SoylentNews

TV Review: The Outer Limits episode “Keeper of the Purple Twilight” (1964)

Eric Plummer (Warren Stevens) is a scientist whose research has hit a frustrating roadblock. He needs one last equation to complete his theory of anti-magnetic disintegration, and cannot derive it. He leaves his lab, its steel paneled walls festooned with mid-20th century techno-regalia, and drives home in angry despair, speeding dangerously, and then seems to hear a voice in the back seat of his car warning him not to kill himself. He arrives home to his girlfriend Janet (Gail Kobe) who tries to soothe him with a martini, but Eric is so fraught with confusion, anger and fear that he locks himself in the study to return to work on the equation.

He hears the voice again, looks up from the drafting table, and sees a strange man (Robert Webber) standing in the room. His voice now seething with fury at the interruption, Eric demands that the man tell him how he got in. The man, curiously attired in a bland one piece jumpsuit, identifies himself as Ikar and offers Eric a deal: I’ll give you the equation if you give me your emotions. Bewildered by the impossibility of the situation in which he finds himself, but desperate for the equation, Eric agrees.

Narrow beams of light appear suddenly between the two men’s eyes as tension and then all traces of emotion drain from Eric’s face.

Eric asks Ikar, with an eerie calmness: “When do we begin?”

The camera cuts back to Ikar, whose figure suddenly transforms into a hideously piscine seven foot tall alien with grey skin, lidless eyes, oversized cranium, and a series of gill-like structures where his mouth should be.

“We have already begun.”

So begins the opening scenes of Keeper of the Purple Twilight, easily one of the best episodes of The Outer Limits ever aired. Against the backdrop of an emerging planet wide invasion with Eric as its unwitting accomplice, we are posed with the question— a question often posed in Outer Limits episodes— of what it means to be human.

With the acquisition of the final equation and help from Ikar, Eric quickly delivers a prototype disintegrator for demonstration to his wise and humane department director, played by the avuncular Curt Conway. Eric turns on the machine, which hums and warbles ominously, and is tethered to an unearthly looking rifle which Eric aims at a small table heaping with what looks like bars of metal stacked upon each other. A white ray of light erupts from the nozzle of the alien contraption, surrounding the table in a pulsating glow of energy. The objects on the table become faint and indistinct, and soon only the outline of the table is visible in luminescent white before finally fading and vanishing completely.

The director is horrified at the destructive power of the new invention, and tells Eric that the experiment must not go forward. Ikar suddenly appears in his native form in the office, along with three other armed aliens with noticeably smaller heads (apparently suggesting to the viewer that they are some kind of security detachment). Ikar makes eye contact with the director as he did before with Eric, but this time the director collapses in an inert heap behind the desk. Eric watches the scene dispassionately, and turns to Ikar for further instructions.

The situation seems hopeless. Although the aliens cannot transport their super weapon to Earth for the invasion, they have a robotic Earthly confederate who will build one for them. But what of Eric’s emotions, which are now being hosted by Ikar as part of the original deal?

Watch it for yourself and see.

Jazz piano player: “The man that hath no music in himself...”

Alexander Scott: “Is fit for treasons, strategems, and spoils.”

Most vintage TV aficionados probably don’t know it, but Alexander Scott, played by the future roofie dunking Bill Cosby, was the first prime time Scotty on the tube— even before Montgomery Scott himself. Scotty is the darkly stone faced, soft spoken sidekick to Robert Culp’s Kelly Robinson who together pose as coach and international tennis pro as covers for their true identities as cool— but very high minded— American spies. Kelly still has the moves both on the court and in hand to hand combat, but his face carries the creases and lines of a man 15 years older. Perhaps this is because of Culp’s relentless overacting, replete with ever changing overwrought facial expressions.

For some reason, all of the episodes so far have taken place in Hong Kong, and have the familiar plot contours of heroes follow leads, heroes fall into mortal peril, followed by heroic triumph. It’s unlikely you’ll see as many pistols karate chopped out of the hands of abductors as you’ll see on I Spy, nor will you find protagonists that look upon the vanquished with as much moral condescension (with the possible exception of Adam West as Batman).

Why you would ever for a moment consider picking a high profile tennis player as a cover for an undercover spy is a question that is quickly overcome by the blinding glitter of American exceptionalism. The theme song is top notch, however, and I’ll probably keep watching intermittently on the remote chance that Scotty will quote Shakespeare once more.

Not long after starting at VaxCo, I found myself in the fortunate circumstance of having some disposable income along with the free time typical of a young man with no dependents. I decided that I was going do something I'd wanted to do since I was a boy: learn how to fly an airplane.

One of the big advantages to working at VaxCo was access to an internal bulletin board system known as Notes, that contained, among lots of other things, employee interest groups, and there was a fairly large group called Flying that contained a wealth of first hand information, and whose posts could be readily traced to individual employees. In other words, no anonymous posting.

This was how I found the flight school I wanted to attend, located at a large mixed use airfield that was originally a US Air Force base. It was quite different from any airport I've passed through before or since. There was one long two story terminal-like building that contained a mix of administrative offices along with the control tower, some small commercial airlines with a very small passenger seating area near the entrance, and a handful of flight schools. On the far side of the airfield were a series of large hangars containing a mix of military transport and corporate aircraft. The runways were smooth, generous and wide: about a mile long and 150 feet wide.

The school I decided to join had a sole proprietor and flight instructor named Mike, a short and wiry man with narrow avian features. He was was friendly and relaxed to the point of laid back, except when any sort of discussion in connection with flight safety ensued, at which point his eyes would widen and fix upon yours until he was sure you had understood the message. His “fleet” consisted of a lone Cessna C172 Skyhawk, one of the familiar workhorses of General Aviation of the time, and probably still is to this day.

Mike outlined the rough cost of the training involved in getting a certification as a private pilot, airplane single engine land (ASEL). This was mostly comprised of Mike's hourly rate (his time plus hourly cost of plane rental), in addition to some training materials which were mostly books and charts, a special purpose calculator I didn't really need, and a few other random accessories I would need but didn't cost much. We scheduled the first lesson for the next weekend, and I pretty much read all the material beforehand.

There was nothing in the way of classroom instruction on that first day, and as we walked out to the plane, Mike reviewed what we were going to do, followed by the preflight check. Upon clambering into the left seat of the plane, my first impression was that the instrument panel seemed really high, like almost at the level of my chin, even though I was six foot two inches tall. As a result, a pilot really can't see what is directly in front of plane, thus requiring the traditional warning of yelling “CLEAR!” before starting up the engine.

Mike handled the radio for the first and only time while I focused on taxiing the plane. After about a decade of driving cars, my first stubborn instinct-- even though I knew better from the training material-- was to steer with the yoke in my hands which offered no resistance and had absolutely no effect on the motion of the plane. As in all planes, you must steer with the rudder pedals while on the ground, and even, to a very large degree, while in the air.

Most of those first lessons were dedicated to getting familiar with the control inputs, using them to hold the plane in a particular attitude for a particular maneuver (basically turns, climbs and descents), and controlling airspeed appropriately. I had lazy feet, of which Mike would frequently remind me by exclaiming “Steer with your feet!” on nearly every one of those early final approaches.

One day, en route to a particular nav point, Mike pulled the throttle closed and called out “Pick a field!” to which my immediate response was “What?”. It was an engine out drill, and he wanted me to pick a place to land (while keeping the wind direction in mind), and make an approach to landing. With the propeller windmilling and no longer delivering thrust, the nose became heavy and wanted to pitch down to maintain airspeed. I called out the field and made the turn, letting the nose come down such that the plane was at best glide speed. In the meantime, Mike recited the engine out checklist which is the series of steps and checks you would make in a real engine failure to get it started again. The landscape began to reveal itself in alarming detail before Mike finally pushed the throttle in full and we climbed out again.

Most of the lessons revisited the concepts of attitude, airspeed control, and how the prevailing wind was affecting your movement over the ground below. We visited a different airport pretty much every lesson for touch and gos. On one visit to a medium size airport, Mike had me taxi off the active runway, and stop the plane. He opened his door, climbed out, leaned in from the open doorway and said, “Now you do it yourself-- two touch and gos, and on the third one land and pick me up here.”

Never one to fail to come up with a pithy reply, I said, “Wait, really?”, whereupon he shut the door and walked away. Time for the first solo! So, following ground control, I taxied my way back to takeoff position, got takeoff clearance, and away I went. In mere seconds, however, things began feeling very different.

Because it was high summer, Mike would always pop open his side window, and instruct me to open mine as soon as the plane had made a landing because it would get hot very quickly without the ventilation provided by the motion of the plane through the air. There were no fans for cabin air like in a car, just vents. The windows were hinged at the the top, and you pushed them out at the bottom where there was a latch to open them. Mike had closed his window, but as I gathered speed down the runway, there was a roar in my left ear that wasn't there before. When the wheels came up off the runway, I also noticed that the plane really wanted to yaw left much more than usual-- shit! shit! shit! I left the window open! Nothing to do but hold it in takeoff attitude as usual, except with a LOT more right rudder to cancel out the drag from the open window on the left side.

I got the window closed and latched when I felt I was far enough off the ground to divert some attention from flying the plane, and did pretty well on those next three landings. Back at the flight school office, I was told to bring back my sweat soaked shirt with my own artistic decoration (after washing) added on the back to commemorate the event, where it would be cut out, and adorn the office wall along with the shirt backs from previous students.

It took me a bit less than a year to get the ticket, and in the years that followed, I rented Mike's C172 along with other planes to go on very short day trips. One of the problems with being a lowly rental pilot is that you are competing with all the other rental pilots for time in a plane, almost always on weekends, and only during acceptably good weather. Actually buying a plane, if you had the financial resources, meant incurring a ton of fixed expenses (inspections, tie-down, fuel, insurance) whether you flew or not. So naturally, most people wind up spending even more dough, so you can “get your money's worth”. This always felt wrong to me, kind of like chasing a bad hand in poker.

And last, but not least, I never really felt comfortably at ease flying-- ever. Single engine planes are noisy, cramped, and really don't have great visibility out of the cockpit. I was really expecting something more open and scenic, and what I felt was a sense of confinement and risk. Mike used to always say that the most important instrument in the cockpit was the “Mark One Eyeball”, and that one's head should always be kept on swivel looking for other planes, especially near airports, which quite naturally were airplane magnets. I always kept this in mind, and so just didn't feel like I could ever relax, and never did.

I did have a few wonderful trips with special guests, but gave it up after about four years, and never went back. I have flown many times since then, mostly in the second seat of my friend's Piper Cub, but I never felt even the slightest urge to take the controls when offered to me. I didn't want to be Pilot in Command.

I wanted to look out the window.

Of my original two bosses, the one who held the title of Director of Hardware proved himself to be a most skilled and resilient manager. Though he had been in the telecom business longer than I had, he had come into the corporation at the same time I had via that gala merger near the Dot Com bubble's peak. Though not really a heavy with regard to tech expertise, he was calm, intelligent, and inscrutable. The weekly staff meetings he presided over had gone from catered lunches to brown bags, but we were still intact as a functioning engineering team. The boss never raised his voice, not once, even to those who had committed the cardinal sins of making the same mistake twice, or failing notify him immediately of a problem that could have an impact on schedule. Many engineers are particularly loathe to confess to the latter, because it often means not just a hit to reputation, but possibly extra time at work to fix the problem. Staff members that failed to come clean would be subjected to the Thousand Yard Stare when such gaffes were revealed, often preceded by the question, “How could this have been allowed to happen?”, and then usually followed by a shirt dampeningly uncomfortable silence.

What the boss lacked in charisma, he made up for in political and organizational acumen when it came to navigating the corporate landscape. Where an ordinary manager might appear to be rattling the proverbial tin cup, our boss was able to successfully close deals in which he effectively sold our services to those projects that had funding from HQ. So adept was he hitching our wagon to the next star, that near the end of his tenure he actually landed us for a few years in the research arm of the company, which offered the highest immunity from layoffs of any place in the corporation. That's a story for another time, suffice to say when we were later pushed out of that sacred corporate womb, our days were finally numbered.

Before that final swan song came to its conclusion, however, Anointed Ones continued to beam down from the corporate mother ship, nearly all of them from start ups that had been absorbed in earlier acquisitions. One of these, a VP in the small pond from which he came, was often spoken of in hushed tones by some of the engineering refugees that had worked for him in his former organization. I couldn't tell, though, whether that meant he was revered or feared, since by this time we had had several other executive imports who seemed to function mainly as hatchet men.

And so it came to pass one day that our team was to present a big slide show to Mr. VP which would highlight the overall product concept behind our next undertaking, and hopefully garner his approval. It was scheduled for the deeply unpopular meeting time of 11 am, but everyone understood that in the interest of maximizing professional decorum and efficiency, it was not to be a lunch meeting. Indeed, once we were all seated and waiting for our exalted guest to arrive, there was nary a granola bar in sight at the tense and unhappy conference table. Unlike the seating arrangements you see in a movie or TV drama where the Big Boss sits at the head of the table, we had a space for Mr. VP in the center of the long side of the rectangle, with my own boss directly opposite. Next to my boss was the designated PowerPoint Monkey who would advance the slides on cue from whomever was presenting.

Our discontentment increased when, after waiting ten minutes, Mr. VP's administrator entered the room to inform us that her boss would be an additional 15 minutes late. Nerves frayed and stomachs rumbled. At last he made his appearance and we got underway. In spite of his professorial bearing, he made very few comments, mainly asking questions and offering little more than bland facial expressions in response.

About a half an hour into the presentation, the administrator swooped into the room with a yellow slip of paper between her fingertips. She presented it to Mr. VP, who read it, and then mumbled something akin to oh, yes, I'll have to take that right away. He raised himself from the chair and they departed.

The level of annoyance was palpable. It went even higher when, minutes later, Mr. VP returned with a take-out style container in hand, from which, after seating himself, he produced a sandwich.

And so ends this product development story, not all that different from stories previous.

Normally in these entries I refrain from identifying by their real name those individuals, or even organizations, whose paths crossed mine in days gone by. But this gentleman went by one of the most extraordinarily appropriate names of anyone I've ever known. To be so named as a science teacher is more than too good to be true because it WAS true.

Yes, his name really was Norbert Lux.

Not only did he have the perfect name, every other aspect of his appearance, personality and powerful sense of theater were perfect for the role of Junior High School science teacher.  He taught physics and chemistry, and especially loved the areas of nuclear science where these two subjects intersected. He wore a starched white shirt with a pocket protector heavily laden with markers and pens on his breast, and sported a necktie on most days.  The appearance of a bow tie heralded a live lab demonstration that was sure to be spectacular.  He wore glasses with very thick black plastic frames containing lenses that were even thicker. So large was the diopter strength of the lens and resulting refraction that you actually could not see through the parts of the glass nearest the frames, only through a region around the center that shrank noticeably if you looked from off its viewing axis. One of his eyes was permanently askew, and he would inevitably explain to a new class without hesitation that his handicapped eye was the result of surgery for retinal detachment that didn't go as well as planned. I remember, too, how he described to class the earlier progress of the affliction: like a dark window shade being slowly pulled down a little more each day.

He had the forward leaning stoop of man with bad vision, and moved around the classroom with a bit of a shuffle that seemed biased toward his good eye. But he didn't give the impression of a downtrodden victim. Rather, he had the look of a man constantly on the watch for something interesting that might come into view nearby, and be worthy of closer examination.  Such was his passion for the subjects he taught that important principles written on the blackboard would be excitedly punctuated by bits of flying chalk as he drove home the conceptual point.

A crackling aura of high intensity nerditry would seem an invitation to classroom management disaster, even back then, but Norbert Lux was such a masterful storyteller that he held rapt nearly every student when it came time to tell one of his science related “true stories,” especially the ones about nuclear radiation.

There were tales of Geiger counters, and other sorts of radiation detectors, some with viewfinders which you held up to your eye. If this latter type, he said, ever showed above a certain number of rems, you should get out of the room and get outdoors right away. Drop the detector and leave it behind-- it might be contaminated. But if you got a reading of this-larger-number-of-rems, however, it didn't matter because you would be dead in hours.

And so he was well known for his vast repertoire of all manner of radiation poisoning stories. Of how the pigments used in bright yellow and orange glazes found on the dinnerware of years ago contained uranium, and was subsequently banned because of health concerns. He spoke of ladies on assembly lines who used to paint the dials of wristwatches with glow-in-the-dark radium based paint, and of how it was their custom to dab the tips of the brushes in their mouths to get a finer point to paint with. “I'd had the habit of falling asleep with my head on my wristwatch from time to time, but when I heard that story,” he exclaimed jovially, “I always made sure to take my watch off before falling asleep!”

Most memorable was the story of the Wood River Junction excursion involving the accidental death of a worker who was mixing radioactive materials in a large vat, and was thrown off the ladder by a strange blue flash that seemed to singe his skin. The radiation exposure badge he wore indicated to the arriving medical team that he had received a lethal dose, and he died two days later.  Badly.

He did, of course, have many other fascinating science stories with far less ghastly outcomes, but he never told them with as much gusto. He was far and away my all time favorite science teacher as he doubtless was for many others.

I was overcome with a wave of nostalgia (and envy) recently when a respected SNer made an off hand comment about his 200 in 1 in Experimenter's kit of yore. In the spirit of sharing distant memories of tech-related children's toys that long pre-dated the Internet, I present to you: Doctor Nim.

Doctor Nim is a special purpose gravity powered mechanical state machine that can count to four for the purpose of playing a simple game in which two players begin with a common pile of 15 marbles. Each player takes a turn in which they must remove either one, two or three marbles. The person who takes the last marble loses.

Doctor Nim looks like a red plastic tray with channels in it for marbles that sits inclined on a horizontal surface like a picture frame. Within some of the channels and hollows are white plastic mechanical levers that can rotate into one of two states, similar to a flip-flop. You load the fifteen marbles in the top of the tray, and depress a lever in the bottom right hand corner for each marble you want to take. Then, when playing against Doctor Nim, for him to move you flip a white plastic selector and press the lever once. If Doctor Nim wants more than one marble, the plastic levers steer marbles back over the button as required, finishing by moving the selector back to the position for a human move.

To fully appreciate not just the ingenious design of the game, but also to get a sense of the vibe of the tech toys of the era, you absolutely have to look at the 22 page (!) “How to Play Dr. Nim” manual: http://www.cs.miami.edu/home/burt/learning/Csc427.152/491_Dr-Nim-Manual5b15d.pdf

It begins as a simple how-to manual, continues with a description of different game variations and how to initialize the machine accordingly, and from there gives a step by step explanation of how it works using logic state tables and equations! If that's not enough to satiate an inquiring mind, the manual goes on for another seven pages touching on such philosophical questions as “Can computers think?” and so on. Also not to be missed are the illustrations of the Dr. Nim avatar scattered throughout the manual which would no doubt be frowned upon for its propagation of the stereotypical portrayal of Chinese-- regaled in pre-Communist imperial clothing, wise, bespectacled and smiling behind an appropriately styled beard... and, wait, what's in that pipe he's holding on page 17?

In the months that followed the layoff, the big network switch project was ultimately canceled for reasons described in previous journal entries. What was left of the team, however, still contained some formidable expertise, and had the reputation of its past successes intact. What was needed, according to top corporate management, was the right set of marketing requirements and a new team of architects to go with it. Those architects who had survived the layoff departed of their own volition, and new ones, mostly from other small acquisitions in the networking business, came in. As I suppose is often the case when major organizational changes come remotely from corporate HQ, politics and perception reigned supreme. That is not to say that any of the New Guard were necessarily technically inept, but there was much more of a sense of top down driven decision making than ever before.

One of the new directors who came from a small company that had been bought out within the previous year was young, decisive and energetic. Under his leadership, the next generation product was to move back to its traditional and highly successful single chassis architecture, but without losing too much bandwidth, and definitely without discarding any of the memory intensive features that distinguished ATM from its rival IP/MPLS protocol.

Well, to some extent, hardware engineering has always been an exercise in putting ten pounds of shit into a five pound bag, but the new product specification gave me the impression that they were looking to add another two or three pounds of shit into the same bag. I got deeply involved in the floorplanning of the next generation cards that naturally had to hold more chips than previous generations, but without being able to significantly expand the card outline. One very big concern was power dissipation, but we didn't have all numbers from the chip suppliers at that time, so we kind of glossed over the issue by putting the stuff we thought would run hottest near the bottom of the card edge that would be closest to the fan tray. Just as an aside, in those days the absolute upper limit for a network processor card of about 8 by 14 inches or so was about 350W. Given that this number is governed by junction temperature limits, heatsink size and airflow (these last two trade off against each other) past a card height that is limited to about an inch, I imagine it's the same number today as it was then.

So right off the bat, since there were no chip specs, we made thermal assumptions about what the chip designers were going to deliver, but these weren't too big of a risk since the chip designers have their own set of limits they need to stay within just like system designers do. The problem that I worried about was that to fit all the stuff into the card outline, I had to place the major components so that they were practically touching each other. This would leave no room for smaller components like bypassing, bias networks, clock buffers or heat sink mounting locations. Some passives you could put on the bottom side, but they would still need vias, which are holes, which rob you of routing resources on every layer where they are placed. Putting some chips on side two would have been nice, but there was no component height allowance there because you have to make room for the next card in the rack. I felt that by going along with this philosophy, I was just kicking the can on down the road to the poor board designer and his layout guy who would then find themselves in an impossible situation. Or worse, you somehow fit it all in, maybe by adding board layers (restricting height still further), but find that everything fails in the chamber at the upper temperature limit.

I presented these concerns repeatedly to the new architect, though, of course, I could not quantitatively predict failure like Roger Boisjoly of Morton Thiokol O-ring bravely took it upon himself to do. Coming from me, a guy with an extensive background in a field devoted to making sure that things don't break in circuit designs (signal integrity), that would sound like FUD. Instead, I told him that I considered it to be an extremely risky foundation for a product development effort.

I can still hear his voice in my head: clear, emphatic, and authoritative with a mid-Atlantic American English twang that turned simple vowel sounds into dipthongs: “Ree-usk is goo-ud!” The idea, of course, being that you don't make ground breakingly competitive products by being 100% sure of your margins.

Maybe he was right, but I'll never know because the product was never built.

I had gotten the stack of white 8 ½ x 11 ½ inch envelopes from my organizational boss the previous day, along with a pointer in email to a deck of Powerpoint slides generated by HR. The slides were very explicit as to the procedure for laying off subordinates: what to say (“business conditions are such that we can no longer maintain your position”), what not to say (“I really think those upper management jerks picked the wrong guy”), how do to it (get a conference room with a phone, and call people waiting in their cubes one by one until finished), and other subtle touches (you might want to have a box of Kleenex on the table). At some point, hopefully before lunch, upper management would give the all clear, and remaining employees would know that they survived.

I had six envelopes with a printed label bearing the name of each employee to terminated. They weren't sealed, and we were told to look inside so that we would be prepared to describe the contents of the “package” to the affected. It was most of the usual stuff you'd expect: a long release form that entitled you to a severance check when signed, a package describing the unemployment benefits available from the state, COBRA info (a way of paying for the continuation of employer insurance benefits), and so on, with one interesting document I'd never heard about before. It was a list of positions, with each nameless line representing a laid off employee, sorted by age in the left had column, along with data about title, salary, start date, and so on. It had been explained to me by the boss that it was data intended show that the company had not engaged in age discrimination in selecting employees for dismissal. I'd never seen nor heard of such a thing before or since. Naturally, I took a good look and it seemed to me that the affected were distributed among employees of all job descriptions and pay grades as well as age.

Most of the folks on the list I had inherited when I agreed to lead the PHY team, except for one fellow whom I had hired on to the signal integrity team. He was about my age, somewhat heavy set, and was born in one of the former French colonies of North Africa. He had emigrated to Canada where he was awarded a PhD in some EE specialty that made him seem well suited. He had a head full of dark black curls of hair, thick and short. He had very light brown skin that would quickly break into a sheen of sweat at the merest question of how the job was going in day to day conversation. With a background that had seemed almost exclusively theoretical in nature, he didn't really seem to me like the sort of guy who would be good at “turning the crank”-- running board simulations, which was really what I needed. My bosses, who were also part of the interview team, seemed to like him, however. Maybe that was because they saw the layoff coming and wanted to pack the bench beforehand. I sure hope not, because this poor fellow had been relocated to the area on either an H1B or a green card (can't remember which), and had only been with us for about six months.

Given his anxious nature and my own level of extreme discomfort at having to give him the heave-ho, once I settled into the conference room, I called him first. In spite of everyone being given very explicit instructions to be in their cubes by 9 am, he wasn't there.

I moved on down the list. With the exception of one severely underperforming senior tech, most of these folks were young, with maybe 5 years of experience. They'd all seen it coming. A few of them actually seemed relieved and spent a little bit of extra time chatting with me as they paged through their documents.

I called the first guy again. He still wasn't in. I briefly cruised the office area asking if anyone had seen him. No one had. I called his cell number, and I identified myself in a voicemail telling him to please see me. Nothing.

Now I was actually getting worried. Had this guy had an accident? Had he done something to incapacitate himself or worse? Was he thinking of arming himself and exacting revenge? I duly reported the situation to my boss who had no insight to offer. He said to keep trying, and minutes later gave the “all clear” to the remaining people in the department. The disturbing irony of this proclamation was lost on me at the time. I was jittery and had my own coating of sweat already well developed.

I tried a few more times, calling without leaving a message. On the last call, anxiety and anger rising, I left a message declaring flatly that you are laid off, to please collect your personal effects, and to pick up your package at the front desk. I told my boss what I had done on the phone, and notified him I was leaving for the day without giving him a chance to reply or protest.

I dropped that last envelope to the man at the front desk after explaining to him what it was all about. I drove to the luncheon chain restaurant just outside the office complex, parked myself at the bar, and downed a double Dewar's on the rocks. I ordered a sandwich off the menu, and another round to go with it.

I heard the next day that the package had been picked up without incident later that same afternoon.

In the course of a year after my start, the business unit I had joined at Little Big Telco went from free lunches and profligate hiring, to a hiring freeze that seriously disrupted the big budget development plan, to eventual layoffs. Keep in mind that this took place in an atmosphere of ambitious people leaving for what they thought were better opportunities at startups. The resulting staffing turmoil was huge, and I was presented-- in some ways confronted-- by opportunities to advance my career. The problem for me at the time was that these potential advances all involved significant management responsibility. That being mostly in the form of: “we want you to lead these engineers”.

Before the hiring freeze took effect, but after I'd had time to size up the place in terms of talent and opportunity, there was a list of job postings circulated that offered big referral bonuses to employees. In the list was a job description that was pretty similar to the one I was already doing. It wasn't an exact match, not everything in my field of expertise was present, but enough was, along with a bunch of other stuff that seemed generic. It was also the next pay grade up. Let's say I was very annoyed, and after considering my options, I went into the bosses office and said that if I didn't get that title based on what I was currently doing, I would change my employment status to independent contractor. They saw such an arrangement as one foot out the door and immediately agreed to my demand.

So a few months after that, they asked me to fill in as the manager of a team card designers that handled the line interfaces (then called PHY cards, as in physical layer interface). They had met my demand, so I agreed to meet their need, and accepted the position. It didn't occur to me at the time what unpleasantries were likely to lurk in that role, especially given the sharply growing downward budgetary pressure being applied by the new parent company.

Sure enough, all the group leaders were called into a big conference room one day for an exercise that I refer to as the Lifeboat Drill.

A hapless middle manager was given a dry erase marker and was asked to man the whiteboard. I had really two bosses at the time, one more on the personnel side of things, and the other, one of the principal architects. They sat together in the center of a formation of tables shaped like three sides of a square with the open section facing the whiteboard. Ceremonies began with the recitation of the names of engineers who had earned a top 1 out of 5 rating on their performance reviews. These employees all got seats in the lifeboat, represented by the list on the left hand side. The 1 performers didn't take up all the seats, so there was some discussion, and pretty soon the list was full of 1's and 2's-- about 15 people out of a department of roughly 50 people. Just hardware.

Next list: life preservers. This means people who will probably survive just this one upcoming layoff that was rumored to be very large. The level of contentiousness began to rise, quite naturally as many managers felt that they weren't defending their turf as much as their people, even though these are often really one and the same. One manager recommended another's team member for Third Class. The second manager, a plain spoken, funny but somewhat tense man who was a part time martial arts instructor, stiffened noticeably in his chair, and tersely recounted a list of vital functions and accomplishments by his man.

Somehow we got through it and came up with three lists of people, and total headcount was checked to make sure all were accounted for. Accounted they were, to the exclusion of the people in the room, of course. Next, the formal layoff list would be finalized by local management, after which it was sent to the mother ship and became irrevocable. After that would come the execution.

The organization I was hired into on the eve of Y2K had launched three very successful network switches that were sold to Baby Bells (Verizon, SBC, Bell Atlantic and all the other spinoffs from the AT&T breakup of 1982) in large volumes. In addition to these established customers, marketeers were expecting that a whole new class of customers known as Competitive Local Exchange Carriers (CLECs) would become viable new customers by virtue of the Telecommunications Act of 1996. It was the rush to supply these so-called “greenfield carriers” with equipment (in many cases in the form of credit!) that was to become one of the main causes of the Dot Com Crash-- something which was to come only a few months after I joined.

Another important contributor to the beating taken by my new employer was the decision to commit future products to a network protocol called Asynchronous Transfer Mode (ATM), which was designed to address to the problem of guaranteeing given levels of Quality of Service (QoS) to traffic that could consist of both packet based based data (like email and webpages) and connection based data (that required persistent low latency service, like live voice traffic). Ultimately, the IP protocol everyone uses today was adapted (then called IP/MPLS) to meet these very same requirements, and ATM has become nothing more than a historical footnote. Though my focus has always been on the hardware side (whose architecture had to be optimized for ATM), it's important to understand that the hardware philosophy I'm about to describe was not the only factor in the ultimate demise of the business. ATM along with the unicorn-like prospect of lots of CLEC demand were also major contributing factors.

Previous models of the ATM switches that had been so successful were entirely self-contained rack mountable units, often called shelves, that could go three to a standard telecom rack. Each unit was self-contained in the sense that a single shelf contained all the functions needed to be a fully operating switch: control, I/O, and a switch fabric that occupied the middle slot of the chassis (note: I'm only going to talk about non-redundant architectures here). The basic idea is that a given packet comes into an I/O port on one I/O card, is sent through the switch fabric, itself under the control of a network processor, and then out to the destination I/O port. The new product was to expand the overall bandwidth and port count by partitioning these functions among different shelf types: one for the fabric, one for control, and many I/O shelves with configurable port types that would connect to these first two shelf types like spokes from a wheel to its hub.

The term switch fabric gets its same from the resemblance of its topology to the fabric used in cheap clothing. For a 32 port switch fabric, imagine an array created by 32 horizontal threads overlaid on 32 vertical threads. It may even be the case that this concept of a thread, today commonly used in many other contexts, owes its existence to this idea. In a switch fabric, any of the 32 x 32 places where a vertical thread touches a horizontal thread can be connected, with only 32 one-to-one connections being made at a time under the control of the network processor. Thus, any port on the horizontal side can be routed to any port on the vertical side. Of course, network architects have to deal with problem of a particular route not being available right away to a particular packet, or you may have instances where many packets on the input side want to get to the same output port. All of these problems fall into the study of traffic management, and are dealt with in part by having memory distributed across the input ports that serve as packet buffer queues. If the congestion gets bad enough, most switches have a signaling mechanism called backpressure that tells the sending switch (that could be anywhere) to stop sending packets destined for that port.

Up until this point in time, a parameter known as port density, which you can think of as being the number of physical ports per shelf surface area, was a primary customer concern along with overall bandwidth. In the growth focused optimism of the Dot Com Salad Days, however, scalability became a major selling point because it would allow network operators the ability to simply add an I/O shelf along with a switch fabric card in the switch fabric (also called “core”) shelf. The problem with this idea, though, was that you still had to buy that entire core shelf, which took up roughly two thirds of a rack in one fell swoop, whether you were going to use all of the bandwidth or only a small fraction of it. Also, even if you utilized all the bandwidth in the core, it had an effective port density of zero. All of its interfaces were proprietary, and went not to the outside network, but only to other shelves using a bundle of multimode optical fiber with a VCSEL based set of optical based transceivers at each end.

When the first protos came in, I was impressed at their size and weight. To provide the connectivity required of a 32 port switch fabric located across 32 different cards in the chassis, the backplane required 44 layers, about twice the layer count of anything I'd seen before, and was almost a third of an inch thick. About half of those layers were plane layers for ground or power, interleaved between signal layers to isolate them from each other, a somewhat higher proportion than you would find in a typical backplane because routes could not always be neatly segregated into X and Y directions. The other contributing factor to the layer count inflation was the sheer number of connections, and the fact that because the 180 nm silicon that made up the fabric could not support clock recovery, every data thread had to have its own clock, which gobbled up routing area. A thick board also creates layer registration problems, and also limits the minimum diameter of holes you can drill for vias (structures that let you connect from one layer to another). And last but not least, a distributed switch fabric on this scale is a real routing problem if you are not allowed to jump between routing layers because of the signal bandwidth. To get a very rough (actually somewhat pessimistic) idea of the problem, take a sheet of paper and draw 32 dots. Now draw lines connecting every dot to every other dot without crossing any lines over each other. When you run out of paths, use a different color that's allowed to cross any other line that doesn't belong to its own color group. How many different colors will you need? This is one of the topological considerations that drove signal layer count for this product.

When the first proto came into the lab, in a genuine display of both awe and naivete, I remarked to my boss, who happened to be one of the principal architects, “Wow! What a behemoth!”. He was visibly put off by the remark in a way I'd never seen before, a harbinger of unpleasantries to come.