The Idea Factory: A Review

It is frequently said that history is written by the victors. This is a history written by the defeated. It is a hagiography of a select few who were at Bell Laboratories (Also Bell Telephone Laboratories, BTL). It also is a paean for the return of structures like Bell Labs despite the overwhelming evidence of the technical and economic success of the entrepreneurial and “creative destruction” mode which has been a dominant driver of our economy since the breakup of AT&T, and even slightly before.

This is a review of The Idea Factory by Gertner. It is placed here rather than on Amazon due to the negative attacks.

Let me commence this review with some bona fides. From 1964 through 1972 I spent a considerable amount of time back and forth from MIT and Bell Labs. I worked in Electronic Switching and in the Defense side; I worked at West Street, just before it closed, then Holmdel, then Indian Hill and finally Whippany. I saw many sides of the business as an MTS during that period. Then again from 1986 through 1990 I became Head of R&D at NYNEX, now Verizon, and thus had to manage parts of Bellcore, the residue of some of Bell Labs. In between I have run CATV companies, Satellite companies, fiber companies, wireless companies, here in the US and in over 20 other countries. I have written over 125 papers and fourteen books, many relating to telecom. Thus my approach to this book brings the “distinct disadvantage of experience”, unlike the author, who appears to have none in this space.

Now as to the many principals in this book I have known some personally. John Pierce taught a semester seminar at MIT, I believe the spring of 1971 when I was finishing my degrees, and I was strongly advised to go. I did and it was an hour of hubris each and every time. At that time the world was already changing, I had worked on Apollo and was essential to its guidance system; I had even worked my first start up the end of 1969, albeit as a consultant yet I saw the future. Pierce was stuck in the past, and the longer he spoke the more I saw that. Big monopoly R&D was still his mantra while we were seeing small startups in the back yard buildings around MIT, companies like DEC. I would see Shannon, a ghost like person, in Building 26 from time to time. He was a total contrast to Wiener, who wandered around with his cigar, listening and speaking to whoever would listen. I never heard Shannon utter a word, and I was teaching Information Theory at the time.

I believe that there is a story in Bell Labs, but that this book grossly fails to tell it. The story is one of a monopoly controlled entity whose actions actually set our technology back, perhaps a decade. Add this to the NASA boom in the 1960s and you had a perfect storm which diverted American entrepreneurialism back for a lost decade, thus it led to the Japanese bubble in the 1980s, but by 1990 the US caught fire again. That is a story worth examining.

Page by Page Comments

The following are some page by page comments and then I will provide some overall observations.

p. 3 The author starts by contemplating how ideas begin. He uses wireless communications and computer networks, and he then uses these two areas as exemplars of what Bell Labs produced that our present and future depend. Unfortunately that is not really how it worked. Let me begin with computer networking. That began with ARPA, first Larry Roberts and then Bob Kahn. Add Vint Cerf, Irwin Jacobs, Kleinrock, and yes even yours truly, I did the satellite connections, and we have the beginning of a new world. But what of BTL? As told to me by one of the principals at ARPA, they went to Murray Hill to seek support to use a Bell modem to test out the ARPA net. A couple of folks from ARPA enter a room filled with BTL executives. Before they could talk about their needs, the BTL Execs told them that if they wanted anything then just give BL an open ended contract, they will decide what is needed, develop it, operate it, and they, ARPA may then determine whatever they want with the data. Fortunately ARPA folks did not comply, the thanked their guests, and went out and assembled a team that in a highly entrepreneurial manner developed the Internet. An Internet that did away with AT&T. It was hubris that led to their ultimate collapse. It was the hubris of the monopoly, of the very structure that the author praises, never asking to see the other side.

p. 5 The author denigrates the entrepreneur by lauding the colossus of BTL. The very nature of technology development is the continual “creative destruction” that is required, the change of the past with the replacement of the new. The typical example is when BTL developed and implemented the Electronic Switching System, ESS No 1, it was a computer controlling massive amounts of mechanical switches, fereeds. Fred Kappel was the CEO of AT&T at the time and he wanted to see this progress to a fully electronic system. Fred tried to get the Labs to develop such a system but the rigidity of the Labs management said that he would get it in due time. So Fred went to Canada. You see ATT owned a large interest in Bell Canada, and they owned Bell Northern Research. He then asked them on the sly to develop an all-digital switch. They did, the first one was sold to New York Telephone and from that came the competitor to Western Electric, Northern Telecom. It was my way or the highway at BTL, even to ATT. Yet one person at ATT said no, and sought a “creative destructive” manner to solve th problem.

p 81 The statement that “it would become a kind of received wisdom that many of the revolutionary technologies …” One must ask by whom was this wisdom received. There were other voices, many other voices. In fact in the 60s there was an explosion of other voices, in many ways driven by the Cold War demands for innovative technologies. Places like MIT Lincoln Labs, developing technology well beyond what Bell Was doing. Silicon Valley was just starting, large Defense contractors were seen many offshoot startups, like Linkabit, STI, and others which were pushing telecom well beyond limits at BTL. Satellite companies like COMSAT had developed TDMA technology, and even developed spread spectrum for use in enhanced telephony, key to all mobile systems today. ARPA was doing packet radio. There was an explosion of technology, much from outside BTL. In contrast when one of my Czech partners, the former Communist head of the Czech PTT, Moscow educated, told me they use Bell System Technical Journal designs to try to reverse engineer systems until they found out Erickson was much farther ahead!

p. 118 The author again and again brings out the mid-west origin of many of the people. This stylistic approach of using the same facts over and over can and does become a distraction.

pp 122-123 Again we have a mid-west hero. But this time it is BTLs adventures in radar. In reality there was a battle between MIT, BTL, and MIT won with the Rad Lab. However BTL learned a lesson and tried to be certain this would not happen again. At the bottom of p 123 the author alludes to Shannon and his work on fire control systems. In reality there were many independent contributors to the enhanced fire control in WW II< and one of course was Norbert Wiener and his use of random processes with control theory. Wiener was a true mathematician, Shannon was at best an applied mathematician. Wiener’s problem was he all too often made certain the answer was acceptable to the best of the best, Shannon gave simple and highly readable descriptions available to the many.

p. 128-129 The classic Shannon papers were so successful not only because of the insight but more importantly because of the presentation and simplicity of paradigms. Before his peers one had complex pure mathematics or hand waving engineering. What Shannon did brilliantly was two things. The block diagram in the book and the one missing, the binary symmetric channel, BSC. From the BSC came innumerable advance in information theory and communications. But Shannon also took the reader through the problem as if it were a simple novel, it was not a challenge until you got to the end. It clearly was one of the most brilliantly written papers ever, beyond its ideas. It made the reader comfortable, you could sit down and become engaged.

p. 130-131 The problem with many non-technical authors is how best to explain Shannon’s results. Simply they were threefold; first, any signal can be defined in terms of some measure call information. Second, knowing the information at a source, then one can also determine the information transferring capacity of a channel, the stuff between source and sink, and this delimits the rate at which you can transfer the information. Third, this is often overlooked, the source coding theorem, which states that you can compress the stuff that you send. The author did not discuss this.

p. 153-154 The author discusses Kelly College and the Techs. Kelly College was a way to educate the people inside. However one thing I could never find was how a Number 1 Crossbar worked. I could do that when speaking with craft folks but not at Kelly College. I did get queuing theory, but from a BTL perspective, namely looking at trunks and overload trunks. Kelly College turned into a Master’s Degree program first at local Universities and then at major Institutions. We had a large group at MIT. Now for Techs, they were good, they had been trained typically in the military and could advance to AMTS, Associate Members of the Technical Staff. But there was a glass ceiling. If you got your degrees after starting at BTL you were almost always never going to become MTS. I always found that disturbing.

pp 184-185 The author’s reference to the Organization Man is quite on point. BTL was a hierarchical system, it made the Catholic Church look like an interloper. Everyone knew their position and if you had a godfather you could make it, otherwise you became a “40 year man” In addition you had no ability to move about unless they asked you to do it. The entrepreneur was anathema. You were told what to do and you did it their way. Kelly clearly was a top down manager. The tale the author tells of his public termination of supervisors after WW II, apparently lauded by the author, is one of the most demeaning and demoralizing acts I have ever seen. You just do not do that, it is not gentlemanly.

p 191 Pierce is introduced with the accent on his being an only child. That characteristic is brought out several more times. It would have been useful to have explored some of these issues: poor families, only child, mid-west, and on and on. In my observations a strong New Yorker would most likely not fit well.

pp 216-219 This is a long discussion regarding Pierce and the satellite. It appears that Pierce was enamored with the technology and not the system, service, and economics. As I will show later he decided on a design to provide excellent voice quality, albeit economically infeasible.

p. 225 This is classic Pierce and accepted by the author as gospel. The author states:

“He (Pierce) doubted satellite engineering could advance at the same rate without Bell Labs …”

That was in reference to the establishment of COMSAT, even though ATT owned some 25-40% of the new entity. The author fails to seek the veracity of such a statement. For by this time NASA existed and NRL and other Government entities have extensive satellite experience. In fact the experience well exceeded anything within BTL. This approach by the author again and again accepting statement at face value is a true and basic failing of the book.

p. 250 The author by this point is becoming annoyingly repetitive. We have heard multiple versions of Shannon stories or Shockley tales.

p 252 The development of the integrated circuits was a massive loss for BTL. It was done by entrepreneurs outside the Bell System. The author’s comment by Ross that “it could never be reliable” is again a reiteration of the not invented here syndrome. But by the time this happened the very foundations of BTL were crumbling. DoD was demanding more advanced systems, technology, and the industry and more importantly the people, were finding new ways to deliver this in a more efficient and cost effective manner. Not the Labs way.

pp 302-303 The discussion about the Drucker paths is worth some note. The alternative of going into business for itself and making money from its patents what in reality is what companies like Qualcomm actually did in the very same time period. When at NYNEX I was the instigator in our venture fund investment in Qualcomm, for after all Irwin Jacobs had been my faculty advisor at MIT.

Specific Observations

Let me make a few detailed observations using the words of people who were there. I choose two, Jack Harrington who was my boss at COMSAT and Paul Rosen who was my boss at MIT Lincoln Labs. I find their observations regarding Bell Labs on point and reflective of some of the serious issues the author failed to consider.

Let me begin by looking at Pierce and his satellite program. He had the big balloon, a jump the shark act, to put BTL out in front. It was good PR but poor technology. Out of that came COMSAT, I was there for six years, and that became the focus for satellite and other communications technologies. COMSAT was set up by law in 1962 and was incorporated with an IPO in early 1963. But they raised some $500 million in anticipation of having to build the satellite system proposed and designed by Pierce and the BTL team. Yet that was a low orbit system requiring tracking antennas and dozens of satellites with many launches to keep them up there. It was the most expensive design ever considered. In 1961 Pierce had written a detailed article in Scientific American detailing his design. He concluded:

The cost of a global satellite communication system will be large-on the order of $500 million-but no larger than the cost of undersea cables that could provide the same geographical coverage. (One cannot say the "same service," because undersea cables do not yet provide the band width needed for television.) Assuming that the experiments of the next few years are encouraging, there should be no lack of capital, domestic and foreign, to share the cost of a world-wide satellite communication system.

Pierce’s proposal was:

A system of 40 satellites in polar orbits and 15 in equatorial orbits would provide service 99.9 per cent of the time between any two points on earth. A.T.&T. has proposed that the system contain about 25 ground stations so placed as to provide global coverage. Bell Telephone Laboratories is building an experimental satellite to test the feasibility of such a system; NASA has agreed to launch it early next year for a fee of about $6 million, to be paid by A.T.&T.

The technical problems were tremendous, many of which I suspect Pierce had no clue about. But just after COMSAT raises the necessary $500 million Harold Rosen and his Hughes team launch Syncoms 1 and 2, synchronous satellites (February and July 1063). Much less expensive and much more effective. They just need simple antennas and they only require 3 for global coverage.

From oral histories at COMSAT Jack Harrington, a former MIT Professor but head of Engineering at COMSAT when I got there states[1]:

NGS: Although just the fact that you knew that you could in fact launch the satellite, put it into geosynchronous orbit, didn't necessarily mean that it would be a commercially viable system because of the time delay.

JH: Yeah, there were time delay effects which under certain circumstances can be troublesome. The Bell people [Bell Telephone Laboratories ] made a big issue of that.

NGS: The "Bell people" meaning John Pierce?

JH: Yeah . The Bell Labs people in general and the AT&T people in general. I wouldn't particularize it, but John Pierce was certainly one of the strong spokesmen of the time delay effects on satellite communications.

NGS: Now, do you feel that there was pressure on the part of the engineering staff by the people at AT&T? Obviously, who had TELSTAR, had had success with TELSTAR.

JH: I think it was clear that if one left it up to the Bell System to specify the nature of the satellite system, the specification would have gone to a medium-altitude satellite, and the reason was to control the delay. Which is annoying, no question. It's very annoying.

Pierce it appears designed the satellite system to meet delay issues and not cost. This was a pandemic issue at Bell. Cost and any tradeoff were thrown out the window if the “best” in their sole opinion could be delivered.

Now to the issue of computer networking. In an oral history with Paul Rosen, one of the MIT Rad Lab team, he states[2]:

At the time Bell Labs had great men working on this, and they said if one drives a telephone line over 600 bits per second, one has arrived at the end of the flat earth, and you've had it. With my simple device I did 1800 bits per second consistently. That was a big deal, because they were going to install thousands of miles of transmission. They (the Bell Labs people) adopted my modem – initially. Jack Harrington and I got a patent on it. I did the grunt work and Jack did the analysis post facto that it would work. One of my colleagues said, "Oh, you're going to get rich on this." We had an incompetent attorney at MIT. You have been with academics. I think you may know that some the administrative people associated with academic institutions believe that they have tenure until ten years after they die. This attorney was terrible. The problem was that he let me write a patent application that was much too constrained.

Much, much broader claims. The result was that when this system ultimately went into the field indeed my design was the one that Western Electric (under Bell Labs aegis) built for its early stuff. There was an Air Force contracts officer who thought that Harrington and I should have gotten something out of it. We got point-triple-O nothing because no one would buy the patent and it belonged to the government. The government gave it to Western Electric and that was the end of it. In the meantime, having worked at these sites first on the components that made up these machines, and second on these machines and third on installing them at sites we began to make a system. I became the leader of the group that built and installed these pieces of equipment at the various sites.

Now what is most startling about this is that Rosen did this in 1945! BTL was still only providing 300bps modems as late as 1980. It was this very 300 bps modem that the ARPA folks wanted to upgrade gratis and that they already knew could and had been done. Where BTL had competent patent attorneys Rosen is quite clear as how MIT just lost a tremendous opportunity. Yet it would be again in the early 1970s when ARPA folks, many from MIT, went to BTL again, they in their normal arrogant manner dismissed them, and the ARPA folks took their demand to people like Linkabit, Codex and others, and thus we had a Hayes modem of the early 80s, and now we have modems of 10s of Gbps on silicon chips. Creative destruction at work, and a missed opportunity for the author to have grasped an important observation and made it worthwhile.

Overall Observations

Let me now make some overall observations than that I have delineated some of my concerns as well as having provided a background of facts which counter the overall theme.

The Message is a Hagiography: This is one continual praise after another for a small portion of what was BTL. There were at times upwards of 15,000 people at the Labs and many just worked away in the tombs of technology. They were buried in routine and endless tasks and creativity was often stifled. I am reminded of an office mate of mine for a shot while at Whippany. I suspect he was mid to late 40s and had  PhD in mechanical engineering. For twenty plus years he had been with the underwater cable group, an area praised by the author. He would come in at 9 promptly each day, open his door, then go down get coffee, read the Wall Street Journal, call his broker, then do some work, then go to lunch at 11:30, then go off to a Bell Labs club of some sort till 1:30, return with another cup of coffee, do some work, the call his broker, then read the Journal again, then leave at 5 PM on the dot! He did this every day! It was like a prison sentence, nothing new, and nothing different. For he was the stress-strain expert on the underwater cable and thus he was good for the next twenty years, he was what we called a “forty year man”. He was the guy who had done X and X had a 40 year life in the Bell System so he was stuck with it forever. I saw hundreds of grey 40 year men! They do not come out in this book.

The Style is Overly Repetitive: As I have noted above the author tells the same tale a dozen different times and a dozen different ways. He tells and retells Kelly, Shannon etc.

It is a Paean to Pierce: One must ask why so much on Pierce. His claim to fame was an analysis of the Travelling Wave Tube, a microwave amplifier. Nice job but sooner or later some PhD student would have done the same thing. Pierce was abrasive, arrogant, and frequently in my opinion wrong. One needs look no further than the satellite tale. Shannon was elusive, Pierce arrogant, and Kelly abusive. Nice group of folks to hang around with.

It Fails to Tell the Other Side: The Rad Lab at MIT made seminal developments in radar and communications. The Labs, according to many writers, always was trying to shut them down and take over the work. One needs look no farther than Norbert Wiener and his feedback systems for fire control. The Labs had invited Wiener solely to pick his brains and then to do an end around run. Fortunately Wiener was smart enough. The other side would be to tell what MIT Rad Lab had done. One needs look no further than the classic Rad Lab Series of books, each of which must have started a dozen new companies over time. I use the Servomechanism Rad Lab book to design Apollo guidance systems, and Antennas for designing satellite systems. There were no such products from Bell. Take Shockley’s book on semiconductors. I tried it a few times but there were much better documents from those outside of BTL.

BTL Was Set Up to Preserve the Monopoly: The Kingsbury Decision made ATT a legal and untouched monopoly. In fact it was written into the Antitrust laws and thus it could do fairly whatever it wanted to and the Labs was a means to preserve that bulwark against any intruder. Their portfolio of patents was a protective measure, less a technological platform to deliver the best available. It was a bulwark against competitors and innovation. The author never seems to be even aware of these issues. Somehow he gives the impression that they actions were purely and always benevolent.

What of the Other People: To understand BTL and to opine on it being the proper paradigm for research one must look at it in its entirety. Not as a single set of a few “strange” people, albeit creative. What of the thousands of “40 year men” who when the Labs went bust were thrown into a world they could not understand. The outside have moved centuries forward and these rejects were cast asunder.

BTL Publications: There were two BTL publications; The BSTJ and Bell System Journal on Economics. The BSTJ seemed to serve two purposes: an in house technology release organ and a vehicle for publishing interesting technical desiderata. The Economics journal was different, it was the house organ attempting to gain peer acceptance in economics addressing issue which were related to justifying ATT positions with regulators. The economics justified the ATT filings but the people were at Bell Labs. One would have liked the author to explore these publications, what they did, how influential they were, and what the staff thought of them.

The General Executive Instructions, GEI: There was a manual, in fact a book shelf of them, called the GEI. It told you what size office you would have the desk type, the chair, the phone, travel restrictions, vacations,  the tiles on the floor, the promotion and salary scales, the organizations structure. It was all regimented to a fine point, and it applied to all. It surpassed anything one could think of even in the Federal Government. It was suggested that this was Kelly and his heavy hand but I was not there long enough to truly understand. But it was pervasive, and no entrepreneurial company could ever exist under such binding regulations.

Thus in my opinion the author had a wonderful opportunity but totally missed it. In my opinion, true historical writing requires both sides, the writer must be competent to understand when they are being used and by whom. Was Bell Labs a sine qua non of R&D. Not really. It did insure the monopoly control over telecom for decades but it also became the seed of its very destruction. Today there is not a single US manufacturer of telephone switching equipment. So this is a legacy?

---

[1] http://www.comara.org/legacy/oralhistory/

[2] http://www.ieeeghn.org/wiki/index.php/Oral-History:Paul_Rosen