Wichita Falls, Texas, was the top engineering graduate of 1964. Maxfield was being heavily recruited to work for IBM in Kentucky and remained noncommittal when he talked with McMurtry about Sylvania during McMurtry’s 1963 fall pilgrimage to Rice. So McMurtry put Ken Oshman on the task of recruiting him. When Maxfield visited the Bay Area, Oshman and his wife wined and dined him in San Francisco. “It was the first time I drank wine that didn’t have a screw-top,” said Maxfield. The dinner seeded a lifelong friendship. Maxfield wanted to be on the front-line of computer technology, but Stanford’s work–study opportunity was very appealing. “IBM promised me I could play with computers in Lexington," Maxfield said. “I told them I’d rather do computer work in San Jose, not Lexington, and go to Stanford.” IBM agreed. Maxfield married ‘Mo’ Harrison (Texas Tech ‘62) a few days after graduation and they headed for San Jose. McMurtry called it his biggest recruiting loss ever, “but it turned out to be a good thing. He was going to need that computer experience.” Maxfield’s class of electrical engineers, among the last to enjoy a tuition-free education at Rice, were still studying electrical generators, transformers, and vacuum tubes. The slide rule was the engineer’s most important tool, always near at hand, offering quick answers to multiplication, division, square roots, logarithms, and trigonometry, though the answers could be reliably discerned to no more than two decimal points. All this would soon change. The new world of computers and semiconductors had sprung up in the late 1950s, largely because of work done by IBM, Texas Instruments, and Fairchild Semiconductor. Its impact was being felt in all facets of industry and was trickling into education. Rice had two digital computers: the IBM 1620 scientific computer and a research computer built by Professor Martin Graham for graduate students to run calculations. Walter Loewenstern had helped build a small part of that computer—his first experience in making a piece of electronic equipment. Graham’s computer occupied a huge room and the power supply made a piercingly shrill noise. The Stanford work–study program enabled all these young men to earn a master’s degree in electrical engineering: McMurtry in 1959, Loewenstern in 1963, Richeson and Oshman in 1965, and Maxfield in 1966. All but Richeson went on to get PhDs, finishing in three years instead of the typical six. McMurtry completed his in 1962. Loewenstern’s PhD, earned in 1966, was from the Engineering Economic Systems department. His dissertation was on the economics of using microwaves for the transmission of large amounts of power. Oshman’s degree, in 1968, was in laser physics, with a dissertation titled “Studies of Optical Frequency Parametric Oscillation” that is still cited in the field today, eleven times in 2011, forty years after publication. Maxfield, who earned his PhD in 1969, wrote his dissertation on “Computing Optimal Controls for Linear Systems with Inequality Constraints,” which could be applied to such problems as how to get a space rover to Saturn on minimal fuel. Not surprisingly, Stanford’s work–study program was a key factor in the foundation of Silicon Valley. It is part of the answer to the question “why there, why then?” often pondered about what makes Silicon Valley so persistently different from other industrial areas.It’s in the Cards Even as a teenager, Walter Loewenstern knew he wanted to start a company. “My father was always my inspiration.” His father left Germany in 1929 because “Jews weren’t allowed to work as engineers.” He settled in Houston and opened his own company in electrical work, though “it wasn’t particularly successful.” Loewenstern had always kept his eyes out for the right business opportunity. He spotted it in mid-1968 when he was working in a small Sylvania group (by then GTE Sylvania) that was chartered to come up with product proposals that used military technology in civilian applications. Loewenstern had thoroughly researched and proposed a police vehicle tracking system that he thought had great potential, but Sylvania declined to consider it. He privately pursued the idea on weekends until he decided he’d better let his boss know what he was up to. He was told, “We don’t want to do the project, and we don’t want you to do it. Either you quit pursuing this, or you quit Sylvania.” It was unclear whether he quit or was fired on the spot. His boss sent him a letter that said, basically, “You may not compete with any product GTE makes. If you form a company, GTE will buy it and fire you.” Loewenstern said, “Of course, GTE made just about every electronic product known to man. The threat didn’t frighten me—I had no assets to lose and there were abundant job openings for engineers.” He took a consulting job and continued pursuing the idea. Monday night poker brought the Texans and a few Californians to the table, with penny antes and maximum pots of $5 because no one had any money. The after-game conversation invariably swung toward, “What are you going to do when you grow up,” said Loewenstern. He rocked the table one night by reporting that he’d just quit Sylvania to pursue a business idea. Ken Oshman wasted no time in calling him after the game. Oshman was working for Burt McMurtry at Sylvania, developing non-linear optical techniques. He said to Loewenstern, “So you’re thinking of starting a company. How about we do something together?” Loewenstern said, “It was the best phone call I ever got.” Kenneth Oshman had been raised by adoptive parents in a tight Jewish community in Rosenberg, Texas. His father was a dentist, but the rest of the family ran small businesses. His mother had a dress shop, cousins had dry goods stores. Other family members were ranchers, cattlemen and cotton traders. His cousin Milton was all three, “a larger-than-life genuine Texan Jewish cowboy,” who Oshman said was like a second father to him. Oshman worked on his cousin’s ranch from the time he was 12, doing anything he was asked—chopping down huisache trees, rounding up and branding cattle. In a 1995 taped interview, Oshman said, “When I was 15, Milton gave me a checkbook and told me to go buy cotton. Buying cotton is a very technical skill. I’d just been sweeping floors in his office and he pushed me into this. He had the audacity to say here, go try it.” Oshman turned out to be better than many of the old-timers at judging the quality of the fibers from a plug cut from a bale, then making a profit selling the bale at the Houston Cotton Exchange. That experience surely played a key role in building the confidence that Oshman carried with him always. In the future, he would often repeat the scenario with others—here, go do this, when they themselves had no idea they could. Oshman and Loewenstern mulled over a number of ideas besides vehicle location, some involving defense systems. Oshman suggested they bring in Gene Richeson, saying, “Gene’s full of ideas. And he’s dealt with the military.” Eugene Richeson, born in Paris, Texas, had grown up in the tiny oil town of Talco in northeast Texas. His mother taught school and his father worked for an oil company. His business experience began at age 11, when he bought a hundred baby chicks. He raised them, sold 50 and kept 50 hens to sell eggs door to door. He grew the business to 600 hens and “signed up half the town for egg delivery in the morning before school started.” He bought a motor scooter with the profit. “That was a trend that continued into adulthood,” Richeson said, “buying bigger and bigger motorized vehicles with my earnings.” When Oshman called him in 1968, he was working at ESL—Electromagnetic Systems Laboratory, Bill Perry’s military contracts spin-out from Sylvania. Before that he'd been at Sylvania for five years on classified military projects. He’d authored several classified papers on electronic reconnaissance and countermeasures. “I jumped at the chance to work with Ken,” he said. The three of them tossed around more ideas, most involving using computers in electronic systems. But they knew nothing about computers. Oshman suggested that he call Bob Maxfield. Robert Maxfield grew up in Wichita Falls, Texas, where his father was an orthopedic surgeon. A career in medicine never interested him, and he saw no examples of entrepreneurship in his family. But he was by nature very competitive, “not in the sense of being power hungry,” he reflected in an interview, “but when something grabbed my interest, I challenged myself to be the best I could be.” He joined the Cub Scouts and became an Eagle Scout. He swam competitively from an early age, and joined the high school swim team despite being smaller than anyone else. By senior year, he was setting records. “I’d be a nervous wreck before swim meets. I had a significant fear of failure, though that was never anything my family instilled. It was just part of me.” From an early age, he enjoyed designing and building things. In the sixth and seventh grades, he drew floor plans of imagined homes, approximating the space to scale. After building innumerable airplane model kits, he drew up his own designs, built them from balsa wood, and tested them out his second-story bedroom window. He designed and built a coop for his homing pigeons and created a two-bit adding computer made of aircraft relays his pilot uncle gave him. Math, science, and engineering were key interests for him, and remain so. At IBM in San Jose, Maxfield had worked with the 1600 process control computer group, but on the periphery of computers without getting to play with them nearly as much as he had hoped. He had designed a custom input-output interface, including circuit board design, testing, and overseeing production, even writing the maintenance manual, not a part of the job he relished. He left IBM in 1967 when he received a fellowship from Stanford to work on his doctorate, though by then he was fairly sure he didn’t want an academic career. “I wanted to be involved with developing products,” said Maxfield. “But a PhD seemed like the ultimate academic challenge and I wanted to see if I was up to it.” Oshman called Maxfield in early fall 1968 and asked what his plans were when his degree would wrap up in January. Maxfield replied that he was looking at research and engineering jobs with IBM and other computer companies. Oshman said, “Gene Richeson and Walter Loewenstern and I are thinking of forming a company. Want to join us?” “I thought for about thirty seconds,” said Maxfield, though he and his wife had an infant daughter and no savings. “Ken’s commitment was all the assurance I needed. I was thrilled at the prospect of working with him, and frankly surprised that I was asked. Among my peers, he was the smartest guy I’d ever met. And not just in technology or book learning. He knew something about a lot of things. He knew about wine and had opinions about art. I knew Ripple and barely noticed art. He was out-going, but not brash, and had charisma, something I always admire in others. I’m fundamentally too shy to have charisma. But I knew I could contribute technically. It seemed like it could be a good balance. We had become good friends as couples, and our wry Texas humor provided plenty of laughs. I thought it could be fun. I was willing to give it a go. “Ken Oshman always knew what he wanted to do,” said Maxfield. “He always knew he wanted to start a company, run a company. He was a natural business genius, and he sensed that even early on. It was hard not to fall in line behind that kind of guy.” Loewenstern was the oldest at 33. Richeson and Oshman were 28, though Oshman looked older, with his black hair and signature black-rim glasses. Maxfield was 27 but looked like a kid in his mid-teens. Richeson was the only one of the four taller than 5 feet 7 inches. All of them were trim, athletic, smart, and eager to get started.Doing What? Loewenstern, Oshman, Richeson and Maxfield set a ground rule that first time they got together in the fall of 1968: whatever they did, it would not directly compete with anything that they had worked on at Sylvania or IBM. They christened their new company Datel. Oshman typed a 33-page “Plan for the Formation of Datel Corporation,” dated January 1969. The plan stated that “Datel is a totally new concept in the electronics business. It will be the first corporation whose central objective is to apply electronic systems techniques…to urban problems.” At the time, the federal government was funding programs for Lyndon Johnson's Great Society. The plan discussed multiple divisions making multiple products in the fields of law enforcement (police vehicle tracking), public safety (electronic control of street lights), transportation (devices to electronically control entry to major highways, to automatically charge bridge tolls and identify stranded motorists), and speech privacy (to block civilian monitoring of police communications). The document could have been the outline of a mature, national corporation. In the plan, Oshman named himself President and General Manager, with Richeson as Vice President and Manager of Systems Division, Loewenstern as Vice President and Manager of Product Division, and Maxfield as Vice President and Director of Research. Loewenstern was surprised to read that Oshman had named himself CEO. He confronted him: “Wait a minute, Ken. I started this ball rolling. Who made you president?” Oshman said, “I did. Because that’s what I can do best.” With hindsight, Loewenstern commented, “I’m a conflict avoider. Within three minutes I’d given in. Thank goodness.” They figured they needed $175,000 to get started. Only a few small venture capitalists (VCs) invested in start-ups with innovative ideas, among them Arthur Rock, Tommy Davis, Sheldon Roberts, and Gene Kleiner. Oshman met with all of them. None expressed interest in the business plan. Oshman next showed the plan to Jack Melchor, who was getting his new VC firm underway. Melchor loves to say, “I’m just a farm boy from North Carolina.” That farm boy earned a PhD in physics from Notre Dame. He ran Sylvania’s Electronic Defense Laboratory from 1953 until 1956, then founded Melabs, which he sold to Hewlett-Packard in 1958. He started a second company, HP Associates, an investing joint venture with HP. In 1967, he took over HP’s floundering computer division and quickly turned it around. He was in the process of setting up his own venture seed firm, Palo Alto Investments, when Oshman visited him at his office above a pizza parlor in Los Altos. After looking over the Datel plan, he told Oshman in his usual blunt style, “You’ve got four markets and five products. No one can succeed trying to do all this junk. Narrow it down.”Honing In In early ‘69, Maxfield began teaching a semester course in electronics circuits at Stanford and took on a consulting gig at Hewlett-Packard, thanks to his friend and former Rice lab partner Gibson Anderson who worked there. (Anderson had also gotten a master’s degree on the IBM/Stanford co-op program and had since moved to HP.) Maxfield’s pay was meager in comparison to what he would have been earning if he’d pursued a corporate research job, but his primary focus was helping to get this start-up under way. The Saturday after Melchor told Oshman to narrow it down, in early February 1969, the four young men met at Richeson’s house. They wracked their brains, with beer and pretzels for inspiration. They crossed off Loewenstern's original vehicle tracking system or anything else that targeted police departments, which notoriously had little money or interest in investing in new technology. All the ideas that were non-military, commercial applications were tossed out. Hours into the session, in mid-afternoon, Richeson said, “What the world really needs is a militarized version of a commercial mini-computer.” Computers built to military specifications (mil-spec)—built to withstand the temperatures and hazards of use in the field—were not new. At Sylvania, Richeson had designed systems using digital, militarized computers. Sperry Univac, IBM, RCA, and others were all soliciting military contracts to build digital computers. Each system took more than a year to develop, funded by the military, cost well over $200,000 per unit, and was uniquely built to meet the specifications of the program it was intended for, including the programming language instructions used to tell the computer what to do. When complete, the buyer, ultimately the U.S. military, owned the design, not the contractor company. That way, if the computer was needed in volume, the government could put the design out for competitive bids to get the lowest price. The long process started all over again with the next military program. Maxfield and Richeson had prowled through the 1968 Fall Joint Computer Conference (FJCC), the premier trade show of the computer industry, conveniently held in San Francisco. That show revealed the burgeoning world of mini-computers. Until then, the computer market was dominated by mainframes, built by IBM and ‘the BUNCH’—Burroughs, UNIVAC, NCR, Control Data Corporation, and Honeywell. Mainframes were temperamental, almost delicate, had the computational power of today’s simple calculator, and were the size of a large living room. Power requirements and expelled heat increased as more disk and tape storage units were connected to expand data storage capacity. Linked in multiples, these units required a specially built room, with raised floors beneath which massive cables ran between the machines. The room had to be chilled to control failure rate because the computer components failed twice as often at 83ºF as they did at 65º. Priced at hundreds of thousands to many millions of dollars, mainframes were used primarily by the government, banks, finance, and insurance firms for processing large amounts of data, such as the information needed by the Census Bureau. The 1968 FJCC opened up the whole new world of abundant, affordable mini-computers. A few had already trickled onto the market, starting in 1960 with DEC’s (Digital Equipment Corp.’s) precursor mini, the PDP-1. IBM had the 1800. Both were the size of a chunky under-counter refrigerator. At the 1968 FJCC, more than a dozen companies announced remarkably small computers the size of a toaster oven that could fit atop a desk. They were typically aimed at corporate finance departments and manufacturing process control and cost tens of thousands of dollars—“about the price of a house in Sunnyvale then,” said Richeson. The leaders were DEC, Hewlett-Packard, and a slew of start-ups. The idea of a mini-computer that met mil specs gained traction as that Saturday afternoon wore on. The market for military products was driving much of technological development. The Vietnam War was escalating. (Loewenstern had fulfilled his military obligation. The other three had draft deferments because they worked in industries critical to national defense.) The Cold War had both the East and the West armed to the teeth. The four guys honed in on the idea of making a computer that could bounce around in jeeps and stand the vibration of helicopters and aircraft carriers, in the freezing tundra and the wilting, humid jungle. They brainstormed about a computer product that would be in continuous production, with off-the-shelf availability, that could run commercially available software programs. They would tell their customers, “Datel can get it to you in a month, none of this one-year nonsense.” Rather than design the computer from scratch, which would be an expensive investment and take a long time to accomplish, they would license an existing commercial mini-computer design and adapt it for ruggedness. Customers could build a prototype using the original supplier’s commercial version, then they could buy Datel’s more expensive rugged version of that computer for deployment in the field. Further, the computer would run all the software programs available on the commercial computer, and could directly connect to standard input-output devices such as disk drives, tape readers, and keyboards. The market would be military intelligence for radars, missile firing systems, data acquisition in the field, and some encrypted applications. Maxfield, the one computer-savvy guy among them, was asked what such a device might sell for at a reasonable profit. He took a stab at “somewhere around $30K.” Whose computer should they license? Large, bureaucratic DEC and HP had obvious solutions but negotiating with them could take months. Or, worse, simply approaching them could pique their interest in the market. They might listen carefully, then decline to offer Datel a license, and pursue their own mil-spec project. They could become deep-pockets competitors against which Datel would have difficulty competing, shutting it down before it barely got started. Interest turned to Data General, the darling of the ’68 FJCC. Data General (DG) was a start-up—a serious negative. One start-up basing its product on that of another amplified the risk. But DG seemed to be doing things in a smart way. Its marketing materials were slick and contemporary. Its Nova mini-computer was powerful and cheap, touted as “the first computer with 4,000 16-bit words of core memory for under $10,000.” The price was right for Datel to ruggedize and sell for $30,000. The Nova had a housing that was stream-lined, glossy, and white, a real stand-out in the gray computer world. Maxfield said, “It looked a lot like what the Apple II looked like eight years later.” At the show, Data General had displayed the Nova running SpaceWar, one of the earliest digital computer games. People crowded around, excited to see moving, shooting icons instead of the ubiquitous static rows of eerie-green words and dots slowly marching across the screen, the standard display of other computers. The Nova used a new generation of semiconductor chips from Fairchild called medium-scaled integration (MSI), which meant that its central processing unit (CPU) had fewer than 200 chips, rather than the typical 1,000 chips. Thus the processor would be relatively small, use less power and be more reliable. Maxfield called it “very sexy compared to everything else.” By late that February afternoon, the young men were energized, feeling like they’d hit on a workable, unique idea. They agreed that Oshman would call DG to begin the conversation, assuming that he’d call first thing Monday. Oshman asked if anyone knew the president’s name. Maxfield did. In what would become his typical no-time-better-than-now style, Oshman picked up the phone, dialed Information, and asked for the home phone number of Edson de Castro in Boston. De Castro had left Digital Equipment Corporation just a year prior to co-found Data General. The man himself answered, at home on a Saturday evening. Oshman gave him a quick pitch, then asked, “Can we come talk to you about a license?” De Castro answered, “Sure, why not? But not during the week.” All thirteen DG employees were working feverishly to ship their first computer. They agreed to meet the following Saturday. Richeson sat stunned as Oshman made the call right then and there. “I wanted to yell, ‘Stop. Put the phone down now. Let’s think about this till we figure out the fatal flaw.’ I would’ve still been talking about the idea a week later,” he said. The pitch was polished and Oshman and Richeson grabbed a red-eye the following Friday, despite Richeson’s wife being due to give birth at any moment. On Saturday they met de Castro and his team at Data General’s office in a vacated beauty parlor in Hudson, Massachusetts. “They didn’t even have furniture yet,” Richeson recalled. They returned home with a memo of understanding. Datel would license DG’s Nova mini-computer under certain conditions: Datel was barred from competing in the general, commercial mini-computer market; Datel’s mil-spec computer would have to be developed without much technical support from DG since DG was busy implementing its own business plan; DG would receive five percent equity in Datel and would be paid a royalty on each rugged computer shipped. Roles were laid out: Oshman was CEO, Richeson VP Marketing, Maxfield VP Engineering and Operations, and Loewenstern VP, working with Maxfield in engineering. Talking about his engineering responsibilities, Maxfield told Rice University’s alumni magazine Sallyport in spring 2005, “It was sheer terror, because once we decided on this idea, it was all on my shoulders, and I knew how little I knew.” Oshman rewrote the business plan focused on a single, rugged, mil-spec computer. He estimated the total military computer market at $100 million. With the new plan and the memo of understanding in hand, Oshman hit the street to bowl over the venture capitalists that had previously rejected him. They hoped to raise $500,000, though they could get going on as little as $175,000. “You don’t want to start out undercapitalized,” said Maxfield.A $75,000 Start-up In mid-February 1969, the venture capitalist Gene Kleiner shook hands on a deal. “We were off and running,” Richeson said, “so I quit my job.” Kleiner backed out the next week. “I felt panicky,” Richeson said, “on the verge of telling ESL that I wasn’t leaving after all, hoping they’d let me stay. But Ken reassured me, ‘No problem, we have a back-up plan.’ It took a lot of faith to just sit tight.” Arthur Rock had expressed interest in the Datel plan but he wanted Max Palevsky’s opinion before making a decision. Rock had backed Palevsky’s Scientific Data Systems (SDS) in 1961. Xerox had recently bought SDS for the record price of $1 billion. Rock had also fared well with Fairchild Semiconductor and had recently invested in start-ups Intel and Rolling Stone magazine. His investment would be a big boost to Datel’s credibility. In The Money Game by Adam Smith, the non-fiction best-seller of 1967, Palevsky was asked what difference his $50 million share of the SDS sale had made in his life. He responded, none, that he had the same house and same friends. Richeson, Oshman, and Maxfield, along with Arthur Rock’s assistant Dick Kramlich, looked forward to meeting this down-to-earth millionaire. They flew to Los Angeles and drove to Bel Air, shocked to enter a grand circular driveway that led to a brand-new, enormous chateau that looked like Versailles to Maxfield. A butler answered the door, in full livery: a white starched shirt, black long-tailed coat, and black pants with satin leg-stripes. They were shown into a mammoth living room, where they perched for some time on the edge of a deep, velvet sofa and were served “little dried up bear claws,” said Richeson. Palevsky finally entered, wearing a baby-blue jumpsuit. In the midst of the conversation, Palevsky’s attractive new wife drifted through in a negligee. Perhaps that took the young men off their game. Palevsky apparently wasn't impressed with them or their plan. Rock declined to invest, as did all other venture capitalists who had been approached. Rock told Oshman why. “Rock said he just couldn’t invest in our miserable track record. I was going to run the company, and I’d never managed but one technician. Bob was going to be head of engineering, designing computers, and he’d never worked in computer design. Gene was going to be VP marketing and sales, and he’d never been in marketing and sales; Walter was going to head research efforts, and he’d never done anything in the computer world.” One potential investor remained on the list. Jack Melchor had turned them down once, but maybe he’d be interested in the new improved plan. They pondered how to sweeten the deal so he would not turn them away again. They decided to offer him one-fifth of the company. He would be an equal owner. Oshman, Maxfield, and Loewenstern called on Jack Melchor. Gene Richeson missed the meeting due to an emergency appendectomy. “Maxfield looked like he was in high school,” Melchor said. When asked what his goals were, Maxfield replied, “I just want to help build a successful company.” Melchor indicated that he was interested in the investment and would think about it. After the meeting, Oshman sought advice from his boss Burt McMurtry at Sylvania, where he and Lowenstern still worked. Many months prior, Oshman had warned McMurtry that he might be leaving to start a company, but they had not talked about it since then. McMurtry said, “Ken came into my office and asked if I knew Jack Melchor. I said I did. Then he said Jack was interested in investing in the new company he and some others were forming but they didn’t know much about Jack. What could I tell them? All I could say to Ken was that I couldn’t talk to him about this. Ken made another stab and I just shook my head. He left puzzled and unhappy. “But Ken worked for me. We were both employees of Sylvania. It’d be a conflict of interest to talk with him about that. An even deeper complication was that I was talking with Jack about going to work for him in Palo Alto Investments! Jack had already asked me what I thought of three guys I might know from Rice and Sylvania who had come to him about forming a company, and a fourth Rice guy who’d gone to IBM. I was happy to give my standard positive references to Jack, but anything I said to Ken would have put me into a huge conflict of interest. I wasn’t evading the issue. I truthfully just could not talk with him about this, and all I could do was sit there and say so, literally. I resigned from Sylvania within a few weeks and went to work with Jack. The situation then became clear to Ken, and our friendship was restored.” Melchor committed $15,000 of his personal money, unwilling to risk impugning the early reputation of his new venture fund with this high-risk proposition. Plus he committed to help secure a bank line of credit. He also promised to help raise a round of venture capital if two milestones were met: Datel had to show a product at the 1969 Fall Joint Computer Conference, a mere eight months away, and orders had to come in soon after the show. Years later Melchor recalled, “They said they’d rather I join on an equal basis than as a VC.” Then he laughed, “And besides, they were cheap.” In March 1969, the four founders and Melchor anted up $15,000 each. Maxfield’s paternal grandmother loaned him the money. Richeson borrowed from his dad. For years afterwards, he put a small repayment into an envelope in his desk drawer once a week. The $75,000 would quickly run out. A $100,000 line of credit, for use as needed, was secured with Bank of America. Years later, Oshman claimed that Melchor had co-signed for the loan, that the bank had required the signature of a “responsible adult,” but Melchor replied, “Not true.” Oshman bet him $1,000 and old records were dug out. Melchor won; he had not signed the note. The bank had believed in the company’s future, probably due to the confidence Melchor expressed behind the scene. Oshman paid off—with a $1,000 check embedded in a Lucite brick that Melchor could display on his desk but never cash. Upon filing incorporation papers with the State of California, the discovery was made that the name Datel was already taken. A new, unique name was promptly needed. The brain-storming session took place in the Oshman’s bedroom, sitting around the floor, on a chair, on the bed. After rejecting numerous ideas, they toyed with the first letter of each last name. Molr sounded like a dental office. They settled on ROLM, in all capital letters. Ken Oshman often joked that the only reason he was allowed in the group was because they needed a vowel. Their lawyer, John Wilson of what is now Wilson Sonsini Goodrich & Rosati, seemed unimpressed by the name. He told them, “Oh well, you can always change it later.” ROLM incorporated in California in May 1969, with Oshman, Melchor, and Wilson as the board of directors. At that bedroom meeting, the discussion turned to how many shares would be divvied up, which had to be filed with the papers to incorporate. Richeson said, “The assumption had been that Ken would have the most, followed by Walter, then myself, then Bob because he had no experience with the military. It was a rather reserved conversation, no one wanted to speak up. Ken finally said, ‘Well, how about let’s all have equal shares,’ though it was clear he was going to run things. That solidified my impression that Ken was one of the most upstanding guys I’d ever known. I repeated that story any time someone expressed doubts about whether they were getting a fair shake in the early days.” The four founders and Melchor each received 15,000 shares at $1 per share, and 4,500 shares went to Data General.A Prune-shed Start Office space was needed, ideally in an industrial complex. The most appealing was Stanford Industrial Park in Palo Alto. In 1951, Stanford University had dedicated a 700-acre portion of its 9,000-acre campus for use in attracting businesses to the region. But space there was very expensive and available only to companies prepared to build their own facilities. ROLM was certainly not ready for that. About twelve miles down the peninsula from Palo Alto, Vallco Office Park in Cupertino was just getting underway. That moniker was an acronym of the last names of the primary developers: Varian Associates, Lester, Leonard, Craft, and Orlando. Hundreds of orchards of plums, peaches, apricots, and cherries were being bulldozed at the corner of the new Interstate 280 and Wolfe Road to make way for technology research, sales, and manufacturing facilities. The same thing would happen on many thousands of acres throughout Santa Clara Valley over the next 25 years as the Valley of Heart’s Delight—a phrase coined in the mid-1940s to promote the area’s agricultural bounty—gave way to Silicon Valley, the new bounty. Rental rates at Vallco were far cheaper than at Stanford, but the multi-tenant building would not be ready for about nine months. The developer had a small retail and business strip mall that was near completion, adjacent to Vallco Office Park. Storefronts would be available there in six weeks. To tide them over, he offered space in a prune-drying shed in the orchard. On June 1, 1969, ROLM began operations in the prune shed, surrounded by the noise and dust of demolition. R, L, and M received salaries of $20,000 (equivalent to $123,000 in 2012), with $500 more for O. Without air conditioning, they sweated away over four desks and one phone. They had all the cherries, plums, and apricots they could eat. Vendors that visited the prune shed seemed dubious of the start-up’s prospects. Many required cash payment. Fortunately, Electronic Memories and Magnetics, the only company that made the mil-spec core memory arrays that ROLM needed, agreed to a regular corporate contract. “They did very well by us for many years,” said Maxfield, “until core memories became obsolete.” While the tiny company was gestating in the prune shed, the rest of the United States was experiencing broad social and political upheaval, with the New Left, counterculture, women’s liberation, and the sexual revolution. The 1967 Summer of Love had left San Francisco teeming with hippies. UC Berkeley and many U.S. campuses reeled with frequent and sometimes violent anti-Vietnam War protests. Weekend parties, even for those with regular day jobs, offered the traditional range of alcohol, with reefers passed from hand to hand and lines of cocaine laid out on a table in a side room. LSD, however, with its ‘enlightening’ capacity, was the drug of choice for those who could get their hands on it, thanks to Timothy Leary of Harvard and Ken Kesey and his Merry Pranksters in the Bay Area. Loewenstern said he “may have participated in the drug culture a bit.” Richeson, Oshman, and Maxfield were oblivious to it. Neil Armstrong and Buzz Aldrin walked on the moon that July 20. In August, with considerable relief, ROLM moved into air-conditioned space in Vallco Shopping Center, next to The Cookie Jar and across from Meyberg’s Delicatessen. A poster-board name plate was tacked to the door of 10925 N. Wolfe Road. In 1982, Technology ran an article about companies that began in strange locations. HP’s garage, ROLM’s prune shed, and Data General’s hair salon were among the highlights. “To reach stardom in technology requires many things—brains, ideas, luck—but Spartanism is mandatory. You must rise from nothing to prove you’re something…Oshman said, ‘We did a lot of very intensive, very cost-effective work in those first six months. Having a very limited budget greatly focuses your mind on getting things running in a heck of a hurry. The money was absolutely going to run out in six months, and if we didn’t have some major milestones to our credit by then, there was no way we were going to get any more.’”ReferencesThe ROADS TO ROLM i Ken Oshman, audio interview by John McLaughlin, “Conversation with Ken Oshman,” Silicon Valley Historical Association, July 19, 1995 [SVHA] ii Chris Warren, “Connecting the Dot.Coms,” Rice Sallyport, Spring 2005, pg. 20 iii SVHA iv Computer History Museum, Competing with Giants, video of Speaker Series event that presented a panel of the four ROLM founders, moderated by Leo Chamberlain, May 19, 2004 [CHM] v Stephan Wilkinson, “Born in a Garage,” Technology, February 1982, pg. 2
Table of Contents
Introduction Leslie Berlin 1
1 The Roads to ROLM 3
2 Targeting the Military 28
3 Seeking a Bigger Market 49
4 Battling a Giant 79
5 Barely Manageable Growth 107
6 A Great Place to Work 127
7 Telecom's Parade of Products 159
8 Mil-Spec Marches Forward 204
9 Drinking from a Fire Hose 213
10 A Pivotal Year 229
11 Thinking the Unthinkable 241
12 Dancing with the Gorilla 264
13 Blindsided 277
Epilogue-ROLM's Legacy 293
Appendix A Updates on the Founders 305
Appendix B ROLM Telecom Products Timeline 311
Appendix C ROLM Mil-Spec Products Timeline 317
About the Author 331