The Fonly Institute

The Very Thought of You

The Very Thought of You

21 Oct. 2009

It's 1979, and I have my first workshop that's not a garage. I had been talked into sharing a garage in Berkeley, where I'd been (taking a year off twice) since I came to school at the university in 1963, by Bob Marsh, a fellow resident of my student co-op from the early days. I didn't really know Bob then, but we met in 1974 when he was looking to set up a workshop and needed someone who could 1) help, and 2) share expenses.

We rented a 1200 square foot garage at 2465 Fourth St., down in the industrial district by the bay for $175 per month. I moved my clumsy wooden workbench in and Bob set up shop in a small office upstairs in the back. I was doing business as an independent engineer under the name “LGC Engineering”, a reference to the utopian poem “All Watched Over by Machines of Loving Grace” by Richard Brautigan, picking up little bits of contract work here and there.

Then the Altair personal computer kit was announced and things picked up. Bob started a company, Processor Technology Corporation, to make plug-in boards for the Altair and I began getting contracts from them to do engineering and technical documentation work. Five years on I had designed their video display board, the VDM-1, which set the standard for personal computer text display, and had designed a computer around it, the Sol-20 “Terminal Computer”, so called because it was shaped like a terminal (someone wrote “like a typewriter with nowhere to put the paper”). The order book was opened for the Sol-20 in August 1976, half a year before the Apple II, and it got displayed in the Smithsonian for 17 years.

I made a deal for royalties in addition to a fixed fee for each product, so for every unit shipped I got $10 (VDM-1) and $12 (Sol-20), paid quarterly. I wouldn't have to give any money back, no matter how the company fared. In 1976 Processor Tech moved out of the garage and into bigger quarters, and I moved my workshop back to my studio apartment. In 1978 they had moved again, 40 miles away, and I decided to use my new money to rent a spiffy office and set up the workshop of my dreams.

I rented a space which would have accommodated two 400-square-foot offices, on the second floor of a building overlooking a parking lot behind the University Avenue Co-op market (of which I was a member). The rent was $800 for 800 square feet, the place was carpeted but otherwise plain. I had some fairly radical plans for it.

My father had worked for a company that makes folding tables and chairs, so through him I was able to buy four “serpentine” tables at cost. These tables describe a 90 degree sector of arc, and can be connected to make circles, U-shapes, or oscillating shapes (hence the name). Each one was 4 feet in radius to the outside, and would fit a 2 by 3 foot sheet of “D-size” drawing paper. Seated at the center point you could reach everywhere on the table, with no “shadow areas” out of reach where clutter could pile up (I had heard architects talking about this phenomenon).

I got four rolling carts each of which held a top-accessible filing drawer (I called them “file pots”) and filled them with hanging file folders. The bottom portion of these carts was just a shelf with clear space up to the bottom of the file pot, and there I placed a simple plastic tray. I put track lighting, with little spot lights that could be re-positioned, on the ceilings.

My idea was for a non-territorial office, in which people would set up their tables where they needed them and then fold them up again before they left. We used director's chairs for seating, and all pencils etc. had to be swept into the plastic trays, the tables and chairs folded up, and all stowed away at the end of the day. We also had a sofa, a drafting table, and a cabinet with a blueprint developer machine on it. It worked well, but only for the people I had hired – when a client placed one of their people in the office to work on a project with us he demanded room dividers to create the equivalent of a little closed office.

In those days, when computer aided engineering (CAE) programs were in their infancy, engineering drawings were still done by hand, with special pencils (called “lead holders” because you could release the thick lead with a push of the button) – that you would perpetually sharpen by twirling them in special pencil sharpeners. You rotated the pencil while you drew the line, so that the point would, in effect, continue to be sharpened (or at least not get dull as fast as it would if you held it still).

To reproduce these drawings, which were done on translucent “vellum” paper, you put the original together with a sheet of “Diazo” paper, which had a yellow photo-sensitive coating, with the coating contacting the back of the original. Then you fed it into the developer, which wrapped it around a rotating glass cylinder with ultraviolet-rich fluorescent tubes inside, giving the paper a good dose of UV through the original.

Where there was a pencil line, the UV was blocked. The coating bleached out in the clear areas and as it emerged you fed the copy paper through another section where it was exposed to ammonia gas, which turned the unbleached coating dark blue. The result was a “blueline print” - the negative of the earlier “blueprint”, which had turned the clear areas blue and the lines white.

The ammonia came in gallon jugs and hooked up to the machine through a special screw-on cap with several tubes connecting it to the machine. The whole rig cost $1000, a lot of money then, but you couldn't do the work without it – Xerox copiers were over ten times more expensive, and there was no such thing as a laser printer for use with a small computer.

If you wanted to make printed circuits you had to do a “tape layout” on a back-lit light box, sticking down thin, black, flexible paper tape to layers of clear Mylar film representing the copper traces on the fiberglass board. You would take the layers of Mylar and tape to an industrial photographic house and have reduced negatives made and reduced from it. If you had dropped your roll of tape on the floor it would pick up lint and hair, which would result in tiny, invisible copper bridges between conductors if you didn't run your layout knife along the tape to clear out the lint.

At that time CAE layout programs were just coming into use. They had been developed inside large corporations that could afford the really expensive precision photo-plotters along with the computers needed for the work – a system might cost $250,000. Eventually some of the used equipment made it outside to “service bureaus” where you could rent the use of it, and the time of a skilled operator, for a hundred dollars an hour or so. They had to work using a digitizing pad from a layout you made on grid-ruled frosted Mylar, using colored pencils, red on one layer and blue on the other (layouts of more than 2 layers were rare).

The grid was 10 lines to the inch, and you usually worked at a 2X magnification, so each grid line represented 0.050” (or “50 mils”). The printed circuit fabrication technology of the day allowed for 12 mil line width and 12 mil spaces – they could do finer but charged a lot more per board. So we could fit one trace on each grid line and one between.

So it happened one night that I was applying myself at the drafting board trying to work out the routing for my “VDM-2” display board – a long-overdue successor to the now-obsolete VDM-1. This product would have 24 lines of 80 characters (the VDM-1 had 16 lines of 64), have two display fields that could be scrolled separately – one scrolled smoothly, and a character font that could be written from the computer (as opposed to a cooked-in ROM character font on the VDM-1). All this took lots of logic, which took lots of integrated circuit packs, which took lots of area on the board. I had to extend the top edge of the board beyond the standard S-100 outline to the limit of what could be accommodated in the Sol-20 cabinet. I called it the “magnum” board outline.

Still, it would not fit. The going was hard and slow, and I had to stay late.

I was listening to KPFA, one of Berkeley's treasures – the world's first listener-sponsored radio station, established 1949. They would let inexperienced interns learn broadcasting skills by taking late-night shifts, and you could always expect to hear unusual things after midnight at their frequency. Tonight was no exception. The program was jazz, and the DJ for some unknown reason had decided to play every available cover of an old standard, each by a different artist.

The song was “The Very Thought of You”:

“The very thought of you/

and I forget to do/

the very ordinary things/

that everyone ought to do...”

It was pleasant enough. Then on it came again, by a different singer with different musicians. Still, I thought that was interesting, too. Then again. Then again – and again. They kept coming with no end in sight.

I was working urgently to get the layout done - quite likely an impossible task. I had to get the prototype to Processor Technology on a deadline, as they were about to leave for the National Computer Conference in New York. I was alone in a spare office with nothing but parking lot lights outside and no living thing in sight.

On and on it went - “The very thought of you...” It was going to go on forever! I was struck by the thought that perhaps I had died, that the sun would never rise, the layout never get done, the song would go on and on and on throughout eternity as I tried and tried and failed...

That was my first glimpse of hell.

The music stopped eventually, the sun rose, though the layout never did get finished as I had to cut it short and fill in for the missing printed circuit traces with wires. Shortly thereafter I found myself standing with the not-yet-working VDM-2 board in the multi-story Javits convention center in New York City, looking for the Processor Technology both, sweating in the summer heat.

I searched and searched - and couldn't find it. There was no Processor Tech booth. I located a bare space where it was scheduled to be - Processor Tech, it turned out, had closed its doors. Future royalty payments would not be made. I would try to sell the design to other manufacturers without success. Winter was coming.

Processor Technology Corporation died for a number of reasons – not all of them clear to me. I do know that they made the mistake of staying with their first products too long. By 1978 the Sol-20 was in dire need of an upgrade from the 8080 processor chip to the Z80 – highly desirable for running the CP/M operating system which had by then become the standard. They never asked me to design any upgrades – the VDM-2 was my initiative. When I asked them too late in the game what they wanted me to do the principals said “we were waiting to see what you came up with”.

I concluded from that experience that the time to start work on the next, improved version of the product is the day after you finish with the previous version. With luck you'll finish before your company needs it – this gives you time to adapt it to the world as it is, or rather, as it has become in the meantime. If you're really lucky and skillful you can create a range of versions so that when the competition pulls their product out of a hat your company will be able to open its closet and choose which product to release that will be better and cheaper than the competition. That and still have higher and lower-end products in reserve.

Of course, that takes a company run by competent professionals – something we didn't have in those days.

Be Now Here

Be Now Here

14 Oct. 2009

It's October 1995, and I am working at a research lab running their prototyping shop. I have about 8 people working “with”, actually “for” me to design and build equipment other researchers will need.

One of those researchers is Michael Naimark, a curly-headed fellow who looks younger than his years and has a massive reputation in interactive video media. In 1980 he had produced something called the Aspen Project, in which four cameras were mounted on top of a Jeep looking out in each direction. A device like an electronic odometer was mounted on the wheels of the Jeep to record how far it travelled and to give a signal when it had moved a certain distance. This signal triggered the cameras and produced a string of four images whenever the jeep traversed that distance, regardless of speed.

Naimark took the Jeep to Aspen, Colorado – a ski resort in an old western town - and ran it back and forth over the grid-patterned streets of the town, covering every street. He then took the photos, scanned them to video and recorded them onto a videodisc. It was rather like like Google Maps' “Street View” without the Internet.

A word of explanation is needed here. In the '70's “laser videodiscs” were supposed to be the coming thing, at least according to IBM. They would be the medium whereby everyone watched movies in their home. A single videodisc could hold 60 minutes of high-quality video, which could be presented as still pictures as well as continuous video. “Like DVDs?” many young people today would ask. The answer is, yes and no. Like DVDs they were read by laser beams and could provide frozen frames. Unlike DVDs they could be produced only at centralized facilities, like LP records had been (there's another technology that's gone into obsolescence).

IBM loved the centralized production aspect of laser videodiscs. Everyone would sit at home, they hoped, getting all their information needs met by these discs, which IBM would happily crank out for a fee. The discs would provide video catalogs of everything people wanted, allowing them to shop from their own living rooms, by “interacting” through a single “BUY” button - the purchase information would be routed through the phone lines to the vendor's systems, also manufactured by IBM, who would take a percentage of the price.

In 1978 I was paid to join a group of luminaries (including Ted Nelson, the prophet of personal computing, and Murray Turoff, a pioneer in online conferencing) at an austere resort in Rhode Island for several days to discuss these wonderful devices and their part in the information economy to come. The consensus of this conference was negative on videodiscs – we felt that computer networks were the proper technology to foster (and we were right). IBM, which sponsored the conference anonymously (a fact to which many participants had been tipped off in advance) was said to be mightily disappointed in this outcome and, true to form for a large organization, they ignored the results. IBM supported the Aspen Project through the M.I.T. Media Lab, where Naimark was on the original team.

The interesting part of the project came on playback. Because the discs could show one picture at a time and seek those pictures randomly (unlike a VCR) the user could, using a joystick, navigate their own path through the town and look in any direction as they did so. It was a sort of “Virtual Reality” in that you could define a unique path that had never been followed and see the scenery in the correct order - “interactive video media”.

Naimark went on to explore in the field of “location-sensitive cameras”. He tapped into a worldwide network of geographers (a concept worthy of a Thomas Pynchon novel) and, by the time we worked together had secured grant funding from UNESCO to record detailed imagery of several sites designated as “endangered heritage sites” by the UN cultural organization. He had committed to taking sets of stereoscopic, high-resolution 3D panoramas of these sites, several over the span of a day, and presenting these scenes to audiences in the most realistic way possible.

He asked my group to build him a rotational platform mounted on a tripod and holding two 35-mm Panavision movie cameras. He wanted to have a motor turn this platform under precise control, so that the cameras would shoot exactly 360 pictures per rotation – one per degree around the circle. We did that, using a precise quartz crystal oscillator divided down by a counter circuity I designed, producing an accurate 360 pulse-per second reference signal.

We connected a motor to drive the platform through a gear train having a 360 to 1 ratio – thus the motor would turn once for each picture taken. A simple “exclusive-or” comparison circuit would look at the 360 Hz reference signal and a signal sensed from the motor shaft. When the two were in a “quadrature phase relationship” the transition of each would split the “on” or “off” portion of the other exactly in half, and the average coming from the comparison circuit would be zero.

If the speed of one signal started to drift from that ideal point where one signal split the other equally the proportion of “high” to “low” in the comparison signal would change. Instead of zero, the average of the compared waveforms would move in the positive or negative direction. Converted into motor speed, this average comparison signal would cause the motor drive power to adjust for any change in external conditions, such as resistance from friction, and the platform rotation would proceed at a speed dictated by the rate of the oscillator signal - and quartz crystals oscillate at very, very precise rates.

We packaged the whole thing up so there was only an on-off switch and an LED and Naimark took it around the world – to Jerusalem (the “Wailing Wall”), to Dubrovnik, Croatia's main square, to the Angkor Wat temple complex in Cambodia, to the center of Timbuktu in Mali, and finally to the park outside the Yerba Buena Center for the Arts in San Francisco, where the system would later be exhibited. Panoramas were recorded in all places five times in a single day, each at the same hour of the day, and the pictures were transferred to a laser videodisc.

But how to exhibit it? The videodisc played back a picture that seemed to move in a straight line, although the viewer might realize it had been viewing a circular panorama when the picture returned to the same point at the end a minute later. But that was a minute too late – the audience had to understand in their viscera that this was tracing out a circle as they watched it.

Naimark and I hit upon the solution at the same time – rotate the audience at the same rate the camera had rotated. A revolving platform was bought – the type typically used to display automobiles at trade shows, adjusted to turn at one rotation per minute and set up as the place on which the audience stood. The 3D image was projected on a stationary screen (the audience put on polarized 3D glasses when they stepped onto the platform). The whole thing was placed in a blacked-out room so there was nothing but the screen to look at. This way, every member of the audience would be shifting their position on the turntable to keep the screen in front of them – in one minute they would all turn 360 degrees relative to the platform surface.

In this way the visceral understanding was created in each viewer that they were turning – they were each shuffling around in a circle as the turntable ran underneath them. The view on the screen seemed to be a window to the outside world that tracked with the viewer's body's own rotation and thus revealed the panorama. It worked!

A few days before the exhibition I led Mike through the intensive process of writing a patent on the system under deadline – the exhibition would “disclose” the method and render it unpatentable afterward. I added my own contribution – a one-person audience on a swivel chair having a rotation sensor, and we both signed as inventors. Since this was Mike's first patent I encouraged him to take the status of lead inventor – he certainly deserved it – and it went through the US Patent Office in record time – clearly the examiners could find nothing anywhere like it.

The system was exhibited in several venues under the title Naimark gave it - “Be Now Here”. At one point I walked into his office as he was talking on the phone and wrote “Be NowHere” on his whiteboard, connecting the last two words into “Nowhere”. Mike broke from his phone call to say “you're the first one to get it”.

The Police Radio Incident

The Police Radio Incident

Oct. 9, 2009

It's fall of 1964 and I'm involved in a revolution. I'm on campus at Berkeley, though I have no classes – it's the work phase of my work-study program and I'm employed as a lab helper working for Dr. Westheimer in the Physiological Optics lab. The revolution is the Free Speech Movement, a 3-month-long struggle to roll back new university rules that would have banned organizing civil rights protests on campus. It's serious because Freedom Summer has just concluded and some of the voting rights workers didn't come back alive.

You can learn the particulars at the website of the Free Speech Movement Archives (www.fsm-a.org). Revolutions generate lots of stories, and this one is just a snapshot from that time, though an important one to me. I feel justified in using the term because revolutions, 1) involve the actions of a large number of people, and 2) overthrow an existing order, in this case the order of “in loco parentis” by which the university arrogated to itself many rights of parenthood.

The FSM headquarters is in a small house in the South Campus student ghetto. Its occupants – major figures in the initial explosion of protest – have retreated to one or two rooms and dedicated the others to organizing the protest activity. The place is full pretty much 24 hours a day and even has a self-appointed housemother named Marilyn Noble directing people here and there. Marilyn is a grad student in sociology and her father is a mechanical engineer who worked on the original Douglas DC commercial planes in the 1930's.

I am hanging around trying to find some way to apply my skills, such as they are. I'm 19, a sophomore in EE, totally intimidated by all these brilliant grad students in fields I do not understand – sociology, psychology, political science – areas dealing with human behavior and relationships unknowable to me. Maybe some of these brilliant people will tell me what they need to have done, and I'll do it.

In the meantime, I gravitate toward the non-existent card file system and attempt to improve it. I want to try my hand at implementing McBee “keysort” cards. These are card with holes punched around the periphery. If you stack the deck square they all line up and you can put a knitting needle through the deck. If you pull up on the needle all the cards will come with it.

If you had punched out the cardboard from a hole to the edge then the card with the punch would not come with the needle when the deck is lifted by that hole. By clever definition of hole fields you could do binary sorting and have a very large sort space. But first you have to have the cards to punch.

I try my hand at taking a stack of index cards and using an electric drill to bore holes through. Don't try this – it only produces holes filled with wood pulp fuzz. The term “drill” is used with paper where the term “punch” ought to be used. You also have to take great care that the cards don't shift and put your holes out of alignment when you're making them.

Still, it keeps me occupied and there might be something more successful I can do. One evening there's a commotion - a group of students burst through the front door shouting about how the police have surrounded the campus. We all freeze – Berkeley's campus is very large and the logistics involved in surrounding it would be tremendous. Everyone wants to know what is actually happening.

Marilyn Noble takes the lead, turning to me. “Quick,” she says, “make us a police radio!”

At this point I enter a surreal space, somewhat at variance with the experience of everyone else. It seemed to me that everyone in the house had barked those words at me in unison. I immediately knew what they were talking about – 25 years earlier police radio was positioned on the spectrum immediately above the 530 – 1700 Khz AM broadcast band. It was possible to take the case off any AM radio, re-tune the local oscillator to a higher frequency and hear the police traffic.

However, 25 years has passed, and the police band has moved up to the 30 – 50 MHz range and gone to narrow-band FM – a totally different proposition. Modifying an FM receiver would be a real problem and require some redesign. I recall that there had been a construction article in one of the electronic hobbyist magazines with plans to build just such a receiver. But you couldn't just snap your fingers and create it.

I try to explain, with all eyes on me. What comes out is a stammering, “You..you don't understand...it takes time..” The crowd all seems to answer in unison again - “Never mind about that! Make us a police radio!”

At that instant two things happened in my mind. At the practical level, I started a project to build a police radio from that construction article. At a deeper, more visceral level, I was stunned to discover that my strategy of “waiting for orders” was fatally flawed. Those frighteningly sophisticated students, knowledgeable in the many ways people interacted, had turned out to be woefully ignorant of anything having to do with technology. If I were to put myself at their disposal it would be guaranteed that I would be working on the wrong thing, too little and too late.

The realization continued, crystallizing at the same instant - I would have to make my own decisions on what technology to pursue that would support the causes I hoped to assist. I would have to take the time required to gather the knowledge and skills, to design and build and test and re-design, to organize the best ways to get things made in the necessary quantities, all without the “leaders” understanding what I was doing. Then, when the demand arose, I would be able to respond, “Well, you can't have that but here's what you can have!”

Whether I wanted it or not, whether I felt qualified or not, I would have to take the lead in any work I did in developing technology for use in society. I could see the path I'd have to take, and it involved educating myself in many of the mysteries of human behavior along with the arcana of technology. It was daunting, but it was better than waiting for orders.

It turns out that revolutions have a third characteristic – they open previously unsuspected ranges of possibilities for their participants and others. The Free Speech Movement won on December 8, 1964 with a vote of the faculty senate that demolished the edifice of “in loco parentis” and demanded that the only authority over civil behavior on campus be the civil authorities, under the Constitution. The crowd of students waiting outside the hall where the senate met formed corridors and applauded the faculty as they emerged.

A few months later I heard a talk from one of the grad students commenting on what had happened that mobilized tens of thousands of students, who were otherwise strangers, to create a community focused around the protest. He said that “barriers to communication between students had come down”, and that anybody could talk to anybody else about the crisis without an excuse. I took that observation and began a long process of trying to find out how to make that an everyday situation, and what technology would be appropriate to that purpose.

If I hadn't had that surreal experience, evoking that instantaneous conclusion, I don't know what I would have become. For the participants, it seems, revolutions never really end.

Becoming Digital

How I Became Digital

1 Oct. 2009

I was analog until the end of 1970. Starting at age 12 when I received an unanticipated gift of a partially-used correspondence course in radio and TV repair, I lived in the analog domain as far as electronics went. Audio was something I could do, and repair radios, but digital, well, in the late 1950's it hadn't progressed to the level of microelectronics and the only transistors I could buy were feeble, noisy little 2N107's that would burn out on the flimsiest pretext. Big, hot, fragile vacuum tubes made computers go, when they went at all.

Remington-Rand had their Univac division in Philadelphia, where I grew up, and I got an after-hours tour of the plant from my instructor at the community center electronics club who was an engineer there. He sat at the console of a Univac I computer and did things I did not understand, but I could see that it was complicated.

In high school my brother, 3 years older, formed the Central High School Computer Club when I was a freshman and he a senior, in 1959. He was interested in software, and got books dating to the 1940's from the public library describing the earliest computers. In these books were cryptic schematic diagrams of electronic circuits, with no specific values of components, which might be helpful for qualified electronic engineers but were nothing that I could use.

“Can you build these?” he would ask, thrusting the diagrams at me. Our relationship was such that it was unthinkable for me to say no. So I said yes, or rather, “I think so”. That was good enough, so he made me Chief Engineer.

With a few friends I tried to build “flip-flop” circuits using vacuum tubes. These were supposed to latch data and count in binary when hooked up right. Unfortunately, our method of creating single, crisply defined pulses – flicking a wire against a metal contact - was one of the few ways guaranteed to generate long strings of random pulses. So our flip-flops worked about 50 percent of the time.

We could have used some professional assistance, and our faculty sponsor should have helped us find it. But apparently computers were so intimidating at the time that the faculty kept silent, and being a lot of frantic overachievers we would never consider doing anything that would make us look ignorant. So we went onward, through the fog.

It's a pity, too. I recently spoke with the son of the legendary John Mauchly, one of the designers of ENIAC and of the UNIVAC I. John lived in Philadelphia at the time, and was a very down-to-earth and helpful person, according to his son. We could have looked up his phone number and called him up, and he would have been willing to help or find someone to help us. Alas, so strong was our instinct to show how “smart” we were that this possibility never entered our minds. One of our teachers could have provided the necessary encouragement, but it was not to be.

I emerged from that experience having sworn off computers and all things digital. One little error didn't count for much in analog electronics, being nothing but a little noise, but in digital it could bring the whole edifice to a crashing halt. Ten years later I had no choice than to embrace it.

One day in 1970 at Ampex we were locked out of the cafeteria at lunchtime - it had been commandeered for an emergency stockholders' meeting. The rumors began circulating – the banks were now running the company, and There Would Be Layoffs. In the fall the first round hit, spreading fear throughout the company. As a $6.00-per-hour Junior Engineer I shouldn't have been too concerned, as they wouldn't have saved much money by canning me. Still, there was a task to be done and no one left to do it but me, and it involved my writing software.

I had to put together a test-bed implementation of the basic elements of our newest “Pyramid” system and write code to make it perform. The hardware wasn't so bad – an array of 12-key keypads, some headphones and some pieces of the demo room system like master tape units and buffer tape units, with cables to interconnect them. Sitting at the center of the web, like a spider, was the Data General Nova 1200 minicomputer. And I had to tell it what to do, and how.

I dug through the manuals and found out about the rudiments of programming this beast. Access to it was through a teletype Model 33 ASR teleprinter, with paper tape punch and reader (the “A” of ASR). You typed and the punch recorded what you typed on paper tape. When you read that tape through the reader it would either be fed into the computer or to the printer, depending upon a switch setting. With the switch in the “remote” position the printer was controlled by signals from the computer.

The first exercise was to get the printer to type one letter. They used “self-modifying code”, which meant that the program section (called a subroutine) was written to command the printing of a character defined by a byte in a particular place – for instance, six locations from the starting point of the subroutine. As written, there was nothing but a string of zeroes (the “NUL” character) in that location, so not much happened when the subroutine was activated.

However, this subroutine was embedded in another subroutine that caused the desired character byte code (passed to that subroutine by others) to be written down the line into the memory location otherwise occupied by the NUL character. This is not considered good programming practice these days – it's not smart to go around overwriting code – you might change something in the wrong place and then you crash.

With my heart in my mouth I ran the paper tapes through the assembler program, converting the mnemonic codes and data bytes into the binary bits that actually worked the computer. Then I pushed the “RUN” switch. The intended character was an”A” (20 hexadecimal, or 0010 0000 in binary).

“CHUNK!” went the Teletype and an “A” appeared on the paper, with the print head obediently moving one space to the right. My God - it had worked! It was a transcendental experience – I recalled the words to a satirical song I had heard about the Pope - “..he don't even have to take dope”. And there was no operating system, no interpreter, no layer upon layer of software to interfere with it. I had written code in simple symbols, assembled it into binary code which I could, with enough effort, understand, and made the computer run that code directly. And it had worked!

With that as reinforcement I forged ahead writing an increasingly elaborate program to run the batch of components I had wired together. Eventually I could do a simple demonstration, using the keypad to order up a program from the master tape to be copied off to the buffer tape unit for that keyboard's position, then to be played back to the headphones from the buffer unit.

I heard that there was to be another round of layoffs. Some time around the year-end holidays, after I had worked over a weekend, the head of the division came back to see what I had to show, and I showed him. He muttered “I'm impressed” and slunk back to his office. He looked back at me and repeated “I am... impressed”, and I had reason to hope that I would survive the layoff. I did.

Thus I crossed the threshold to the digital world. It took ten years, within which computers had shrunk from roomful to table-top size, a generation of transistor circuits had shrunk into digital chips and a computer could be supported and operated by one person. But my transcendental experience was probably no different from that experienced by generations of my predecessors, and hopefully by generations of successors. If, however, it's gotten so easy that it's no thrill at all I'll do something to let people experience it the way I did.

Small Valley

Small Valley

23 Sept. 2009

It's 1970 and I'm a 25-year-old Junior Engineer (no degree, a Berkeley dropout for now) at the Special Products Division of Ampex Corporation. I have a desk out on the floor of an open area used for staging systems and equipment – the “bull pen”. I earn a little over $6 and hour and I wear a suit and tie to work.

My responsibilities are “maintenance of product design” for the computer interface electronics of the “Pyramid” system – a “programmed instruction” system meant to be installed in a university, that delivers illustrated lectures on demand to students and which can respond to students' keystrokes on a touch-tone type keyboard.

Each lecture is actually recorded on 8-track tape, and the tape can stop when a question is asked, wait for a response, and move to another place on another track to deliver an appropriate response, most of which end up back at the original question. If the right answer was given, the tape continues on its original path. This is known as “branching” and it's supposed to be the coming thing in programmed instruction.

The logic for all these responses is handled by a small computer built to be installed as part of an industrial system. Known as a “minicomputer” this one is the Data General Nova 1200. It's a box that can be mounted in an industry standard 19-inch-wide rack. The outside world connects to it through multi-pin connectors on the back panel.

I own the “I/O” (Input/Output) box that Special Products built. It has a cable going to the computer back panel and contains a number of circuit boards performing various functions – some receive data from the keypads, some send out control signals to the tape drives, and some – the tone detectors, listen to audio playing to the student and decode bursts of 55-Hertz (cycle per second) waves added into the audio stream when the tape was mastered. These bursts - using a method similar to Morse code - cue the computer that an event is ready to happen and tells it what should be done. So when the tape stops after a question is asked, that's how the computer knew when to stop and what responses to give.

The design of the tone detector was beyond my capabilities, using magic known as “active filters” to selectively amplify narrow bands of frequencies, but I only have to deal with circumstances in which it doesn't work and fix it. Soon enough one such circumstance brought itself to our attention.

The prototype system was installed in a community college in Oak Park, Illinois, and the tone detectors were misbehaving. An engineer named Jim Ferguson was sent out to find a fix, and he returned with a tone detector board to which a piece of cardboard had been added as an extension. On this cardboard were two adjustable resistors (“potentiometers” or “pots” in the terminology) and some other components connected to the board circuitry to allow the adjustment of certain delay timings.

Ferguson, quite full of himself, was telling everyone repeatedly how he had adjusted the board operation just so. I was less delighted, since the specifications for the board called for no such adjustments, but I was instructed by management to incorporate the new adjustable pots in a new version of the design, and so I did.

When we tried the new board out at our development system it didn't work. Now the problem was mine. I finally found the cause of the problem after what seemed to be weeks of frantic effort (but was probably only a few days), by using an old paper chart-recording oscillograph – a device which has been totally rendered obsolete by digital oscilloscopes now.

On the chart I could demonstrate that when the burst of 55-Hz tone began something was shocked into a kind of oscillation and the audio signal began sloshing up and down at a very slow rate – about 1 Hz. You couldn't just define a signal level for the 55-Hz bursts at which the comparator said “this must be a filtered and amplified signal at 55 Hz, so I will start a digital pulse marking the event”. The reference level used for comparison was bouncing up and down at the 1 Hz rate.

Further investigation revealed the culprit – the master tape recorder on which the program materials were created. Each system had exactly one of these – a studio-quality model AG-440 tape recorder from Ampex's distinguished studio product line. Unfortunately, Oak Park had a different model number, which had different response characteristics from ours.

When asked about the anomalous performance of the master recorder an audio engineer would say “Who cares what the recorder is doing at those low frequencies – they'll never get through the channel and do any harm!” (High fidelity audio was considered to end at 20 Hz in the low direction) And this sloshing behavior to the equivalent of incredibly loud, incredibly low-frequency disturbance was no doubt the byproduct of some clever engineering that was quite effective at higher frequencies. “Bump and thump” percussion effects were still decades away.

I knew that, unless we were going to tune the adjustment of the tone detectors to the specific master recorder used at the site of each system we could never get it working reliably. And what would happen when a different master recorder was used some time in the future – resulting in a mix of programs with different baseline shift characteristics!

I reached the conclusion that we had to do something in the tone detector that would straighten out the sloshing effect and adjust for inevitable variations in level that could be expected in the future. This meant AGC – Automatic Gain Control. I set about trying to modify the design to include this feature.

For several weeks I continued to develop the AGC solution, but support was not forthcoming from management. I would stand up in the weekly review meeting and tell how I was still trying to finish the AGC version and would be told – in front of all the engineers present - “never mind about that – freeze the design the way it is.” After two such episodes I relented, and three other Pyramid systems were shipped with tone detectors that were invitations to trouble.

Two years later, after having returned to Berkeley and finishing my degree, I returned to check in with the guys who were left (Ampex was undergoing a long and painful contraction phase, which has lasted to the present day) and see what the outcome had been. During my last few weeks on the job the Pyramid system had been passed over to something called the Videofile Division, who had been trying to use videotape recorders to store and retrieve huge amounts of digital data.

My former colleagues told me that the solution to the tone detector problems had been to add AGC. I looked at the schematic diagrams to better understand what had been done and saw the name of the responsible engineer – one Al Alcorn. I had never heard of him.

Two years after that I found myself standing in front of Al Alcorn's desk asking for a job at his new company, Atari, of which he was Vice President for Engineering. He had gone with a fellow Videofile engineer named Nolan Bushnell and the two had founded what became the first coin-op video game company, with a winning product called Pong, designed by Alcorn.

I tried hard to explain why the tone detector debacle hadn't been my fault, but Al was not interested. He was looking for manufacturing engineers, not designers like me. I didn't get a job at Atari. The young man who had pre-screened me and escorted me into Al's office, wearing a white shirt and tie, turned out to be Steve Jobs before his India sojourn. I think I left on my own.

In retrospect I don't see what I could have done differently, other than trying to build political support within Special Products for my AGC conclusion. But short-term thinking ruled there as it does in many other places in Silicon Valley, a valley that's a lot smaller than it seemed. “Don't worry about that”, indeed!

What Engineers Sometimes Must Do

The year was 1970, and I was a Junior Engineer at the Special Products Division of Ampex Corporation in Redwood City California. I had applied for a job as a technician after dropping out of Berkeley in a depressive crash, but with my work experience while in school I was routed to a job that was sort of a combination of technician and engineer. 

Special Products, as its name implies, was the place where oddball jobs came to be done. We had to be able to design and build almost anything, and the division was tiny by Ampex standards. I learned the craft of engineering there, and consider myself quite lucky to have wound up there. 

In 1970 we were building the "Pyramid System" - a contrived acronym referring to a system for delivering illustrated audio lectures on a college campus and allowing the lectures to take various branches in response to the students' responses on a numeric keypad. Today it could be handled by a DVD player with a keyboard, but then it required racks ands racks of equipment installed in a basement room and wired to multiple locations around campus. 

A system like that had to be controlled by a computer, and we were using the Data General Nova 1200 - a box with maybe 16K of 32-bit magnetic core memory, a panel with switches and lights and a number of connectors on the back. One of those connectors was the "I/O bus" - the means by which external electronic equipment could be placed under control of the computer. The architecture of the Nova I now consider to be rather inferior for reasons I won't go into here - back then I knew no better. I was placed in charge of "maintaining" the design of the external interface electronics - a job no one else wanted.

We had a demo room in our building - the largest conference room available in which the system was being staged. Our software guys worked in there putting together the machine code that would make everything run. They used a Teletype Model 33 as a terminal, connected by another plug to the back of the Nova. 

When prospective customers came to visit they were first received in a front conference room and told the wonders they would behold. Then our engineering manager, Maynard Kuljian, would conduct them back through the doorway labeled "Engineering" to the demo room. Far too often the system would crash as soon as they entered the room. I remember Ray, one of the software engineers, softly crying out "Wha' hoppen'?" as the disaster struck, as always at the worst possible time. Maynard would have to take the visitors back out and entertain them for another half hour while Ray and others scrambled to re-load the software. I would retreat to my desk out in the bull pen and try to stay busy.

Eventually Maynard made a hard decision. He would not go back to the demo room with the clients, but would send them in someone else's charge. Somehow, he figured, he was the cause of the crashes. And apparently it worked! The crashes ceased!

That was engineering at work. In a process far older than science, one takes a guess as to what might work and tries it. The question of "why" is irrelevant (though possibly useful) - the operative question is "how". If the "why" is supernatural, that's fine, as long as the effect is consistent for as long as it's needed. And it was.

For two years I worked on that system, while Ampex began a long, slow process of contraction. Six years later I sought out Maynard and tried to recruit him to be engineering manager of the personal computer company I was helping. He turned me down - he was much happier, he said, working on building products than on the management track.

So when you look at engineers at work, don't think that they are all so very smart. Sometimes they're just foolhardy - and lucky. But you'll never know when that is.

Professional Services Available

Lee Felsenstein

Fonly LLC

2460 Park Blvd. #2

Palo Alto, CA 94306-1917

650-814-0427

lee (at) fonlyinstitute (dot) com

Whether it's fixing a marginal design or making a working one practical to manufacture, creating a system of devices or designing a device that fits into an existing system, or designing an entire product from the ground up, I have been doing it for more than 40 years and can do it for your project.

My design for combining video display circuitry with a computer has propagated so that it is the standard in personal computers today. I designed the first portable computer that was commercially successful. I have redesigned an existing medical device to make it portable, as part of a 4-person team, and managed the circuit board layout to make its sensitive EEG circuits less noisy than the original.

I rescued a company from oblivion by proposing a different technology for its product and was listed on its patent as an inventor. I upgraded piece of pharmaceutical research equipment by adding an embedded computer, display and sensors as well as choosing a mechanical design subcontractor and working with the machine shop to build new parts. I ran the prototype development shop for a research lab and chaired its intellectual property committee.

I know every step of the process of product development, from conceptual design at the marketing level to circuit and mechanical prototyping, circuit board layout, fabrication and assembly and design of plastic enclosures. I have put products through agency approval testing and designed circuit boards to meet UL standards.

Let me look at your requirements – I may be able to ask the questions that help you improve them. Let me review your intellectual property approach – I may be able to help you write a patent that protects a broader range of what needs protecting to cover your future product family. Let me look at your needs and propose an electronic solution – and then implement that solution.

I do:

circuit design

device design

system design

product design

- and bring these designs to the level of working, manufacturable prototypes that put your company on the path to success.

Disciplines:

Analog

Digital

Control

Video

Power supply

Embedded computer

Medical device

PCB layout - signal integrity

Functions performed:

Conceptual design

Prototyping

Packaging and mechanical design

Circuit Capture

Printed Circuit Layout

Design for Manufacture

Reverse Engineering

Machine Language Software

Failure mode analysis

Documentation

Agency Approval

Intellectual Property Strategy

Patent preparation

Patent analysis

Technologies used so far:

Vacuum tube

Transistor

Integrated circuit – analog

Operational Amplifier

Timer

Linear regulator

Switching regulator

High Power Op Amp

Phase Locked Loop

Video Amplifier

Video Sync Extractor

Video Sync Generator

Video Color Encoder

Switched capacitor filter

Laser Driver

LED / photodetetor

Integrated circuit – digital

TTL

LSTTL

STTL

ECL

CMOS

BiCMOS

NMOS

SRAM

DRAM

PLD

FPGA

Digital to Analog Converters

Analog to Digital Converters

Ultrasonics

Magnetics

Laptops or Phones - both, but start with phones!

Here's a suggestion I posted recently on an "Educational Technology Debate" website sponsored by InfoDev and UNESCO. The topic, run by Wayan Vota, was whether laptops or mobile phones offered the best platforms for educational use in developing countries. The discussion pushed me to think in the direction of a synthesis:

I had attempted to post this earlier, but managed to lose my text and ran out of time. It does seem, though, that this is not a bad time to argue for a synthesis of the phone and the laptop. 

I have long argued for using technology that is economically helpful to the family of the educational user, so that it becomes a tool seen as worth supporting. For this purpose the mobile phone would be best, as it has been shown to assist in raising the economic level of users. But for the purposes of education we need to add some of the functions of the laptop. Depending upon the level of education that it should provide, this could be done with as little as the addition of a stylus touchpad. 

I say this because it appears that the initial big application would be basic literacy - both for children and for adults (presumably with their childrens' help). The introduction to the letters of the alphabet (whichever one is used locally), the correspondence of sounds to the characters and combination of letters, the ability to practice drawing letters and to have the program correct or otherwise indicate improvements, all are functions performed by every teacher in every classroom for the first year of education. The advantage here would be that parents could engage in the same lessons and learn literacy where they never would have the time or be willing to suffer the loss of pride to come into a classroom - even if they would be permitted there by school rules. 

It's certainly true that adults learn more slowly and differently than do children, and the programs would have to take this into account. But the basic technology required is pretty much covered by the mobile phone - given a speakerphone mode and a stylus touchpad. Different modules of the programs could be downloaded from the network, eliminating the need for mass storage, and the screen would have to be capable of a certain amount of resolution, but given the ability of the program to speak and to hear pronunciation of letters, such an enhanced phone could provide the start for great increases in basic literacy (and numeracy as well, since numbers and arithmetic can be handled with the same device). 

To create such a "literacy phone" and the necessary programs in a sufficient number of languages and cultural settings would be a big task, but it would produce results much faster than the expansion of numbers of laptops. Later applications could make the same phone an electronic book, which would give it a place in the education process well beyond the basic stages. 

Laptops are indeed necessary for the higher levels of the educational process - but don't forget you have to start simple.

An Idea for Chicago Transit

Through an email list I was introduced to Innocentive, a clearinghouse matching inventors and developers with organizations in need of innovative solutions. Often there is a cash award for the ideas and solutions submitted.  In this case the request was for "Ideas for Increasing Public Transportation Use to Reduce Greenhouse Gases in Chicago".

I had done some thinking about this kind of problem back in 2000, immediately following the closure of Interval Research, and had written wrote 14 pages of scenarios on the subject of using wireless Internet on public rail transportation. I took some of the ideas I had developed then, updated them to take into account the growth and penetration of applications on cellular phones, and turned out the following  ideas for making public transit more usable and attractive.

The deadline for submission has now passed so I am publishing the submission here, since ideas have little value if kept secret, and require a lot of value added before they become reality.

Lee Felsenstein

Introduction and Background

A major inhibition to the use of public transportation is the fact that scheduling is very poorly understood by the rider and that time in transit is generally considered wasted (except possibly for reading). If riders could be kept informed of their projected arrival times at transfer points (with real-time performance taken into consideration), the realistically projected arrival times of connecting routes, and could make arrangements for the disposition of their time between connections, there would be significant incentives to chose public transit over private vehicle use.

The growth of wireless communication creates an opportunity in improving these aspects. Penetration of cellular phones (technically known as “handsets”) has become significant, and the arrival of a new generation of “smart phones”, with the Apple iPhone followed by a number of competitors’s offerings has moved the “cell phone” experience well beyond simple telephone voice messaging.

Detailed Description of the Solution

At the base would be real-time access to the database of schedule information, both for intended schedules and actual progress of runs. On rail sections this information can be updated from the signaling system, but on street transportation there are various other ways to update progress of a run relative to the schedule. Since the author is not familiar with the capabilities of the CTA in extracting actual run performance we will make some general assumptions here about possible solutions.

Ideally there would be a central database of schedules for each run, updated by input from signal or sensor networks with real-time information about actual performance. This database would be accessible on a read-only basis by an external application which is itself accessed from wireless phone providers and from the Internet. This application would have the capability to calculate projected arrival times at any schedule point from the schedule modified by performance information. This information would ideally include projected arrival times for the next several runs at any given stop.

The utility of such a system is obvious for users who are awaiting an arrival, however other modes become interesting when one considers use by riders in transit. The obvious question for a rider is whether a connection will be made, and how much time is available before the connecting arrival. If there is to be more than a brief gap, the user would be helped by information about merchant services available at the transfer point during the gap. This would create a market for local advertising from merchants located within a short distance of the transfer point, and this market could be served by another wireless application accessible from the schedule progress application.

Since such a local merchant application would require a sales and marketing function, it would best be implemented as a private profit center. Merchants would be encouraged by the merchant application operator to create special offers for transit riders, who could place orders or reservations through text messages generated through the advertising application.

In this way the rider’s process of decision making about use of time would no longer be deferred until arrival (with uncertainty as to time limits due to the connection) but would be extended into on-vehicle time and given optional connection limits well in advance.

Another area where the trip coordination information could find a use is in social networking. Riders who will be in the same area at the same times would have something temporarily in common, and might want to arrange face-to-face meetings under differing degrees of anonymity. Providers of social networking application service should be interested in extending their offerings to this new venue.

A limited subset of the schedule and connection service could be made available to riders without wireless service through a built-in publicly-accessible wireless unit running only the relevant application. For all on-board users it will be necessary to make the run number available and to provide a relatively easy method to identify stops. Numbering of stops on the in-vehicle route map or an interactive map accessible from the wireless schedule application are two non-exclusive approaches to this requirement.

We will mention here, without further detail, the possibilities for wireless-based charging for merchandise or services. There are technologies available for displaying specialized bar codes on wireless handset displays which can enable charge transactions to complete sales. At the least, the merchant application can provide confirmation numbers to display on the handset by which riders can pick up merchandise ordered earlier.

Conclusion

The effect of these applications would be to decrease the friction inherent in fitting one’s schedule to those of the transit system. In addition, it would provide the opportunity to extend one’s productive time into trip time, as well as reducing uncertainty and anxiety.  In fact, it should be possible to reverse the equation and make travel by public transit less anxiety-producing than travel by private vehicle.

Seeking Stamp Collectors for OLPC

My brother Joe, who is famous in the world of population genetics, tells how for decades those working in his corner of biology (phylogenetic inference – the science of constructing inheritance trees) were scorned by the reigning molecular biologists as “stamp collectors”. While the molecular biologists pursued the secret of life itself, the stamp collectors puttered around with statistics and large data sets, working out how to make sense out of data patterns.

Then came the crowning triumph of molecular biology – reading the human genome. Note that I do not say “decoding the human genome”, as it suddenly became clear that no one knew how to make sense of gigabytes of gene sequence data. Who, the moleculars wondered, could make some order of all this data?

Then everyone looked at each other and exclaimed in unison, “the stamp collectors!” Joe and his colleagues were showered with money and attention. Their grant requests were now favored for approval, and at Joe’s university a brand-new Department of Genome Sciences was created which welcomed his august presence.

The parallel is this – the OLPC project is about as far as it can go without empowering its own “stamp collectors”, by which I mean those who have long labored in the field of experimental education. Yes, there are others besides Seymour Papert, and the official OLPC line on the topic, that the educational research had already been done and that the engineering was all that was left, was always blatantly untrue.