Hackers

by Steven Levy

  But someone would have to, because the Altair was the basis for a fantastic system to build new systems, new worlds. Just as the PDP-1, or the PDP-6, had arrived at MIT as a magic box without a satisfactory operating system, and just as the MIT hackers had supplied it with assemblers, debuggers, and all sorts of hardware and software tools to make it useful in creating new systems and even some applications, it was up to these as yet unorganized hardware hackers to make their own mark on the Altair 8800.

  Bob Marsh understood that this was the beginning of a new era, and a terrific opportunity. Sitting on the cold floor in Gordon French’s garage, he decided that he would design and build some circuit boards that would plug into one of the blank slots on the Altair bus.

  Bob Marsh wasn’t the only one with that idea. In fact, right there in Palo Alto (the town next to Menlo Park, where the meeting was being held), two Stanford professors named Harry Garland and Roger Melen were already working on add-on boards to the Altair. They hadn’t heard about the meeting, but would come to the second meeting of hardware enthusiasts, and be regulars thereafter.

  The two Ph.D.s had first heard about the Altair when Melen, a tall, heavy man whose wittiness was only slightly impeded by a recurrent stutter, was visiting Les Solomon in late 1974 at the New York office of Popular Electronics. Melen and Garland had done articles outlining hobbyist projects for the magazine in their spare time, and were just putting to bed an article telling how to build a TV camera control device.

  Melen noticed a strange box on Solomon’s desk and asked what it was. Solomon informed him that the box, the prototype Altair that Ed Roberts had sent to replace the one lost in air freight, was an 8080 microcomputer that sold for under four hundred dollars. Roger Melen did not think that such a thing was possible, and Les Solomon told him that if he doubted it, he should call Ed Roberts in Albuquerque. Melen did this without hesitation, and arranged to make a stopover on his way back West. He wanted to buy two of those computers. Also, Ed Roberts had previously licensed a project that Melen and Garland had written about in Popular Electronics and had never gotten around to paying them royalties. So there were two things that Melen wanted to talk to Roberts about.

  The Altair computer was the more important by far—the right toy at the right time, Melen thought—and he was so excited about the prospect of owning one that he couldn’t sleep that night. When he finally got to MITS’ modest headquarters, he was disappointed to find that there was no Altair ready to take home. But Ed Roberts was a fascinating fellow, a dyed-in-the-wool engineer with a blazing vision. They talked until five in the morning about the technical aspects of this vision. This was before the Popular Electronics article was out, though, and Roberts was concerned at what the response might be. He figured it would not hurt to have some people manufacturing boards to put into the Altair to make it useful, and he agreed to send Melen and Garland an early prototype, so they could make something to connect a TV camera to the machine, and then a board to output a video image as well.

  So Garland and Melen were in business, naming their company Cromemco, in honor of the Stanford dorm they’d once lived in, Crowthers Memorial. They were delighted to find similar spirits at the Homebrew Club, among them Marsh, who had talked his friend Gary Ingram into helping start a company called Processor Technology.

  Marsh knew that the biggest immediate need of an Altair owner was a memory bigger than the lousy 256 bytes that came with the machine, so he figured he’d make a board which would give 2K of memory. (Each “K” equals 1,024 bytes.) MITS had announced its own memory boards, and had delivered some to customers. They were nice memory boards, but they didn’t work. Marsh borrowed the PCC’s Altair and looked it over carefully, read the manual backward and forward. This was a necessity because he couldn’t initially afford to spend the money to make a Xerox copy. He figured that he would run the company the way Roberts was apparently running MITS—announce his product first, then collect the money required to design and manufacture the product.

  So on April Fools’ Day, Marsh and Ingram, a reclusive engineer who didn’t go to Homebrew meetings (“It’s not the kind of thing he did,” Marsh later explained), officially inaugurated the company. Marsh was able to scrape up enough money to Xerox fifty fliers explaining the line of proposed products. On April 2, Marsh stood up at the third Homebrew meeting, handed out the fliers, and announced a twenty percent discount to anyone who ordered in advance. After a week, he hadn’t heard anything. As Marsh later said, “Despair had set in. We felt, we’ve blown it, it’s not going to work. Then our first order came in, for a ROM [memory] board costing only forty-five dollars. A purchase order asking ‘Net 30 terms,’ from this company called Cromemco. We thought, ‘Who is this Cromemco? And why don’t they pay cash?’ Despair set in once more. IT’S NOT GOING TO FLY! The next day three orders came in, and within a week after that we had twenty-five hundred dollars cash. We took a thousand, ponied up for a sixth-page ad in Popular Electronics, and all hell broke loose after that. It took us only two months to get a hundred thousand dollars in orders.”

  The irony was that Marsh and the other hacker-run operations were not setting up to be huge businesses. They were looking for a way to finance their avocation of playing with electronics, of exploring this new realm of little bitty computers. For Marsh and the others who left the first few Homebrew meetings with board-building fervor, the fun was beginning: designing and building stuff, expressing themselves by the twists and tangles of a digital logic integrated circuit board to be attached to Ed Roberts’ byzantine bus.

  As Marsh found out, building a board for the Altair was the Homebrew hacker’s equivalent of attempting a great novel. It would be something that harsh Homebrew reviewers would examine carefully, and they would not only note whether it worked or not but judge the relative beauty and stability of its architecture. The layout of circuits on the board was a window into the designer’s personality, and even superficial details like the quality of the holes by which one mounted the board would betray the designer’s motives, philosophy, and commitment to elegance. Digital designs, like computer programs, “are the best pictures of minds you can get,” Lee Felsenstein once said. “There are things I can tell about people from hardware designs I see. You can look at something and say, ‘Jesus Christ, this guy designs like an earthworm—goes from one place through to the end and doesn’t even know what it was he did in the middle.’”

  Bob Marsh wanted Processor Technology to be known for quality products, and he spent the next few months in a frazzled state, trying not only to finish his projects, but to do them right. It was important for the company and for his pride as well.

  The process was not a terribly simple one. After figuring out what your board would do, you would spend long nights designing the layout. Looking in the manual that described the workings of the 8080 chip, you would jot down the numbers for the various sections you wanted—designating this section for an input, that one for memory—and the labyrinthine grid inside that piece of black plastic would begin to reshape inside your head. The effectiveness of your choice of which sections to access would depend on how well and how accurately you kept that vision up there. You would make a pencil drawing of those connections, with the stuff destined to go on one side of the board written in blue, stuff for the other side in red. Then you would get sheets of Mylar, lay them on a grid on a light table, and begin laying out the outline of the connections, using crepe paper tape. You might find out that your scheme had some problems—too much traffic in one part, the interconnections too tight—and have to realign some things. One mistake could blow everything. So you’d be sure to do an overlay of the schematic: placing that on top of your taped-up design, you could see if you made some grievous error, like hooking three things together. If the schematic itself was in error, forget it.

  You would design it so that the board would have several layers; a different set of connections on the top and the bottom. You would flip the layout back and forth as
you worked, and sometimes the tape would peel off, or you would have little pieces of tape left over, or a hair would get stuck somewhere: any of these uncalled-for phenomena would be faithfully duplicated in the sepia reproductions made for you at a blueline house (if you didn’t have money for that, you’d do a careful Xerox), and result in a disastrous short circuit. Then you’d mark up the layout for the board company, telling where to drill and what needed gold-plating, and so on.

  Finally, you’d go to a local board house with drawings in hand. You’d give it to them. Since it was still a recession, they would be happy for the business, even business coming from a scruffy, small-time, glassy-eyed hardware hacker. They would put the thing on a digitizer, drill the holes, and produce on greenish epoxy material a mess of silvery interconnections. That was the deluxe method—Bob Marsh at first could not afford that, so he hand-etched the board over the kitchen stove, using printed circuit laminate material, making barely discernible lines that the material would melt into. That method was a tortuous courting of the bitch goddess Disaster, but Marsh was a compulsively careful worker. He later explained, “I really get into it. I become one with my schematic design.”

  For this first memory board, Marsh was under particular pressure. Every other week at the Homebrew meeting, every day on the phone, frantic people were gasping for static memory boards like divers gasping for air. Marsh later recalled their cries: “Where’s my board? I need it. I GOTTA HAVE IT.”

  Finally Marsh was done. There wasn’t time for a prototype. He had his board, which was the green epoxy rectangle with a little protrusion of etched gold connectors underneath, sized to fit into a slot in the Altair bus. He had the chips and wires which the kit builders would solder onto it. (Processor Tech would only sell unassembled boards at first.) Marsh had it all ready—and no Altair to test it out on. So despite the fact that it was three in the morning he called that guy Dompier he knew from Homebrew and told him to bring the machine over. Dompier’s Altair was at least as valuable to him as a human infant offspring would be if he weren’t in Bachelor Mode, so he carefully wrapped it up in a little red blanket to bring it over. Dompier had gone by the book in assembling the machine, even wearing a copper bracelet around his wrist when he soldered (to keep static down), and taking care not to touch the fragile 8080 heart of the machine. So he was stunned, after lovingly setting the machine down in Marsh’s workshop, when the hardware veterans Marsh and Ingram began handling chips like a couple of garage mechanics installing a muffler. They’d grab chips with their grubby fingers and throw chips around and pull chips out and stuff them back in. Dompier watched in horror. Finally they had the board all ready, and Ingram flicked the switch on, and Steve Dompier’s precious computer fizzled into unconsciousness. They’d put the board in backward.

  It took a day to fix Dompier’s Altair, but Steve Dompier harbored no anger: in fact, he loaned his machine to Processor Technology for future testing. It was indicative of Homebrew behavior. These were a different breed of hacker than the unapproachable wizards of MIT, but they still held to the Hacker Ethic that sublimated possession and selfishness in favor of the common good, which meant anything that could help people hack more efficiently. Steve Dompier was nervous about his Altair, but he wanted little in the world more than a memory board so he could run some real programs on the machine. And then he wanted I/O devices, display devices . . . so that he could write utilities to make the machine more powerful. Tools to Make Tools, to go deep into the world that centered on the mysterious 8080 microprocessor inside his machine. Bob Marsh and the others in Homebrew, whether they were offering products for sale or were simply curious hackers like himself, were all in this together, and together they formed a community that may not have been as geographically centered as MIT’s PDP-6 community was—it stretched from Sacramento to San Jose—but was strongly bonded nonetheless.

  When Bob Marsh showed up at a Homebrew meeting in early June with the first shipment of boards, the people who ordered them were so thankful you might think that he’d been giving them away. He handed over the little plastic blister-wrapped packets of board and ICs, along with the instruction manual Lee Felsenstein had written. “Unless you are an experienced kit builder,” Lee warned, “don’t build this kit.”

  There was very little experience in the world at building those kinds of things, but much of the experience that did exist in the world was centered in that meeting room, which was now the auditorium at the Stanford Linear Accelerator (SLAC). It was four months after the first casual meeting of the club, and its membership had grown almost tenfold.

  • • • • • • • •

  The little club formed by Fred Moore and Gordon French had grown to something neither could have imagined. It was the vanguard of a breed of hardware hackers who were “bootstrapping” themselves into a new industry—which, they were sure, would be different from any previous industry. The microcomputer industry would be ruled by the Hacker Ethic. (The term “bootstrap” was indicative of the new jargon spoken by these hackers: the term literally describes the process by which a computer program feeds itself into a machine when the machine is first turned on, or “booted.” Part of the program will feed the code into the computer; this code will program the machine to tell itself to feed the rest of the code in. Just like pulling yourself up by your bootstraps. It is symbolic of what the Homebrew people were doing—creating a niche in the world of small computer systems, then digging deeper to make the niche a cavern, a permanent settlement.)

  But the club’s founders were both soon outdistanced by the technical brilliance around them. In French’s case, he suffered from what seemed to be a latent bureaucratic attitude. In some respects, his mania to keep the club progressing in an orderly, controlled manner was helpful. He acted as secretary and librarian, keeping a list of everyone’s phone number and what equipment everyone owned. As he later recalled, “My phone rang off the hook. It was incredible. Everybody needed information, and they needed each other in order to get going because there was an absolute paucity of equipment. For example: ‘If you have a terminal could I borrow it for a couple days while I get my program in it so it’ll read my punch paper tape reader?’ That sort of thing.”

  But in other respects, particularly in the way he moderated the meetings, French’s style was not consistent with the hacker spirit brewing in Homebrew. “Gordon was a didactic sort,” Lee Felsenstein would later recall. “He would try to push the discussion to where he wanted it to go. He wanted it to be an educational event, holding lectures, teaching people about certain things, especially stuff he was expert on. He was very upset if the discussion strayed from people literally teaching other people in a schoolish sense. He would jump into whatever people were saying and get involved in the content, injecting his opinions and telling them ‘There’s an important point that shouldn’t be missed, and I know more about this kind of stuff.’” After the first part of the meeting, in which people would introduce themselves and say what they were working on, Gordon would stand up in front of the room and give what amounted to a tutorial, explaining the way the machine uses the code you feed into it, and informing the restless members how learning good coding habits will save you headaches in the future . . . and sooner or later people would get so impatient they’d slip out of the meetings and start exchanging information in the hall. It was a touchy situation, the kind of complex human dilemma that hackers don’t generally like to confront. But the feeling emerged that a new moderator should take over.

  The logical choice might have been Fred Moore, who sat in the front of the room for the first few months of Homebrew with his tape recorder and notebook, capturing the meeting so he could summarize highlights in the newsletter he put out every month. He was putting a lot of his time into the group, because he saw that the hackers and their Altairs were on the verge of what might be a significant social force. “By sharing our experience and exchanging tips we advance the state-of-the-art and make low-cost computing possi
ble for more folks,” he wrote in the newsletter, adding his social commentary: “The evidence is overwhelming that people want computers, probably for self-entertainment and education usage. Why did the big companies miss this market? They were busy selling overpriced machines to each other (and the government and the military). They don’t want to sell directly to the public. I’m all in favor of the splash MITS is having with the Altair because it will do three things: (1) force the awakening of other companies to the demand for low-cost computers in the home . . . (2) cause local computer clubs and hobby groups to form to fill the technical knowledge vacuum, (3) help demystify computers . . .”

  Moore explicitly identified the purpose of the club as an information exchange. Like the unfettered flow of bits in an elegantly designed computer, information should pass freely among the participants in Homebrew. “More than any other individual, Fred Moore knew what sharing was all about,” Gordon French later recalled. “That was one of the expressions he was always using—sharing, sharing, sharing.”

  But the majority of the club preferred a path that diverged from Fred Moore’s. Fred was always harping on applications. Every so often in the early meetings he would urge the members of this basically anarchistic group to get together and do something, though he was usually vague on what that something might be.

  Maybe using computers to aid handicapped people, maybe compiling mailing lists for draft resistance. Moore might have been correct in perceiving that the thrust of the club was in some way political, but his view seemed at odds with the reality that hackers do not generally set about to create social change—hackers act like hackers. Moore was less fascinated with the workings of computer systems than with the idea of bringing about a sharing, benevolent social system; he seemed to regard Homebrew not as a technical stronghold of people hungry for the pyramid-building power of in-home computers, but as a cadre devoted to social change, like the draft resistance or antinuke groups he’d been involved in. He would suggest cake sales to raise funds for the group, or publish cute little poems in the newsletter like “Don’t complain or fuss / It is up to each of us / To make the Club do / What we want it to.” Meanwhile, most of the club members would be turning to the back of the newsletter to study the schematics in the contribution called “Arbitrary Logic Function Generation Via Digital Multiplexers.” That was the way to change the world, and a lot more fun than a cake sale.