Tools For Thought

by Howard Rheingold

April, 2000: a revised edition of Tools for Thought is available from MIT Press including a revised chapter with 1999 interviews of Doug Engelbart, Bob Taylor, Alan Kay, Brenda Laurel, and Avron Barr. ISBN: 9780262681155, 1985 ed, ISBN: 0262681153

The idea that people could use computers to amplify thought and communication, as tools for intellectual work and social activity, was not an invention of the mainstream computer industry or orthodox computer science, nor even homebrew computerists; their work was rooted in older, equally eccentric, equally visionary, work. You can't really guess where mind-amplifying technology is going unless you understand where it came from.

- HLR

Chapter Eight:
Witness to Software History: The Mascot of Project MAC

When he tried the doorknob and found it unlocked, then opened the door to Building 26 and poked his head into a room full of weirdos having a high old time with candy bars and computer programs, David Rodman knew he had discovered something. The year was 1960. David Rodman was ten years old. And 1960 was still at least four years too early for weird people to be anything but a rarity, even on college campuses.

It turned out that these pasty-faced, hollow-eyed, jargon-spewing, insanely cackling young men were the first, founding generation of dropout programming wizards to call themselves "hackers," and Building 26 was where the hotshot hired programmers of MIT's artificial intelligence Project MAC were caged until they all moved to the ninth floor of 545 Technology Square, in the early sixties.

Technology Square was MIT's space-age temple of sci-tech. The geographical move from outpost to the pinnacle of the technohierarchy reflected an elevation in the importance of the whole field of man-machine systems. MAC was set up originally by Licklider, later administered at various times by Fano, Minsky, and Papert, and the ambiguity about the meaning of the acronym was deliberate. On the level of the hackers' employers, it meant both "machine-aided cognition" and "Multi-access computing," because in the early 1960s computer system design and AI research had not yet parted ways.

Down in Building 26, where the dirty work went on, where this motley group of exceptionally gifted programmers got their fingers into the logical guts of machines and made them do their bidding, they were Maniacs And Clowns, Men Against Computers, and numerous unprintable variations. They were the unruly but indispensable hired craftsmen of the projects directed by the likes of McCarthy and Minsky and funded by Licklider — the ones who built the software probes their employers launched into the frontiers of machine intelligence.

At the moment David walked in, a young man named Richard Greenblatt, who lived on the stereotypical hacker diet of soft drinks, candy bars, and Rolaids, and who didn't stop to sleep, much less to wash or change clothing, was explaining to a circle of awed admirers, which included some of the computer scientists who had hired him, how he intended to write a chess playing program good enough to beat a human. Greenblatt's thesis advisor, Marvin Minsky, tried to discourage Greenblatt, telling him there was little hope of making progress in chess playing software.

Six years after he first stumbled upon the inhabitants of building 26, sixteen-year-old David Rodman, by now a dropout, acidhead, and professional AI programmer of his own, albeit smaller, repute, was in the group that watched Greenblatt's "MacHack" program demolish Hubert Dreyfus, the number one critic of the whole AI field, in a much-heralded and highly symbolic game of chess. The MacHack versus Dreyfus duel has become one of the hacker legends, and MacHack became the first program to be granted honorary membership in the American Chess Federation.

The Dreyfus chess match was only one of several historic moments in AI history that David witnessed from his vantage point of mascot, then apprentice, then full-fledged hired hacker, during the heyday of MAC, between 1960 and 1967. He was there when his motley colleagues began to build the programming and operating systems for the TX-0 and PDP-1 computer hardware, thus establishing the first software thrust into the age of interactive computing. David was also there when Joseph Weizenbaum, to his later regret unveiled ELIZA, probably the most widely quoted and widely misunderstood program in history — the program that seems to be an uncannily perceptive psychiatrist, but is actually a programmer's semantic trick.

David came upon the hackers through a mixture of mischief and happenstance. He was one of those prodigies who was angry about having a brain like his trapped for another eight years in the body of a child. Since he was six, he had been an exceptional musician, but he gave up the piano at ten because he despised performing for adults. He was a loner, a wanderer, a looker through doorways, an urban spelunker — a snoop, but not a thief, unless you consider knowledge of how to find your way through a complicated system as a stealable property. By the age of fifteen, David and his friends could find their way into any building in the MIT complex, via the system of underground utility tunnels.

Wandering through the halls of MIT, where his father worked in the medical school, was one of his favorite pastimes. He liked to try doors and see what was behind the unlocked ones. When he cast his eyes on those strange guys gathered around an odd-looking television set with wires coming out of it, and then joined them at a game called "Spacewar," using a control panel made out of a cigar box, and nobody seemed to notice that he was ten years old — David knew he had found his new intellectual home.

"They treated me with some subtlety. I think it was a kind of recognition. They had all been through it, but they weren't about to tell me anything before I figured it out for myself," David recalled, twenty years later. He just sat down and there was a keyboard and someone got him started, and although they were the first people he had met who didn't make a fuss over his intelligence, they noticed how quick he picked it up, all right.

After David returned a few times, and demonstrated his ability to find his way around the computer, the hackers made him a mascot, and when he was a full-blooded initiate ("when they started calling me 'Rodman' instead of 'hey, kid'"), they started giving him small tasks in machine language, eventually showing him tricks in the sexy new programming language known as LISP invented specifically for AI programmers by John McCarthy, one of the project's founders.

Marvin Minsky's secretary took a liking to this wiseass ten-year-old who seemed to take to programming as some kids take to chess or tennis or ballet, and Minsky, who had always been the hackers' patron in MIT computer circles, let David use his password.

Today, having grown up through the early days of the hackers and AI research, the ARPAnet years, the consulting contracts and security clearances, the regular escalation of his income, and the transformation of the social status of computer programmers from weirdo outsiders to millionaire culture heroes, David Rodman is the president of a microcomputer software company whose primary product is a system of programs he wrote himself. His personal odyssey from the inner sanctums of AI hackdom to the rough-and-tumble capitalism of the microcomputer industry is a kind of capsule history of the whole strange journey of interactive computing from laboratory curiosity to home appliance.

But like many others who are now in their middle thirties and who didn't wear suits and carry briefcases, the early history was colorful and not a little painful: "At the age of ten, I was like a coiled spring inside — lonely, uptight, angry, cynical. I was unable to balance my intelligence against the rest of the world. Then suddenly, here were people not unlike myself, who showed me a device that would respond to me when I sat down to program it. Those people knew what was happening to me, and when I began programming, they encouraged it."

MIT, to begin with, was the engineers' school of engineers' schools, where the undergraduates hold an annual "ugliest man on campus" contest — an unashamed, self-proclaimed, national haven for supernerds. The campus population was primarily composed of the people from all the high schools in the country who stayed home and learned integral calculus or built ham radios while everybody else was at the sock hop. Amid all this self-styled rejection of conventional youth culture and the atmosphere of cultivated unfashionability, computer obsessives were considered oddballs even by the other outcasts. Their standards were entirely their own. They and their computers, and a few people in ARPA, were the only ones who knew that the top hackers were really the insiders. Although they were outcasts from the wider society, from their fellow techies, and even from most other computer scientists, they happened to be the people who were creating the future of computing — the first time-sharing systems.

They were having so much fun with what they all knew to be the hot technology of the future that they seemed to deliberately encourage their unappetizing image. You don't just barge in and make yourself a hacker. You've gotta hack. And that means making a computer do things its manufacturers never expected it to do. (This kind of programming is known among hackers as "black magic.") It also meant surviving what the other hackers could do to the results of all your work if you weren't clever enough to prevent them.

There was a matter of intellectual style. Boldness and speed and raw power were as important as (critics of hackers would say more important than) elegance and efficiency when it came to "cutting code" (writing the detailed machine language or high-level language lists of instructions that make programs do what computer users want them to do.). One common comeback when an outsider asked what "hacker" meant was "somebody who makes furniture with an axe." Orthodox programming style was hardly de rigeur in this crowd. The challenge was to think of a clever way to do something that most normal computer experts would do some other way or not at all. The performance standards were idiosyncratic and subtle, but all-important. These people judged each other by criteria that the rest of the world didn't even understand, and the hackers didn't mind keeping it that way.

They were other kinds of outcasts besides social outcasts, self-selected or otherwise. Their values were entirely their own: academic or commercial success was too trivial to be considered a driving motivation; the opportunity to work with like-minded colleagues on state-of-the-art equipment was paramount. They had their own culture, their own ethic, even their own dialect. The eighteen-year-old MIT dropouts David Rodman wanted to emulate were distinguished from the hippies and radicals they superficially resembled because they all happened to have a talent that was particularly valued in those days, and still is — the ability to write code that makes computers useful to nonprogrammers.

While all their former classmates were on to their doctorates and assistant professorships and corporate research laboratories, the misfits suddenly found their conventionally successful peers, at a job where they weren't relegated to working out a payroll system or an airline reservation service. The hackers knew, even if nobody else did, that they — and not IBM, or even their straighter "FORTRAN type" colleagues in computer science — were the test pilots of the computational frontier, pushing the limits of what could be done with each fresh generation of hardware.

Their mandate was to dream up new things for computers to do, and in the process what they did was invent a whole new computer system and computer-oriented society, a technology and social order in which their own little fraternity of ex-outsiders, and not the conventional computer types, were privileged to know the inner mysteries. When the rest of the world caught up with them, they knew they would be on to something even more mysterious to the outsider and more exciting to the hacker. None of them would deny the charges of addiction. Some of the same people who were in that room when David walked in, almost a quarter of a century ago, are still sitting in front of a computer terminal, somewhere on the upper floors of 545 Technology Square.

Spacewar

Their superiors were smart enough to know that the best of the hackers would come up with amazing things if they were left to their own devices. Spacewar, which spread from MIT to other campus computer centers, was one of the rites of passage and defining characteristics of any den of hackers. It was invented by a MAC hacker named Russell, known as "Slug", and was perfected in a communal effort by generations, and it survived wherever it sprouted, like some antibiotic resistant organism, because every computer laboratory manager in the country learned that productivity dove when Spacewar was banned and shot back up when the game was reinstated.

It was Spacewar that influenced Nolan Bushnell to create, over a decade later, a much simpler version called Pong, a commercial venture that created the first incarnation of Atari Corporation and a billion-dollar video game industry. Before Pong succeeded, however, Bushnell had failed to get people interested in a more complex game, a more direct derivation of Spacewar. But in those days, the people who put quarters into video machines at bars and arcades hadn't yet been educated in their video game sensibilities by the Space Invaders and Pac-Man phenomena of the late 1970s and early 1980s.

But fun and games were only part of the fun and games. One of the things the hackers were building when David arrived was the software for one of the first time-sharing systems. They were writing a time-sharing operating system that they intended to use to create the greatest hacks, the biggest pranks, the most amazing demonstrations of programming virtuosity in hacker history. The fact that they were pioneering a whole new way to use computers that would eventually bring the outside population in on it was not the first thing on their mind. They wanted to get their own hands on the system, so they built it in record time.

Actually, there were two MIT time-sharing projects. The more staid project was CTSS — Compatible Time-Sharing System, so named because it was designed to be compatible with other systems that were being constructed elsewhere. The MAC hackers were designing an operating system they called ITS — the Incompatible Time-sharing System. They couldn't care less about making it easy for outsiders to use. They were having too much fun to share it with the kind of straight-arrow programmers who could stand to eat or sleep before finishing a good hack.

There were hackers and there were metahackers. Richard Greenblatt, because of what his program did to Dreyfus, and because of his ability to improvise great code without fully understanding how he did it, was at the top of the hacking order. He was a dropout and looked the part of the "Pepsi-guzzling, nonsleeping, single-minded programming addict who ate only food that came from a vending machine and whose skin had not absorbed anything but fluorescent light in three years," as Rodman fondly remembered him, three decades later. But Greenblatt's peers knew him as a Nijinsky, a Frank Lloyd Wright, a Johann Sebastian Bach of LISP programming.

The matter of pranks, of what the hackers called "wheel wars" — mucking up each others files, trying to thwart each other or "crash" the operating system — was part of the working environment. Crashing the system could be accomplished by running some kind of unrunnable self-swallowing program that the programmer who designed the system hadn't made precautions for. When such a prank succeeds, everybody connected to the system can lose important data. In the early sixties, at places like MAC, it was understood that, despite its unfortunate side effects, crashing was an allowable test of the system if the hack revealed an important system vulnerability.

Two decades later, when mischievous and sometimes vandalistic teenagers with home computers started calling themselves "hackers" and crashed the files of nonhackers via the telephone, they were doing something quite different in its ultimate effect, if not in its outward appearance, from what the first such outlaws at MAC were trying to accomplish. The excuse was that they were "just exploring" an interesting vulnerability in the system had some real validity back when the hackers were creating and testing new time-sharing systems, and when their expertise was aimed toward a common goal. But when the system that crashes, as nearly happened in 1983, is an operational computer used by a hospital to keep track of patient medication records, it is a somewhat different matter. The same kind of iconoclastic mischief that had one meaning in the 1960s took on another meaning in the 1980s.

"Phone-hacking" was another kind of prank pioneered by MAC hackers in the early 1960s that was to spawn anarchic variants in the 1970s. The self-taught mastery of complex technologies is the hallmark of the hacker's obsession, the conviction that all information (and information delivery technologies) ought to be free is a central tenet of the hacker ethical code, and the global telephone network is a complex technological system par excellence, a kind of ad hoc worldwide computer. The fact that a tone generator and a knowledge of switching circuits could provide access to long-distance lines, free of charge, led to a number of legendary phone hacks. But the mythology didn't die there.

In California, the Stanford AI Laboratory (SAIL) and the proximity to Silicon Valley led to the growth of another phone-hacking subcult of "phone Phreaks" in the 1970s, whose hero was a fellow who went by the name of Captain Crunch. A gap-toothed, crazy-eyed, full-bearded fellow who now writes software and stays away from illegal activities, Crunch traveled the highways in the late sixties and early seventies with a van full of electronic equipment, playing virtuoso pranks from roadside phone booths — until he was caught, prosecuted, sentenced, and jailed. One of Crunch's phone hacking buddies from the outlaw days, Steve Wozniak, went on to bigger fame when he invented the first Apple computer. Captain Crunch, also known as John Draper, now makes very decent legitimate money as "Cap'n Software," the sole programmer for the microcomputer software company of the same name.

At Project MAC, and at the subcultural counterparts at Stanford (where they began to blend some of their California brand of craziness into the hacker formula) and elsewhere, you had to suffer in order to be admitted to the more interesting levels of hacker wisdom. As in any closed subculture, the hackers spared no one their own kind of rites of passage. David was the youngest initiate, but they didn't go any easier on him than any other newcomer. You just weren't part of things at MAC until you met the now-legendary "cookie monster" and some of its nastier relatives.

Crashing the system was a fact of life and an implicit challenge at the higher reaches of hackdom — if you were smart enough to come up with something that the system programmers hadn't guarded against, it was more of an honor than a misdemeanor to bring the computer to a halt, dumping hours or weeks of someone's work. By comparison, the cookie monster was relatively mild. Unlike an operating system crash, the cookie monster struck only selected victims, rather than everybody who was unfortunate enough to be using the system when a crash was perpetrated.

The cookie monster would strike most often at four in the morning. (All-night hacking began with time-sharing systems, not only because it fit in with the hacker's weird self-image, but because time-shared systems run faster at night, when all the nonhackers are out having dates or studying poetry or sleeping or whatever nonhackers do at night in the real world.) You would be looking for a bug somewhere in the two-thousandth line of your program. Suddenly, without warning, the words "I WANT A COOKIE!!" appear on your monitor screen — and all your painstakingly crafted code is relentlessly munched into oblivion by the word COOKIE!!, multiplied over and over until you finally figure out or (horror of horrors) somebody has to tell you: you have to type the word COOKIE!! on your keyboard.

In their own way, the MAC hackers were the forerunners of other kinds of psychic desperadoes who appeared on college campuses in the 1960s. A contempt for middle-class values and an abiding interest in the workings of their own mind were two characteristics that hackers were to share with later subcultures who had nothing to do with computers. David Rodman was a confirmed hacker in the late 1960s, when he began to dabble in a very different yet strangely similar outlaw subculture that was springing up in the Cambridge student community.

"I would characterize my first acid trip as a quantum leap into the innards of my own psychology," David recalls today. "Suddenly, there I was — inside myself. I didn't know the path to get in, but there I was. I could observe myself playing the guitar or writing code, and think to myself while improvising. 'Where am I going and how do I know how to go there and what am I really expressing?' It was the trip of all trips."

David thinks that "for my peculiar cognitive style, programming was a perfect preparation for psychedelics, because it allowed me to model a little piece of my personality in the machine, and interact with it. The older hackers would tell me 'never mind what the main program does, we want you to write a program that moves a chess piece on a chessboard,' so I wrote a small, gemlike part of the utility package that went into one of the chess programs. The next time I found myself in one of those gemlike structures on my first acid trip."

The small "gemlike structures" that David created were incorporated into early versions of Greenblatt's MacHack, the program that eventually became an emblem of the hackers' sovereignty within the AI community when MacHack met Dreyfus in 1967. It all started when Hubert Dreyfus had the temerity to question not only the chances of success but the very legitimacy of AI research. The entire field of artificial intelligence had been challenged as a fraud, and very serious efforts that went beyond the usual acrimony of academic debate were being made to cut off funding for the foolishness Minsky et al. were attempting. The Dreyfus affair began in the summer of 1965, when Hubert Dreyfus — a philosopher, not a computer scientist — spent a few months at the Rand Corporation. The paper that Dreyfus wrote at the end of that summer, entitled "Alchemy and Artificial Intelligence," was informally circulated as a Rand report.

Dreyfus thought that AI was a crock. He specifically attacked some of the claims AI enthusiasts had made about the future of their field. He claimed that the "progress" the AI folks had been citing was an illusion, and attempted to prove that their goal was a delusion. An IBM researcher, Arthur Samuels, had recently created a pretty decent checkers program that was on its way to becoming a champion. To Dreyfus, saying that the checkers program represented a step toward a true human-like machine intelligence was like saying that an ape who could climb to the top of a tree was making progress toward flying to the moon.

Dreyfus challenged the idea that a chess playing program of any significance could ever be built, pointed out that in 1957 Herbert Simon had predicted an unbeatable chess playing program within ten years, and noted that the time was about up. Greenblatt came out of nowhere with his carefully constructed chess hack, and Seymour Papert, then codirector of MAC, maneuvered Dreyfus into a public match.

David and other witnesses remember the game as a dramatic and unpredictable match — a cliff-hanger that was far more suspenseful and ingenious and less mechanical than what any of them had expected. This was more than a friendly rivalry. The source of their funds was being attacked, and it was just possible that this . . . this . . . philosopher might manage to get people so stirred up that they would take their precious terminals away. It was a grudge match, no question about it.

MacHack won. Gleefully, the bulletin of the Special Interest Group in Artificial intelligence (SIGART) of the Association for Computing Machinery reported the results of the match under a headline taken from Dreyfus' paper: "A Ten-Year-Old Can Beat the Machine — Dreyfus." The SIGART editors amended it with a subhead of their own: "But the Machine Can Beat Dreyfus." The SIGART article touched off a series of letters to editors, accusations, and counteraccusations, and Dreyfus ended up writing a book, What Computers Can't Do in which he admitted: "Embarrassed by my expose of the disparity between their enthusiasm and their results, AI workers finally produced a reasonably competent program. R. Greenblatt's program called MacHack did in fact beat the author, a rank amateur."

MacHack went on to become an honorary member of the U. S. Chess federation, and the Dreyfus-versus-AI controversy has dragged on for decades, albeit without the hand-to-hand fury of 1967, when a hacker rose brilliantly to the defense of his art with a legendary hack, then retreated back to his terminal while others argued the significance of what he had done. The event had more than symbolic significance: the formal paper Greenblatt wrote about the program was of historical value to those who still hope to fulfill Turing's, von Neumann's, and Shannon's dreams of playing against a true master chess-machine

Eliza

MacHack was actually the second of two historic software births David Rodman witnessed during his apprenticeship at MAC. Joseph Weizenbaum showed up at MIT in 1963, and when he created ELIZA between 1964 and 1966, he changed the way everybody thought about what computers can't do — and that included changing his own mind about where the whole computer-AI enterprise was heading. ELIZA was a clever way of mimicking human interaction through a computer-mediated dialogue; what the inventor hadn't anticipated was people's willingness to be taken in by the mimicry — even people who should have known better. By the time Weizenbaum recovered from the shock of seeing the way people reacted to his program, he was convinced that something very dangerous lurked in the much-heralded computer revolution.

The reaction to ELIZA eventually led Weizenbaum to question the ultimate value of the changes that computers were introducing to the general population — changes he felt we might all later regret. He also declared that we would soon be faced with important decisions about what computers ought and ought not to do. He specifically cited the hackers as a symptom of a sickness in the heart of computerdom. Weizenbaum's assault on some of the most fundamental premises of the computer culture with the 1976 publication of Computer Power and Human Reason set off a continuing, oft-heated public debate between Weizenbaum and the AI community.

The Dreyfus-AI debate had been largely a technical argument, which helped make MacHack's technical victory so sweet. Weizenbaum's was a moral argument, and it carried a passionate force far different in effect from that of Hubert Dreyfus, flying in from California with his phenomenology. This was Joseph Weizenbaum, honored professor of computer science at MIT, saying that AI might not be a crock, but we better be a lot more careful with computers, and watch out for the hackers in the process.

Remember when those funny-looking "computer letters" started appearing on the bottom of checks, in the early 1960s? That was part of Joseph Weizenbaum's work in the days before he came to MIT. As a software expert for General Electric, he was centrally involved in Bank of America's ERMA project, a milestone in the computerization of the world's banking system. When Weizenbaum later spoke about the morality of using computers in ways that might change millions of people's lives, he was speaking from experience. His creation of a program that gave the illusion of a wise, all-knowing, computerized psychiatrist — and his shock at seeing how willingly even his computer-sophisticated colleagues were taken in by the illusion — triggered Weizenbaum's dissension.

Weizenbaum started out at MIT with what he thought would be an abstract interest in programs that used simple programming tricks to answer questions posed in English. He put together a working version that the hackers had fun with, and which seemed to be a step, albeit a very primitive first step, toward a genuine language-understanding program. The hackers enjoyed the pretense that they were "conversing" with a computer, even though they knew that the program was parsing very simple sentences with no real underlying understanding of their meaning.

While he was working on a more elaborate question-answering program, with greater sentence-parsing capabilities than his first version, Weizenbaum met a psychiatrist from Stanford named Kenneth Colby, who was interested in using computers to model mental states and mental disorders in order to find clues to more effective forms of psychotherapy. Colby helped Weizenbaum create a special-case version of ELIZA, known as DOCTOR. In the introduction of his 1976 book, Weizenbaum described how this most successful and disturbing version of ELIZA came into being, and conveyed a little of the flavor of the "ELIZA effect":

The work was done in the period 1964-1966, and it was reported in the computer-science literature in January 1966 and August 1967. To summarize it briefly, I composed a computer program with which one could "converse" in English. The human conversationalist partner would type his portion of the conversation on a typewriter connected to a computer, and the computer, under control of my program, would analyze the message that had been transmitted to it, compose a response to it in English, and cause the response to be typed on the computer's typewriter.

I chose the name ELIZA for the language analysis program because, like the Eliza of Pygmalion fame, it could be taught to "speak" increasingly well. . . .

For my first experiment, I gave ELIZA a script designed to permit it to play (I really should say parody) the role of a Rogerian psychotherapist engaged in an initial interview with a patient. The Rogerian psychotherapist is relatively easy to imitate because much of his technique consists of drawing his patient out by reflecting the statements back to him. The following conversation betwen a young lady and ELIZA playing doctor illustrates both the Rogerian technique of encouraging a patient to keep talking and the operation of the computer program ELIZA. The first to "speak" is the young lady. The computer's responses are printed entirely in capitals.

Men are all alike.
IN WHAT WAY
They're always bugging us about something or other.
CAN YOU THINK OF A SPECIFIC EXAMPLE
Well, my boyfriend made me come here.
YOUR BOYFRIEND MADE YOU COME HERE
He says I'm depressed much of the time.
I'M SORRY TO HEAR YOU ARE DEPRESSED

The first thing that alarmed Weizenbaum was the way people who knew enough about computers to know better began to get drawn into serious conversations with it about their lives! Even his own secretary had fallen into the habit of conversing with it! People were asking to be left alone with the machine to spill out their most intimate thoughts! Weizenbaum was so horrified that he started rethinking everything he believed. But that was just a symptom of how gullible we all might be to what he knew was just a clever hack.

More serious, to Weizenbaum's way of thinking, was the fact that Colby and others thought that "computer-therapists" might soon be applied to some form of "automatic psychotherapy" — an idea Weizenbaum considered "obscene." Weizenbaum's ethical debate, although it originated in the same laboratory that spawned so many important innovations in AI and computer systems design, will not be discussed at length here. His books and the ideas expressed by Weizenbaum and his critics deserve consideration on their own accord.

David Rodman was one of those who spent time conversing with ELIZA when it was still in its infancy, while he was employed as a research assistant in the same laboratory. Some of David's earliest LISP hacks were attempts to emulate ELIZA. And although Weizenbaum didn't know it, some of David's early acid trips were spent in "conversation" with ELIZA.

While Minsky was a kind of patron saint of hackdom, and Greenblatt was an unkempt hero, and McCarthy had his own brand of AI prodigies, Weizenbaum was not very fond of some of the hackers who shared his working quarters, to put it mildly. In his book, he mounted a direct assault on the inner circle of hard-core hackers:

Wherever computer centers have become established, that is to say, in countless places in the United States, as well as in virtually all other industrial regions of the world, bright young men of disheveled appearance, often with sunken glowing eyes, can be seen sitting at computer consoles, their arms tensed waiting to fire, their fingers, already poised to strike at the buttons and keys on which their attention seems to be as riveted as a gambler's on the rolling dice. When not so transfixed, they often sit at tables strewn with computer printouts over which they pore like possessed students of a cabalistic text. They work until they drop, twenty, thirty hours at a time. Their food, if they arrange it, is brought to them: coffee, cokes, sandwiches. If possible they sleep on cots near the computer. But only for a few hours — then back to the console or the printouts. Their rumpled clothes, their unwashed and unshaven faces, and their uncombed hair all testify that they are oblivious to their bodies and to the world in which they move. They exist, at least when so engaged, only through and for computers. These are computer bums, compulsive programmers. They are an international phenomenon.

Weizenbaum attacked those particularly obsessed specimens among the hackers he called "compulsive programmers" on several grounds besides their unorthodox appearance and dietary habits. But he also took care to note (parenthetically) that "(It has to be said that not all hackers are pathologically compulsive programmers. Indeed, were it not for the often, in its own terms, highly creative labor of people who proudly claim the title 'hacker,' few of today's sophisticated time-sharing systems, computer language translators, computer graphics systems, etc., would exist.)"

The compulsive programmers, according to Weizenbaum's criteria, spend far more time playing with their computers than using them to solve the problems they are being paid to solve. They are often superb technicians, he admitted, but he also charged that they are very often so sloppy when they document the programs they have written that other programmers, when they later have to use or modify them, are unable to make sense of what they did.

The obsessed hacker's motivation is not problem-solving, but the raw thrill of interacting with the computer, and that, Weizenbaum charged, was a sign, not of prodigy, but of pathology. "The compulsive programmer," he insisted, "is merely the proverbial mad scientist who has been given a theater, the computer, in which he can, and does, play out his fantasies."

Minsky and others rose to the hackers' defense, pointing out that they should be considered with some of the same suspension of normal standards that society reserves for artists. And just as it is true that a hollow-eyed dropout is not a particularly pleasant sight, and perhaps there is truth to the charge that many of them find it easier to relate to the machine than to other people; isn't there also a chance that they are being unfairly maligned?

Hackers would rather be judged by their creations than by their behavior, and nobody cares about van Gogh's habits of dressing, or whether Mozart went without sleep for days at a time. Minsky deplored public stereotyping and scapegoating of people who happen to be passionate about programming instead of violin playing or basketball or making money.

Weizenbaum was undoubtedly right about the temptation to use computers for stimulating fantasies of omnipotence over fantastically controllable worlds. The value to society of obsessively converting sophisticated computers into toys and games has been a matter of extended debate. Nobody would deny that hackers love fantasy. That these fantasies can be fascinating to nonhackers as well has been an inside secret for years, ever since the hack known as "Dwarf Hall of Mists, XYZZY and the Infamous Repository," created by Will Crowther and Don Woods, now more commonly known as "Adventure," surfaced at MAC and SAIL.

After they introduce you to ELIZA, "Adventure" is what hackers show you when you ask them why they are addicted to computing. They hit a few keys, sit you down in front of a monitor and a keyboard, and come back in a few hours to forcibly unplug you. Even in this age of more dazzling computer-generated effects, the sheer temptation to explore the computer-stored fantasy remains strong.

After you are told you can give simple instructions like "drop sword," "go up," "cross bridge," the following words, still famous at every computer center, appear on the screen: "You are standing at the end of a road before a small brick building. Around you is a forest. A small stream flows out of the building and down a gully . . ."

Without warning, and without any high-resolution graphics or sound effects, you are drawn into Colossal Cave, where a labyrinth of chambers containing treasure, dwarfs, magic, strategy and dangers awaits your command. It can take weeks to finish a game. More than one commentator has used "Adventure" as a metaphor for hacking: This is a complex pathway hidden inside the computer, and it is up to the hacker to use all his or her skill, knowledge, and magic to find the treasure and bring it back.

A high regard for programming skill, a mischievous bent, and a predilection for playing games seemed to accompany the spread of the hacker culture, along with Spacewar and Adventure. Weizenbaum might have been the first, but he wasn't the last computer scientist to voice concern over the possible dangerous side effects of this way of thinking.

One famous debate erupted at Stanford, years after Weizenbaum's original diatribe. Stanford has been a West Coast headquarters for hackers since the mid-1960s, although significant outposts have long existed at UC Berkeley, Los Angeles, San Diego, and Santa Barbara, at Stanford Research Institute, and even at Rand before the Ellsberg affair. But LOTS — Stanford's Low Overhead Time-sharing System — is where the undergraduate hackers hang out. It was here that another, more recent major hacker controversy surfaced, in the form of a dialogue on the medium that was known by the mid-1970s as "electronic mail." It was the option of everybody on LOTS to post and read messages, either to specific individuals or groups, or to anyone who was interested, via the "bulletin board" sector of the mail program. People could read and add messages whenever they were logged onto the computer.

Sometimes serious issues were discussed in this manner, and sometimes long impassioned graffiti (known as "flames") were launched against a variety of targets ranging from the profound to the utterly inconsequential. Sometimes serious issues were disguised as flames, and vice versa. Branches and subbranches of such exchanges could continue for months, making up a kind of electronically embedded ad hoc literature. That was where the "hacker papers" came from.

This particular counterpoint of flames on the subject of hackers, written by hackers, came to the attention of the "real world" because a Stanford professor of psychology named Philip Zimbardo discovered the dialogue and published it, with commentary, in Psychology Today magazine in 1980, twenty years after Rodman met Greenblatt et al. in Building 26.

The exchange of flames began with a hacker's version of Luther's 95 theses, nailed, metaphorically, to the door of the electronic temple. A self-sworn ex-hacker who called himself "G. Gandalf" (the tradition is to give oneself a pseudonym on the public mail channel, like the "handles" used in the citizen's-band radio subculture) posted a bulletin entitled "Essay on Hacking," that said, among other things:

In the middle of Stanford University there is a large concrete-and-glass building filled with computer terminals. When one enters this building through the glass doors, one steps into a different culture. Fifty people stare at terminal screens. Fifty faces connected to fifty bodies connected to 50 sets of fingers that pound on 50 keyboards ultimately linked to a computer. . . . These are the members of a subculture so foreign to most outsiders that it not only walls itself off but is walled off, in turn, by those who cannot understand it. The wall is built from both sides at once.

These people deserve a description. In very few ways do they seem average. First they are all bright, so bright, in fact, that they experienced social problems even before they became interested in computers. Second, they are self-contained. Their entire social existence usually centers around one another. . . . Third, all aspects of their existence reinforce one another. They go to school in order to learn about computers, they work at jobs in programming and computer maintenance, and they lead their social lives with hackers. Academically, socially, and in the world of cash, computers are the focus of their existence.

As might well be expected, this diatribe did not go unanswered. As usual, opinion was heatedly divided. Some — a minority, of course — agreed wholeheartedly with the heresy. Hackers as a group harbor a love for heresy, iconoclasm, and debating whether something is or is not heretical, even if — especially if — the topic relates to hackers themselves.

Of those who rebutted Gandalf, the one known as "A. Anonymous" offered the West Coast version of the "Minsky defense":

We are dealing with an infinitely malleable tool. People who choose to develop and use that tool, whether for work, play, or both, have that choice and cannot be denied it. A person who chooses to be a musician must devote hours and hours to gain adequate expertise. But would you consider the computer hacker any less creative than such a person? I certainly wouldn't. The computer serves not only as a workhorse, but also as an easel for exercising one's creative abilities. Therefore, in my opinion, the hacker has not limited myself at all. Rather, he has expanded his intellectual horizon because now he has the infinite tool.

As for the charge that it disrupts one's social life, I would tend to agree with this to a point. But it depends on how controlled the individual is. At any time, he can withdraw to a more normal schedule. Why doesn't he? The reason is obvious. The infinite tool that knows few boundaries is accessible to a much higher degree, and thus he can devote more time to it. Why is this wrong? I think it is definitely a bonus, since the usual restraints of 9-to-5 are eliminated and the person is allowed to expand beyond boundaries to do what he wants.

Now we come to the human versus the machine factor. Gandalf stresses the necessity of human interaction and the inherent evil of the machine. Would you stress the evil of instruments in an orchestra, or the instruments in a laboratory, or the typewriter of an author? All of these occupations demand extraordinary amounts of time for excellence. But I see no greater human interaction in these fields than in computers. I feel that people who disparage computers for a seemingly decreased human interaction are not at all familiar with the true import of the computer. Not only is it the infinite tool, it is also an extremely fluid medium of communication.

The publication of the controversy set off an avalanche of electronic mail over the ARPAnet and at local computer centers. The hacker debates had spread to the amateur "bulletin board computers" by 1983, when the movie WarGames and the real-life young computer-systems "crackers" who subsequently surfaced brought the word hacker to widespread public attention, in this newer, unpleasantly restricted sense.

One of the oldest rules of the game is "thou shalt not do unto ordinary computer users what thou hast done to other hackers." Almost all of the old-time hackers deplore what the young computer trespassers and crashers did — "dark-side hacking" — although the anarchist minority still insist that the ultimate freedom is the freedom to figure out how the communication-computing system works, and declare that the burden of protection against trespassing ought to be on the system programmer who has files to protect, not on the explorer who might tap in during some midnight jaunt through the network.

Real computer criminals aside, the concern of the noncomputing public over the hacker controversy does seem a bit strange. After all, these people aren't accused of mayhem or arson — just of being very smart when it comes to knowing how to operate computers. The capacity for scapegoating is very high in a culture where most people have been led to believe that computers are either smarter than they are or too complicated for ordinary people to use. James Milojkovic, an associate of Zimbardo's at Stanford who was writing his psychology doctoral thesis about the cognitive and motivational impact of the microcomputer, came to the hackers' defense.

In a 1982 interview, Milojkovic said he spent plenty of time around hackers, and saw nothing pathological about what they were doing. In regard to all the public concern about what threat (noncriminal) hackers might pose, he said "clearly it's nonsense. I think what's happening is that there's some sort of fear that maybe what they're doing with the machines is aimed against us." Like "A. Anonymous," Milojkovic sees nothing wrong with a little compulsiveness in regard to learning: "I can think of nothing more natural than to fall in love with knowledge," he said, "and hackers are so deeply in love with knowledge of the computers that they're just swept off their feet."

A case in point: David Rodman. When last we saw him, lurking in the background of the MacHack versus Dreyfus match, an acidhead teenage dropout hacker, he was almost certainly headed for a sunken-eyed, computer-nut future. In fact, quite the opposite turned out to be the case. He was doing quite well for himself, even at sixteen, as a freelance programmer. He got some offers to set up computer systems for social service bureaucrats, so he moved to D.C. in his early twenties.

By 1972, David found himself up to his ears in the same problem that plagued Herman Hollerith — handling huge data bases. In fact, designing probes of the U.S. Census information, now stored on magnetic tape, was David's specialty. He moved back to Cambridge to work for a software think tank, did more than a few jobs for agencies he doesn't want to name, and in 1978 he decided it was time to turn what he knew into a marketable product.

David Rodman ended up creating and marketing a tool for managing data bases, a program that he designed to be usable by microcomputer owners. Thus he was one of many formerly sequestered programmers who joined the software business at the beginning of the consumer computing boom, when it was still possible for a programmer-turned-entrepreneur to go far and fast. A couple of other, older, MIT hackers put out VisiCalc in 1978 — the "electronic spreadsheet" that allows users to ask "what-if" questions about numerical data — and millions of people who had never touched a keyboard before began tackling problems that had formerly been reserved for mainframe programmers.

I first met David Rodman in the early 1980s, because of his strange grin. I knew his name because it was stamped onto the plastic card that was pinned to his lapel. His rumpled suit and convention badge didn't exactly mark him as a high roller, but his smile projected a self-assurance of almost demented intensity. We were standing in the magnificent casino that is conveniently located between the Hilton lobby and the indoor walkway to the Las Vegas convention center. Upward of fifty thousand people attending Comdex, a national convention for the microcomputer industry, trooped through the casino every day. The arriving computerists didn't mind spending their money, and they were an amiable group. A lot of them seemed downright happy. David Rodman, for example, was still smiling after he turned away from the craps table.

"Why do you look so damn cheerful?" I couldn't help inquiring.

"I was wrong about the dice," he replied, "but I'm too far ahead to complain."

"Craps?"

"Data management systems."

"Not my game," I said. "What's the product?"

"About forty pages of zeroes and ones."

"The market pretty good for zeroes and ones?"

"The software market, as of today, is nothing less than astounding."

Considering the fact that he had just dropped a hundred dollars in less than ten seconds, he must have been doing very well indeed to be making money at the convention faster than he was losing it in the casino. The crooked grin on his face, a variant of the slightly demented expression that attracted my attention at the craps table, made it clear that he didn't mind talking about his business.

We got to know each other, and eventually I learned about what he did before he was the prime mover and chief asset of a software corporation. There was no sign that he was an ex-MAC hacker, ex-acidhead, ex-consultant to unnamed intelligence agencies. He was freckled, balding, and what hair he had left was short and neatly combed. He was clean-shaven, and his attire wouldn't have been out of place on an accountant or a widget salesman. But in his heart, he was still a hacker, and an evangelistic one at that.

By the time we got through the story to the point of talking about his current product, it was clear that he had not turned his back on the programming priesthood, but was merely interested in expanding it, to his own profit, by giving millions of people a direct taste of the same experience that hooked him back in Building 26.

"I remember the way I learned jazz improvisation, and how that affected my programming. When I was first learning, I said to myself, 'Here I am in this chord, and I've got to get to that chord.' The transition, the way you hop from note to note or pass a variable from procedure to procedure — that's where the individual style of the musician or the programmer comes in. Nothing happened, a lot of the time. But when my teacher showed me something I hadn't realized before, pointed out that a certain note would work in a way I wasn't expecting, for instance, I would get a little shock of understanding, and the next time I came to a transition I'd loosen my grip on my conscious effort and try to recapture that shock, and there would be the note or the line of code I needed.

"Now I think of the person sitting in front of his computer with a keyboard. What this person needs is a profit and loss statement, or information about sales accounts, or a breakdown of stock in inventory. What I need to do is to create an environment for that person, structured in such a way that it is natural and easy to translate his or her desire to the actual P & L statement, or a sales report or inventory account, and even show them how they can improvise along the way. Not only should this tool work better than their old pencil and paper and calculator and filing cabinet — it should also give the user one of those pleasurable little shocks. I want my file management system to enable that person to become a jazz musician.

"A really good program designer makes an artist out of the person who uses the computer, by creating a world that puts them in the position of 'Here's the keyboard, and here's the screen. Now once you learn a few rudimentary computer skills, you can be a superstar.' "

It was an unexpected, but perhaps not inappropriate philosophy to hear from a LISP hacker turned software vendor. He has yet to carve out an empire like Bill Gates or Steve Wozniak, but David Rodman knows that most of the potential consumers of microcomputer software are still in the earliest stages of their progression toward obsessive software intoxication. David sees a niche for people like himself as toolmakers and trailblazers, leading the way for the emergence of an entire population of programming artists. He wants programming to become a performing art.

But long before hackers started thinking about using their computers for intellectual improvisation — before David Rodman was born, in fact — a dreamer out in California was designing his own kind of mind amplifier.