Wildfire in Silicon Valley
Processor Technology was a nexus for hobbyists making a transition, trying to be serious about it all and not always succeeding.
–Lee Felsenstein, designer of several microcomputer products
The Fourth Street garage in Berkeley was a busy place that spring. Lee Felsenstein was making a meager living from odd jobs, including repairing friends’ Altairs, while Bob Marsh was tearing open checks, writing ad copy, and doing his best to convince hobbydom that Processor Technology was a million-dollar company when, in fact, it still existed mostly in his head.
Fixing the Altair
Felsenstein had gotten himself in trouble that spring. In spreading the word about the Altair through an article for PCC, he based his description of the machine’s workings and capabilities on information he received from Homebrew and from a telephone interview with MITS president Ed Roberts. Irate letters soon poured into the PCC office contending that Felsenstein had not been critical enough of the product. The Altair had serious problems, the letters claimed. Steve Dompier, for one, showed Felsenstein the difficulties he’d had with the front panel of his Altair and even got Felsenstein to fix it.
In a PCC article he titled “Criticism and Self-Criticism,” Felsenstein apologized: “I lied folks; this thing has problems.” He detailed the computer’s flaws and how to correct them. He also began fixing Altairs for friends and PCC readers, working on them in his half of the garage. Loyal to other hobbyists and feeling guilty about misinforming people, Felsenstein did the work cheaply. In the process, he learned a great deal about those early Altairs.
Meanwhile, Marsh and Ingram were using their half of the garage to create the Altair boards they were getting checks for. But they were stalled early on: they needed a sharp engineer to draw up the schematics for the boards that Marsh had conceived. The engineer had to be willing to work in a cramped and messy garage, and he had to work cheap.
Marsh knew just the man.
Felsenstein had made it clear that he did not want to join Processor Technology or any other company. He had better things to do. Although he was working long hours for little pay, he was doing what he wanted and felt beholden to no one. And long hours for little pay was about all Bob Marsh could offer him. Nevertheless, Marsh put forward a new proposal. Would Felsenstein just do the schematic for the first board, as a consultant rather than an employee?
Felsenstein thought it over, agreed, and offered to do the schematic for $50. This price, Marsh thought, was pathetically low. It was a $3,000 job and Felsenstein, the poor goof, was offering to do it for $50. Marsh refused to go below $500. Felsenstein accepted the compromise.
It was fast work, and by June they were shipping boards. One of them was first meant to be a 2K memory board for the Altair, an ambitious project given that MITS was shipping only an eighth as much memory. Then, at the last minute, Marsh changed the design, doubling the capacity to 4K. MITS’s first real competition came from those 4K memory boards, which definitely cut into MITS’s profits. Ed Roberts wasn’t pleased.
But MITS’s defective memory boards and delivery backlog had already kicked the door open for some real competition. Bruce Seals, a Tennessee hobbyist, flew to Albuquerque in July to discuss an East Coast dealership, and returned to Tennessee with the entire state as his territory and a promise of three-day delivery. When MITS couldn’t move the products fast enough, especially the memory boards, Seals saw the same need—and opportunity—that Marsh had, and he, too, designed and began to sell a 4K memory board. An industry of sorts was developing.
Processor Technology continued to market memory boards while moving on to new designs. The VDM, or video display module, Felsenstein’s next contract for Proc Tech, was an interface board that allowed the Altair to display output on a television screen. Chuck Grant and Mark Greenberg, who had left UC Berkeley with George Morrow and were now doing business as G & G Systems, did the software for the module, and Steve Dompier wrote Target, a video game that showed off the VDM. Dompier later asserted that it was the VDM that made video games possible.
In the fall of 1975, a local computer show took place at UC Berkeley’s Lawrence Hall of Science, where the East Bay spin-off Homebrew Computer Club first got together. MITS was represented by two area Altair dealers, Paul Terrell and Boyd Wilson, who proudly showed Felsenstein and Marsh the hoops their machine could jump through. Marsh was more impressed by the fact that the Altair was filled with Processor Technology memory boards. Harry Garland and Roger Melen were also present, demonstrating how their Cyclops camera could be used with the Altair.
Sharing Ideas
Before the original Homebrew Computer Club had grown large enough to need the auditorium at the Stanford Linear Accelerator Center (SLAC), Popular Electronics technical editor Les Solomon visited the club at the nearby SLAC Orange Room. He was the star of the evening, telling somewhat far-fetched stories of his own experiences. Sometimes he sounded like a counterspy, other times like a vaudeville magician. “It was unclear which country he was working for,” joked Lee Felsenstein, who was among Solomon’s admirers. At one point, Solomon led the Homebrewers outside, did some hocus-pocus, and instructed them to lift the huge stone table in the yard. They were surprised to find that they could hoist it right up, although Felsenstein noted dryly that the group hadn’t tried lifting it without the hocus-pocus.
Some nights at Homebrew, a tall, dapper, charismatic man could be found at the back of the room selling books out of a cardboard box. He was Adam Osborne, a chemical engineer born in Bangkok of British parents, and the same Adam Osborne who had been doing technical writing for Intel. He had since self-published a book called An Introduction to Microcomputers. It was, in fact, an introduction to microprocessors, such as the Intel 8080. In the early days, microprocessors were commonly referred to as microcomputers, especially by the public-relations departments of semiconductor companies.
Although the people from IMSAI, the leading microcomputer company, almost never attended club meetings, IMSAI cofounder Bruce Van Natta was at Homebrew one night when Osborne was hawking his books, and bought a copy. His decision to include a copy of Osborne’s book with every IMSAI allowed Osborne to start a publishing company that would eventually be purchased for millions by McGraw-Hill. Ironically, it would be Osborne who would first announce IMSAI’s demise in a column in a computer magazine.
After Homebrew meetings, its most fanatical members went to a Menlo Park beer-and-burger place named the Oasis but known to the savvy as “the O.” They sat in wooden booths surrounded by the deeply carved initials of generations of Stanford students, drank beer, and argued computer design. They ignored the fact that they were competitors. There were a lot of things to learn in developing this new kind of product, and they weren’t about to let economic issues get in the way. Marsh and Melen regularly traded insights on design, and Grant and Greenberg sometimes joined them at the O.
Wildfire grep -nH -e Beyond the Valley
By the end of 1975, new microcomputer companies were poking up everywhere, with the most furious activity still in the San Francisco Bay Area. IMSAI was located in San Leandro. Bay Area--based Cromemco was designing boards for the Altair. MOS Technology had released its KIM-1 hobby computer, based on its bargain-basement 6502 microprocessor, equipped with a hexadecimal keyboard in place of binary switches. Microcomputer Associates in Los Altos had its Jolt, a 6502 kit.
Southern California was also a center of growing hobbyist activity. In Gardena, Dennis Brown was selling his Wave Mate Jupiter II, a computer based on the Motorola 6800 microprocessor and designed to attract “serious hobbyists,” for less than $1,000. Although the Altair had sold for less than half that, a realistic price for an assembled Altair system, including some sort of I/O device, adequate memory, and a storage device, was well over $1,000. In San Diego, Electronics Products announced another 6800-based computer, the Micro 68.
On December 31, 1975, Rich Peterson, Brian Wilcox, and John Stephensen quit their jobs to form their own company. Peterson and Wilcox had built an Altair, Stephensen had built his own 8080 machine from scratch, and they found themselves designing boards to make the Altair run better. Deciding that their hobby could just as well be their vocation, they formed PolyMorphic Systems and started working on a computer kit. They first called it the Micro-Altair, and later, under duress, changed the name to Poly 88.
Elsewhere in the West, MITS in Albuquerque was offering a 4K static memory board for its 8080 system and was developing a computer based on what was emerging as the “Southwest chip”: Motorola’s 6800. Systems Research in Salt Lake City sold a 6800 microcomputer board. Mike Wise’s Sphere, operating out of a small factory near Salt Lake City, was offering its 6800 computer with a built-in terminal and plastic case. Southwest Technical Products, run by Dan Meyer in San Antonio, also offered a 6800 system. The Digital Group in Denver was selling a variety of boards.
In the Midwest, Martin Research was offering its Mike CPU boards with 8008 or 8080 chips. Ohio Scientific Instruments in Hudson, Ohio, had 6800 and 6502 kits. Heathkit in Benton Harbor, Michigan, had a computer in the works.
In the East, the hobbyist movement grew up around the Amateur Computer Group of New Jersey. Scelbi, in Milford, Connecticut, put out a popular kit based on the 8008, and Technical Design Labs in Trenton, New Jersey, was developing a computer kit around a new chip, the Zilog Z80. Hal Chamberlin in North Carolina, Bruce Seals in Tennessee, and Georgia Tech student Ron Roberts were active hobbyists working on systems, components, or software.
But the fire burned most strongly in Silicon Valley, with its atmosphere of symbiotic information sharing. New companies that created circuit boards for the Altair popped up almost daily. By the end of 1975, one of these, Processor Technology, was on its way to parlaying its substitute for the defective Altair memory board into financial wealth and, within a curiously anticorporate industry, a kind of corporate respectability.