The Story of the ThinkPad: Building a Global Success
Chapter 1 The Philosophy, Struggles and Technology behind the Birth of the ThinkPad
A “Made in Japan” portable PC
When I joined IBM (Japan), their newly established Japanese R&D center was in Fujisawa. Back then, IBM’s reputation was as a manufacturer of mainframes, but the Japanese laboratory mainly developed terminals.
The first job I was in charge of was to modify terminals for banks to Japanese specifications. At the time, the direction that Japanese and American magnetic stripe card readers scanned in was different, so terminals developed and manufactured in the United States could not be used as-is in Japan, therefore requiring a number of changes in specifications. I was put in charge of terminal communication technology at that time.
For me, not yet being a computer specialist, this was an extremely valuable experience in terms of gaining an understanding of computers and learning about logic circuits. I was also fortunate that a number of senior people in the lab took me under their wing and taught me a lot.
It was a single-byte (8-bit) world back then, even for computers, meaning that kanji, the Chinese character set used in the Japanese language, could not be displayed. Due to the insufficient number of bits, though it was possible to display the phonetic katakana characters, English lowercase characters could not be typed.
Tackling this challenge, IBM promoted the development of a two-byte kanji code system and established a new product development division to devise Japanese specifications for hardware enabling these complex characters to be displayed and printed. There was a call for people to work on this, and being hungry for new challenges, I immediately volunteered. And so my journey toward the ThinkPad began. IBM terminals displaying kanji for the first time in the world! I thought this was a very exciting prospect and an excellent learning opportunity. Looking back, the world of computers started undergoing significant changes around that time. It was the dawn of the age of distributed, compact, and open systems.
The relationship between mainframes and terminals was simple: mainframes are the brain, and terminals are the limbs. All the processing was done by large-scale computers called mainframes. But around that time, a shift in direction began to occur, it moved towards making terminals more intelligent and performing local tasks locally. This led to the emergence of the personal computer (PC).
Yet there was no concept of the “PC” in Japan at that time, and we called such machines “workstations” (WS), meaning, “work terminals.” As the plan was for these to be endowed with the intelligence not only to display kanji characters, but to print them as well, we called them workstations with CPUs. Kanji support was part of this trend.
PCs then did not have fully developed capabilities. CPU speeds were low and memory capacities insufficient, spurring IBM, among others, to begin pouring its energy into the development of workstations capable of displaying and printing kanji. The result was the 5550, a uniquely Japanese product with specifications completely different from those of U.S. products.
Until that point, I had been working on the development of IBM communication terminals, such as the 3270 and 5250. Communication terminal technology was my specialty, so I took charge of the development of the communication adapter for linking the 5550 to IBM’s network. Later, I took on the development of main units, overseeing derivative systems such as the 5540 and 5530. This phase corresponded to the development of kanji-capable workstations.
My instruction at the time was to merge Japan’s original workstations with the PCs of IBM US. The rationale was that continuing to follow a separate course would keep us from making the most of IBM’s power. I was called to the Boca Raton labs in Florida and was assigned to a project team that would work on how to develop IBM PCs and Japanese-language workstations based on the same architecture.
IBM’s latest PC at the time was the PS2. In 1982, building on the hardware and design of that machine, a desktop PC called PS55 was developed in Japan. It was in the process of creating this machine that the engineers at the Japanese laboratory learned PC technology, forming a base that would allow them to become the group of engineers that went on to develop world-class PCs.
Seeing this, IBM thought it would let the Japanese laboratory develop PC products for the global market, not just Japanese versions. However, the Boca Raton laboratory in the U.S. was already in charge of desktops and floor standing computers. So what should we Japanese create? We thought about a portable PC, which was something nobody had attempted before. And we thought we’d create this product in Japan.
There were also foreshadowing developments. In 1978, Toshiba had developed the first Japanese-language word processor, and in the ‘80s, such word processors began to spread in earnest, leading to the emergence of portable Japanese-language word processors in the second half of the decade. There was already an industry of suppliers of parts supporting the development of such products, including plasma display units and light keyboards.
In 1985, the laboratory at Fujisawa was relocated to the city of Yamato, and its name changed from the Fujisawa Development Laboratory to the Yamato Development Laboratory as a result. From that point on, IBM’s personal computers have been associated with the name Yamato Development Laboratory.
I, however, was stationed at IBM’s headquarters in New York. The development of AC-powered portable PCs was being launched at the Yamato Development Laboratory, and I was in daily contact with the Yamato lab driving development while in New York. But one day, perhaps out of exasperation, the Japanese side suggested that although it had planned to let me stay a while longer in the U.S., it now wanted me to go back immediately. There was no specific mention of what my responsibilities would be, and indeed they may have intentionally been keeping me in the dark. As I got back to Japan, I was told that I would be in charge of the development of battery-powered PCs. This was the start of my long and hard struggle.
A number of battery-powered portable PCs were already on the market by then. The first to appear was Toshiba’s Dynabook, a solid product. Compaq was another early pioneer. Adding IBM to this duo, we became one of the Big 3. Thus, IBM was clearly one of the main players in the early days of the portable PC. And Japan, more specifically the Yamato lab, spearheaded development.
Driven to pioneer
Our main development center was the Yamato lab, but the focus of our work was not specific to Japan. Yes, the Yamato lab was at the heart of the R&D effort, and had overall development responsibility, but all of IBM’s formidable capabilities were marshaled toward producing revolutionary new products for the world market, and a global network was set up for this purpose.
For example, mechanical design was done at the Boca Raton laboratory in Florida. The design of keyboards was done by the keyboard manufacturing division in Lexington, Kentucky, while the Raleigh Plant in North Carolina and the Greenock Plant in Britain were responsible for their manufacture. Naturally, a large number of partner companies were also involved in development. I was selected to serve as the Development Manager for this organization.
Battery-powered portable PCs are what are now known as notebook computers. This was a completely new type of product for us at the time. Moreover, I was suddenly charged with the major responsibility of managing a project of global proportions—no simple task.
Although I had officially returned to Japan, I made a point of being on location during the various phases, from development to production. For instance, at the initial concept design stage, I would stay in Boca Raton or in California, where a collaborating company was located. Later, at the joint design stage and also at the actual product completion stage, I was in Yamato. For the start of production, I was at the Raleigh Plant. Thus, I was moving between various plants as the development and production sites were distributed all over the world, with the main centers in the U.S. and Japan.
Back then, we had no notebook PCs, no internet, and no email. While in the U.S., I was unable to find out what was happening in Japan even if I wan
ted to, I had no access to Japanese newspapers. I ended up buying a full-band radio and would place it by the window to listen to short-wave broadcasts from Japan when I was desperate for news.
For internal communications, however, I think that IBM was very innovative. A few years after I joined the company, it established a worldwide online system for all employees. There were no video or audio capabilities, only text, and since email went through a large-scale computer, it could only be used at the company. We could make conference calls, but video conferencing was of course not available at the time. This was an age in which going onsite and talking face-to-face were still very important. Conference calls were also part of my daily routine. In those days, I led product developments over the phone with Boca Raton or Raleigh when I was in Japan; and with Japan when I was in Raleigh. So although I belonged to the Yamato lab, I was almost never there.
During that period, Takenobu Yonemochi (currently a managing director at Lenovo Japan) managed design activities at the Yamato lab, on a de facto basis. Though a keyboard specialist, he had general oversight of mechanical design. One time, as I had just returned to Japan, I received a phone call from Yonemochi’s wife.
“Is this Naitoh-san? Kenshin (what his wife called him) has been hospitalized.”
I rushed to visit him, and was relieved to find out that his condition was not serious. This goes to show how everybody was struggling like mad at the time.
Another time, I got a call from the wife of a junior member staff, asking me where her husband had gone. He was actually at the company, but he had not gone home in three days and failed to let anyone know where he was. His anxious family thought he had gone missing.
Everyone on the team went through a lot of hardship. This indicated how difficult it is to develop a new area. Our motivation was not due to company allegiance or salary. The conditions were so extreme that we could not have overcome them had we not been following an internal desire. I’m convinced that we were all driven by a deeper wish for self-realization, the challenges of exploring and pushing back the frontier, and the sense of achievement in opening up new territories. Our bodies were completely exhausted, but there was a fire in our eyes.
The war on EMI
Notebook PCs are now common. They can be carried around, allowing people to work away from their desks, in conference rooms and meeting rooms at companies, and also toted outside for use on business trips, or at cafes or anywhere, really. It’s easy to connect to the world, wherever you may be, over a wireless LAN or 3G connection. Today, these technologies appear simple and straightforward, but many obstacles lay in the path to their development. Hurdles needed to be cleared before any of these technologies can be achieved. We need to overcome these challenges in order to achieve the smallest possible size, the lightest weight, and the greatest user-friendliness, all combined with long battery life and high processing speed.
As an example, minimizing interference from unwanted electromagnetic radiation waves, known as spurious emissions, was no easy task. It is necessary to reduce this occurrence, which is commonly referred to as EMI (electromagnetic interference), to prevent the “noise” that compromises electronic transmissions such as TV reception, and meet country regulations.
Spurious electromagnetic wave radiation peaks in a high frequency range that is two, three, or even 10 times the basic frequency of a given device, such as a CPU operating at a given number of gigahertz. Nowadays, we have a nifty technique called spread spectrum that spreads out this basic frequency over a frequency range that does not trigger system errors. In the case of a basic frequency of 10 GHz, the peaks can be made shallower by spreading the frequency over the range of 9.99X to 10.00X GHz. In other words, this approach reduces the energy concentration for a given frequency. This technique was patented a long time ago, but only recently became practical. Before that, reducing interference by spurious emissions was a major challenge. Every electronics company had a large EMI testing room and was equipped with various highly sensitive testing devices and verification apparatus.
Often, we would create a new board, perform measurements and find out that it generated spurious emissions. We would try to contain this radiation mechanically, but after finding this impossible, we’d go back and redesign the board, through the process of trial and error. The greatest difficulty lay not in board design, but in the fact that the torque with which a single screw was tightened might influence the result. So even if we created 10 units of identical designs, it might be difficult to complete all to a satisfactory level. And of course no units could be shipped until spurious emissions could be suppressed.
The issue did not end with spurious emissions, however. The problem of interference within the machine proper was also a major hurdle. We were trying to fit various components, each interfering with the others through wave emissions, into a very small space. Even if we used the same parts, slight differences in how a coil was wound in each part could lead to a completely different result. All too often, just when we thought that things would definitely work, a glitch would appear in the production line, leaving us speechless. Even during final product testing, we would see products flagged with defect tags cropping up one after another.
A birthday surprise
As mentioned earlier, during the production stage, the supervision of the Yamato lab was entrusted to Yonemochi, while I was stationed at the Raleigh Plant. It would have been preferable to have had larger allowances for our designs, but in the initial stages of production this was impossible, as we had to cram so many components and parts into a small space. As a result, there were many problems at the plant. As I had to oversee the production line, I was unable to even go back to my hotel. At an even earlier stage, we would often be running the production line as designed, but the line would suddenly stop, forcing us to make specific changes. Moreover, this was the first time the production line operators were making such small products, and we had many freshly minted employees working on the line wearing shiny new IBM badges. So we had to deal not only with design problems, but inexperienced operators to boot. At first, I chipped in with advice on various aspects of the job. The production line operated in three shifts around the clock, and line operators were able to clock out when their shifts ended. But I had to greet the people of the next shift so I frequently had no chance to return to my hotel. The best I could manage was to rush back to my room once in a while for a shower. This went on for about two months. (Although I was treated as a VIP at the hotel where I was staying, I wasn’t able to enjoy the benefits and remember feeling down rather often.)
Our first product was scheduled for a commercial launch in September 1990, but this ended up being pushed back until March of the following year. As in so many things, the work dictates the schedule. We even worked on New Year’s Eve. I remember calling Yonemochi in Japan from Raleigh on that day to let him know that I had sent a fax about some problem. I was trying to be casual describing it, but of course he would have to go in to the office to read that fax. Essentially, I was asking him to go to work on New Year’s Eve, an important day for families in Japan. On top of that, my fax said that this had to be taken care of by the next day. Yonemochi immediately called his people in, and thus the Japanese development team’s work into the New Year began. We were already past the planned shipping date, and we were inevitably under intense pressure. It got to the point where I could almost see an electronic billboard in my mind displaying the number of days we were behind schedule.
After a succession of struggles, the day on which we finally received authorization to ship fell on March 13, 1991. By coincidence, this happened to be my birthday. I relayed the good news to Yonemochi in Japan, and remember that after I hung up the receiver, my Raleigh co-workers sang me “Happy Birthday To You.” They then took out some items and gave them to me, saying they were presents. It was all the parts that had failed during development! Everyone had set them aside somewhere for the occasion. One ite
m, for example, was a cover that had ended up blocking the slot of the floppy disk drive, a device one rarely sees these days. Another item was presented to me with explanation note saying, “This is a cap made of wrapping foil you used on the production line to prevent the electromagnetic waves from a coil from causing floppy disk errors.”
The PS/2 L40SX was the result of overcoming these many difficulties. This PC, which could be called the predecessor of the ThinkPad, was marketed as a battery-powered laptop PC.
ThinkPad built on 700C’s success
I believe that the success of the ThinkPad series was assured by the popularity of the first model, the 700C (known as the PS/55note C52 486SLC in Japan). This was launched in October 1992. No one knew at that time the extent to which notebook PCs would be embraced by the market, or how much market share they could capture. The degree of acceptance of the first model would therefore determine the level of future investment in the series. At the time, there were only around three manufacturers of notebook PCs, and they all shared this uncertainty.
The 700C incorporated all the technologies we learned during the development of the PS/2 L40SX. Technologies for minimizing EMI, saving energy, and battery management technology, as well as the newly developed MCA (micro-channel architecture) chipset and a 10.4-inch color liquid crystal display, the largest class at the time, were all built into the 700C.
The PS/2 L40SX, predecessor of the ThinkPad
released in March 1991