Jef Raskin's story "Meeting Merlin"
Merlin's console and tape units
Meeting Merlin
It is a commonplace now for a youngster to learn to use a computer long before they've gone to high school. Most of them use computers as appliances for game playing, web surfing, writing homework papers, and emailing and IMing (Instant Messaging) with friends. Computers are nothing special any more (thank goodness), but a handful of youngsters do go beyond the computer-as-appliance and learn to create web sites. Of those a still smaller group come to the wonderful realization that a computer is the most plastic medium yet devised. You can program it to have almost any behavior you can dream up. Some limit their imagination to creating game worlds, others realize that they have a supreme tool at their hands and apply it in all different directions (and usually write a few games on the way, too). By the 1990s, most kids had been exposed to computers. Not everybody used them or had one, but everybody knew about them.
I was not particularly different, I ran into a computer shortly before I was to go to high school, and was dazzled by the power and potential of programming. However, by extraordinary good luck, I met Merlin. Merlin's magic warped time for me so that my introduction to computers and programming occurred in the 1950s, instead of the 1990s. This put me 40 years ahead of my time, and I was later to make use of this privilege to create the Macintosh project, which may well have jumped the progress of personal computers ahead by a few years.
Many science fiction stories relate the difficulty of time travelers when they try to convince people of an earlier time about some fact from the future, or to get them to change their culture or technology in some way that, to the reader and protagonist, is clearly beneficial. For me, this is a reality, not a fiction. I experience it today, as I try to make the computer industry realize that its products are obnoxious and need radical revision, not incremental improvements. We all know this: Personal computers and other microprocessor-based products (cellphones, PDAs, high-end cars) are the butt of a continual stream of jokes. I know how to redesign their software and thereby fix a lot of the problems, but changing the status quo is difficult. It is amazing to me, but most people who are familiar with today's technology cannot understand the solutions I present, though the solutions are simple and straightforward, because they are so thoroughly embedded in the present. I guess it would be worse if I were trying to explain how a radio works to a farmer of the 16th century.
I am not being arrogant. This has just been my experience ever since I met Merlin. For example, getting the people at Apple Computer to understand the concepts behind the Macintosh was extremely difficult. At first, nobody at Apple could grasp the ideas, which were alien to the way they viewed computers. This held up the product for a year or two and resulted in its being dumbed down from what it could have been. They did the best they could.
Miriam Meisler (nee Horowitz), now a professor of genetics at Ann Arbor, was, always, my favorite cousin. Almost alone in the great morass of our extended family, she was destined to be a scientist and her questing mind was more open to the future than most. I confided in her, and too often I wished that she were not my cousin (it didn't help that she was a very attractive young woman). Recently she wrote this to me, "I have that vivid memory of you on a subway station (West 4th street) in the early 60s. talking about computers as household appliances." At the time it was a fantastic notion, a vision held by very few. Science fiction writers and buffs saw the burgeoning of the computer age coming, but they had it all wrong. Where they saw huge, wild behemoths, I saw tiny, tame household pets. Read the SF from that period, feel the hum of the massive transformers, hear the whirring of the huge reels of tape, see it burst into flame when presented with a contradiction.
My life would have been very different, and I would not have met Merlin, if it were not for a gym teacher who switched over to mathematics; L. Roland "Ronnie" Genise. His conversion was profound, and he loved mathematics with a rare passion and a gift for conveying it. When I email to my former classmates, he is the one teacher that stands out in all their memories. He lives not far from here, we both having switched coasts, and we still occasionally chat. He moved up in teaching grades much as I progressed, so that his influence and support were present for more years than is usual between a student and a math teacher. I was a geek and a hacker before those terms were invented, and got little emotional support from my fellow students. But Mr. Genise (as I then called him) was my life preserver. My parents were another line of support, and I cannot say too many good things at my luck in having the parents I did, but that would take this essay too far afield.
I met Merlin on a field trip organized by Mr. Genise. In a class of mostly unruly students, I was one of the few that gave teachers little concern. I would show up for the bus on time, never got lost, did not pull pranks, did not fight or throw things, and in general was a reliable little nerd. On this field trip, we went to Brookhaven National Laboratory, then a hotbed of the latest in atomic physics and all things futuristic. I don't remember the date, but from all the evidence it was most likely 1956, when I was 13. It might have been a year earlier or later, but it is not important. We looked at this or that, and a docent explained things at a level that I found condescending and which my classmates variously found irrelevant, dull, or incomprehensible.
We were led into a computer control room and just a few of us made a side trip to see the innards. I remember someone opening a door so that we could see an array of dots on a screen, making a kind of fat cross. It was a "Williams tube" we were told, and each of the dots was one bit of information. The tube could hold 1,024 bits. "Why 1,024" asked a classmate and I blurted out that it was a power of two.
Ron Genise taught the "new math" that is, for some reason, still called the "new math". Its promise has become tarnished and ridiculed, mostly by those who do not understand it or have seen it misused. While it will always remain true that to do arithmetic by hand swiftly and accurately takes practice (to be blunt: drill), understanding does not come from that. Now that four-function (with square root!) calculators are throw-away cheap, there is no reason for a student to be snappily proficient at long division, but if you can do it, however haltingly, in various bases you understand division and will never forget it. I had learned to do arithmetic in any base, and Mr. Genise would ask tantalizing questions that were outside the book: what would it mean to use base zero, how about base -5, or base 1/2? "Useless questions" the detractors would cry, and if I were to have a career as a clerk making change, it might be useless. But it was intriguing, inspiring, and opened vistas. Thinking about base 1/2 leads quickly to thinking about infinite sequences, which is pretty heady stuff for a kid (how would you express 1/3 in base 1/2?). And not everybody in class was as excited as I was, but Mr. Genise had invented "nobody left behind" a long time before the Bush administration made up the phrase (and then forgot to fund it). Genise was somehow able to induce understanding in even the thickest head. He is still my model as a teacher, and when I remember to follow his techniques when I am teaching, I do especially well for my students. In any case, I knew my powers of two because I had to in order to do base 2 arithmetic and convert to and from base 10. Because of this I understood exponentiation, and found the concept of logarithms trivial. Do not denigrate the "new math". It leads to thinking instead of drudgery.
But knowing about exponents also led to one of the most crushing moments of my life when I was in sixth grade. I do not remember where I was when Kennedy was assassinated, nor where I was or who I was with when I watched the first moon landing, but I remember this moment with what seems like photographic detail. I know from studies of memory that my description some of this event (and any event I describe) must be at least in part a reconstruction, we are not video cameras, but the images are vivid, and the memory of the hurt is real. My parents, along with my honorary aunt and uncle, Millie and Jim, and some other friends were at the dining room table, playing penny-ante poker as they often did on Saturday evenings. I was in my room reading the New York Times, when I ran into an article about a new astronomical discovery. I think it was that some galaxy or other had been found to be much larger than expected, and farther away, and the universe was now known to be bigger than anybody had thought. Having read a lot of science (we had a few encyclopedias around, including a Britannica, and I had read those, cover to cover). I found the news staggering and momentous. So I ran into the living room to my father and said something like "this galaxy is ten-to-the-sixth times larger than we thought it was, and..." He stopped me in his gentle way, and said, "you know that I don't understand 'ten-to-the-sixth' or whatever you said." He was ready for an explanation, which is what I'd usually try, and I usually could explain things in terms of concepts he did understand. But this time I turned around and ran back to my room because I had suddenly realized that I had gone beyond my parents' knowledge. My all-knowing, wise, and well-read parents were not all-knowing. In fact, there was a lot I knew about science and math, and had known for a while, that I'd never be able to discuss or share with them. They could hear about it, but never understand it. I don't know for how long I grieved and cried over what I much later realized was a portion of childhood suddenly ended. But things were never the same. I was on on my own, I could never again -- in these fields which were to become the center of my life -- turn to my parents for answers or advice. This had been true for a while, and my father had tried to sit down and follow what I was saying, but I must have pushed the thought that it might be beyond him forever from my mind. There was no less love, and things were the same in all other realms. You should know that my parents were exceptionally brave and just people. The way they stood against segregation and racism BEFORE the civil rights movements of the 1960s was incredible; they faced the wrath of the whole town and did not flinch because they knew that they were right; we lost our livelihood, we lost many friends, we went from comfortable to nearly bankrupt, from driving Packards and Buicks to cheap imports, from owning a store in the tree-lined center of town to owning a store in the Puerto Rican barrio, all as a result of sticking to ones principles. I am incredibly proud of my parents, and they gave me a model of righteousness that I will always try to live up to. But a touch of reality had come into a young mind, the first intimations that one would have to be responsible for oneself and not be taken care of by father and mother. I was a bit young for this to have happened. So I cried that night, and many later nights when the realization would come upon me again.
Not every child chooses to memorize, for example, 20 powers of two, the first hundred primes, or the first 50 digits of pi. It is a skill no different than remembering the names of the members of your favorite baseball team and their batting averages. Depends on whether your hero was Mickey Mantle or Carl Friedrich Gauss. My comment about 1,024 being a power of two didn't explain anything to my fellow students or even the person who was showing us around. But it did catch the attention of a man sitting behind a chicken-wire partition. Nowadays it is not unusual for a grade-school kid to recognize a power of two, everybody has heard of "binary arithmetic", and has some vague idea that that's the language computers are supposed to dream in. But it was not common then. I suspect that from my tone of voice, it was clear that I knew why it was important that 1,024 was a power of two. As the others moved on to rejoin the main tour, he invited me to see what he was doing. I decided that that would be more interesting than the stock tour, so I walked into his "office."
It was there and then that I was formally introduced to Merlin, and to programming. My memory of Merlin had faded somewhat, even the name of the computer had vanished from my memory, but the son of one of its builders happened upon my web site and wrote to me. Soon I was awash in data and memories. The Merlin was based on the Maniac, which was related to the Eniac, which was the first electronic (as opposed to electric or mechanical) stored program general purpose digital computer. The Eniac was born in the same year I was -- I am precisely a child of the computer age.
.
Merlin's Racks
The web and its search engines allow people to make connections that would never have happened in earlier times. In the earliest days of the Macintosh project (1979) I saw that the main use of personal computers would be based on a rich network full of varied information and allowing person-to-person mail, and I wrote a proposal that Apple develop something based on the Arpanet that sounds very much like the world-wide web. The people at Apple were not convinced. (If you doubt me, the paper is available at www.jefraskin.com as an appendix to an article called "Holes in the Histories." Devoted skeptics can seek out the original physical piece of paper, which is in Stanford University's History of Technology collection. The web, pretty much as I had foreseen it, finally became a reality over a decade later. Merlin's influence, no doubt.
When I mentioned some of my history in a published article, William Higinbotham contacted me (the Web is wonderful for this kind of thing). His father, head of instrumentation division at Brookhaven, helped to build the Merlin's memory unit. William kindly sent me information and pointers to much more data about the machine and its times. He also sent me the photos of Merlin that accompany this essay.
But I know of no photograph of the desk the programmer had me sit at or the paper printouts that he showed me. Soon, he was explaining how programming worked, and after a while he had me writing a few simple instructions that moved data from here to there, or added them. I do not remember the details, but the power of iteration and the ability of a program to make a decision as to how to proceed depending on an earlier result or an input made a profound impression on me.
I was prepared for this not only by Genise's mathematics, but by my father's love of tinkering. My father was not one of those legendary figures who fritters away the family fortune on vain pursuits. He was a very practical man, spent his time running the family store, and there was no family fortune to fritter. In World War II he had worked at the Brooklyn naval shipyard where he helped build aircraft carriers. He once told me that at first he crawled over the narrow planks that bridged between sections of the nascent carriers, but that he eventually was able to walk or even run along them like everyone else even though they might be a fatal 50 feet above the metalwork below. He also worked for a company (whose name I have long since forgotten, if I ever knew it) which built cable testing equipment. He had in our basement workshop a wonderful gadget that he had helped invent which allowed the testing of the many wires in a cable to be done very quickly. It was electromechanical, and I understood it so clearly that I could reproduce one today if I wished. I loved producing mechanical linkages, and once made a wooden clock of my own design. The escapement and pendulum worked, but I think I never got around to providing it with more than one hand.
I still love to build mechanical things, but in my childhood I read a book comparing mechanical and electrical devices. "Which can carry more power," asked the caption to an illustration, "this one inch diameter steel rod or this one inch diameter copper bar?" I knew that copper was soft and steel hard, and I fell into the author's trap, and thought that the steel bar was the winner. But, as he explained, if we talk about conveying electrical power, the copper bar was the clear winner. And the amount of power it could convey electrically was orders of magnitude greater than the steel shaft could convey mechanically.
Mechanical linkages are magical, a motion of this kind here results in a motion of that kind (just what you needed) there. But electrical circuits are much more magical. Nothing visibly moves, it looks like nothing is happening, but sound comes out of a speaker or images move on a display.
So my attention moved from the mechanical to the electrical, and I started building circuits. My father dug out an old soldering iron and got me some solder. I scrounged parts from old radios and TV sets. By the time I was 14 I had a Tektronix oscilloscope (that's a story in itself) and had built myself a nice electronics lab with lots of outlets and a large AC ammeter to see how much current my whole lab was using (I still have the meter), and I began to design electronics. The door was a hanging piece of cloth, the walls were pegboard on 2 X 4 studs my father had helped me put up so that I could hang up my tools. He loved pegboard.
When I was 15 I published my first article (in Radio Electronics) wherein I had deduced, from understanding the electrical properties of barium titanate, how I might be able to turn a certain kind of capacitor into a random noise generator. It worked, and the circuit and article were published. An early success is a great motivator.
Then, too, there was phone hacking. This was before phones were "dialed" with tones, but by a make-and-break electrical connection sending what were essentially clicks up the line. As I learned more about the way the phone company's switching circuits worked, I began to experiment. From my daily piano practice, I had a good sense of timing and manual dexterity, and I could, with a handheld switch, dial numbers without a phone. It is sometimes amazing how seemingly unrelated studies coalesce.
It was at some point along my autodidactical studies of electronics that my earlier experience with the Merlin began to well up, combined with what I was learning about the phone company, and with another hint provided by Mr. Genise. The hint was a classic paper by Claude Shannon on information theory. "I don't quite understand this," said Mr. Genise (he was always clever about how to motivate me) "but it looks like something up your alley, and when you figure it out, explain it to me."
This paper was another waypoint along my life's path. It was not easy going. The realization that information, which I had thought a wishy-washy concept not worthy of scientific discussion, could be handled as a branch of physics surprised me (and, as I was later to learn, much of the scientific world). That information theory was, in fact, exactly the same thing as thermodynamics was astounding. I was forced to redouble my efforts to understand physics, the difficult concept of entropy (that is so often abused in the popular press), and to learn some statistics. I was inspired to begin to learn calculus while my peers in 9th grade were tackling elementary algebra.
Mr. Genise was always encouraging, putting me in contact with college professors when he couldn't answer my questions. He was also always taking new math courses and had new wonders to share with me. I discovered a few theorems in number theory (modular congruences) and was partially elated to learn from a math professor that I was correct, and partially deflated to learn that Gauss had discovered them all, and lots more, hundreds of years earlier, when he was a teenager just a few years older than I was then.
All of these studies began to come together into a coherent structure. The computer was even more amazing than other electronics. This is how I put it to myself: In mechanics the parts moved but a machine would do only one or a very small number of tasks. To make it do something different required designing a new machine, usually from the ground up. In electronics, the electrons moved and could be redirected by switches. The speed and reliability were far higher, to design and build much cheaper and easier. The computer was something on a higher plane entirely. It was, in a sense, all possible circuits, all possible electronic devices. Instead of making it do ones' bidding by assembling and connecting parts, you made it do what you wanted with words. Words you typed, words you put in with punch cards or tape (in those days). They were not English words, but they were strings of symbols, and it was clear to me that they were, in essence, no different.
I come from a culture where the Word was held in esteem, words had nearly magical powers. Books were precious, one did not write in them or deface them in any way (a taboo that would hurt my studies in later years, it took me a long time to shed it). That words could communicate ideas, that words could preserve the thoughts of people long dead, that words could raise emotions and move people to action, said that written language was one of the more incredible inventions of all time. Reading and writing are taken for granted by almost everybody, but the Jewish culture my non-religious family was embedded in had not forgotten how special language is.
We even call our orders to the machine a "programming language". That day in the dark cage of the computer, the programmer showed me the first few incantations. I understood that each was a step that the machine would follow, and I raced to learn each new command (in "machine language"). Hidden away from any windows, we lost all sense of time. Programmers keep long hours, and work late into the night. This happens to me still when I am programming. It happened to me for the first time then. The bus left, but nobody thought to check to see if I was aboard because I was the faithful student who was always on the bus, never causing trouble, never running off. I normally rode my bike home after school, so the first person who missed me was my mother. She called the school -- it was closed. She remembered that I was on a field trip and called Mr. Genise. He was sure I had been on the bus. I was always on the bus after a field trip. But he hadn't looked specifically for me, he had to pay attention to the spitball-throwing herd and keep them in their seats. I was probably in the back corner as always, where nobody could sneak in behind me, reading a book.
My bike was still at school when my mother drove to see if I had gotten hurt. Had it been daytime, she would probably have stopped at the library, where I had been known to get lost forever in a book. Finally, they called Brookhaven. The guards there followed the tour path and finally found me, at 11:00 PM, totally unsleepy, not at all hungry, and programming away. A call to my parents let them know that all was well, and they somehow knew when they came to get me that a display of anxiety was in order, without anger. I didn't get into trouble, nor did the programmer or Mr. Genise. Things might be different today, but there was an assumption of good will all around back then.
I had no satori immediately, though I had learned the basics of what programming is about. It did take a few years to realize the implications: by putting words into a non-living machine, it could do your bidding. No cutting, no filing, no screwing nut onto bolt, no oil, no sound, no soldering, no calculations of P = IE or E = IR. Just words, as in the Jewish creation myth where everything was called into being by God's pronouncement, "Let there be...". Now I wanted computers. I had met Merlin at a crucial moment in my life, and a few years later I finally understood what Merlin could do. Put in the right spells, do your incantations (but you must do them exactly right, as the fairy tales warned us) and your wish would come true. It was like magic, it was real, and it was mine to command.
Technical notes on the Merlin
Based on the Los Alamos Maniac II, the Merlin had 4096 words of 48 bits each, or 24,576 bytes of memory, and some check bits sufficient for one-bit error correction. Williams tube storage was chosen over magnetic core memory because the former was $0.05 per bit and the latter $0.20 per bit. By way of comparison, the memory card I keep in my pocket has a capacity of 512,000,000 bytes and my computer's memory is the same size (each represents the memory of over 20,000 Merlins). The cost per bit is of my memory card is $0.0000024. With its air conditioning, The Merlin required 56,000 Watts of power and could do an addition in 8 millionths of a second; the computer on which I am writing this requires 170 Watts and can do an addition in about a billionth of a second. My machine is thus at least 8,000 times faster (in fact, for other technical reasons, the ratio is much higher). My complete system cost under $6,000. The Merlin cost $600,000, or a hundred times as much (unadjusted for inflation, which would add another zero to the Merlin's cost). Some of the Merlin data is from http://tingilinde.typepad.com/starstuff/2003/09/computing_at_bn.html
The Stats on Merlin
The first computer I ever programmed (I was about 12 at the time) was the Brookhaven National Laboratory's Merlin. The son of one of its designers sent me this information:
PROGRAMMING AND NUMERICAL SYSTEM
Internal number system Binary
Number of binary digits per word 48
Number of binary digits per
instruction 48
Number of instructions per word 1
Number of instructions decoded Approx 90
Arithmetic system Floating point
Fixed point
Instruction type One address (mostly)
Two address (some)
Instruction word format
+----+----+----+----+-----------------+--------------------+
| Y | Z | b | b' | m | m' |
+----+----+----+----+-----------------+--------------------+
YZ = command (two hexadecimal characters)
b B box address for lst address, 2nd address
b' (4 bits each)
m lst memory address (16 bits each)
m' 2nd memory address
Automatic built-in subroutines:
Square-root
Fetching of next
sequenced instruction begins before completion of operation.
MERLIN is patterned after MANIAC II (Los Alamos).
ARITHMETIC UNIT
Incl. Stor. Access Exclud. Stor. Access
Microsec. Microsec.
Add 8 us (3.5) -
Mult. 140 us 130
Div. 330 " 320
It cost about $600,000 and used 40 Kw of power. No power-saving sleep mode back then.
The above information (there's more) can be obtained from: http://ed-thelen.org/comp-hist/BRL61-m.html
.