kobe university repository : kernel · entrepreneur, in a narrow sense according to shane (2004),...
TRANSCRIPT
Kobe University Repository : Kernel
タイトルTit le
University Start-up or Technology Transfer?:Japan-US comparison oncommercializat ion of new technology at the dawn of the computer age
著者Author(s) Takase, Susumu / Ito, Chiaki
掲載誌・巻号・ページCitat ion ,:
刊行日Issue date
資源タイプResource Type Journal Art icle / 学術雑誌論文
版区分Resource Version author
権利Rights
DOI
JaLCDOI
URL http://www.lib.kobe-u.ac.jp/handle_kernel/90002474
PDF issue: 2020-09-18
-1-
University Start-up or Technology Transfer?
-Japan-US comparison on commercialization of new
technology at the dawn of the computer age-
Susumu Takase and Chiaki Ito
Kobe University Graduate School of Business Administration
Abstract
At the present time, Japanese university start-ups cannot be considered successful in starting a
business. This paper aims to provide a historical account of this situation, through a Japan-US
comparison on commercialization of new technology at the dawn of the computer age, with the
focus on differences between the eco-systems of Japan and the US.
The first successful university start-up in the US is Digital Equipment Corporation (DEC), a
manufacturer of minicomputers. Around the same time in Japan, the commercial computer was
created through technology transfer from Electronics Test Laboratory (ETL) to large companies at
the initiative of the Ministry of International Trade and Industry (MITI).
Computers back in those days were the core of military technology. In Japan, one of the
defeated nations of WWII, an eco-system for supporting commercialization of new technology was
formed without involving universities. This resulted in a difference between the US and Japan in the
commercialization path of new technology: "university start-ups" by university-based researchers in
the US, and "technology transfer" to large companies in Japan.
Keywords: university start-up, technology transfer, eco-system, and transistor computer
-2-
1. Issues in question
This paper aims to provide a historical account of why Japanese university start-ups are not
successful in starting a business, from the viewpoint of the eco-system (1), through a Japan-US
comparison on commercialization of new technology at the dawn of the computer age.
Since 2000, with a plan for creating 1,000 university start-ups announced by the Ministry of
Economy, Trade and Industry, Japan has been encouraging university start-ups at the political level.
As of the end of fiscal 2008, the total number of university start-ups in Japan was 1,954, of which
1,339 were founded by researchers (Ogura, 2010, p. 2; p. 28). Looking at these numbers alone, one
may think that the original goal has already been achieved. This is not the case, however, according
to Tadao Kagono, a business scholar, and Kazuo Taki, an information-communication researcher
and founder of a system LSI designing company. In a panel discussion during the 11th national
convention of Japan Academic Society for Ventures and Entrepreneurs held in November 2008,
they presented the following three points of view in the light of the fact that most university
start-ups have not yet started a business (2):
First, there is a large difference between the role of a researcher who is responsible for research
and education at a university and that of an entrepreneur who identifies business opportunities in the
market. Secondly, no system is established in Japan for selecting university start-up entrepreneurs.
Thirdly, most university start-ups aim at social contribution or local cooperation, which is
incompatible with American-style venture capitals.
With the above problems in mind, we will compare representative cases of Japan and the US
about introduction of the transistor, which was a new technology at the dawn of the computer age.
The discussion will be centered on Ken Olsen, a founder of DEC, in the US's case, and on Shigeru
Takahashi of ETL in Japan's case (3).
2. Previous studies on the university start-up
2.1 Technology of the university start-up
-3-
Several definitions have been proposed for the term university start-up (Roberts, 1991; Kanai,
2010; Kirihata, 2010). This paper will be based on a definition which focuses on technology: "a
new company founded to exploit a piece of intellectual property created in an academic institution"
(Shane, 2004, p. 4, translation, p. 5).
Having thus defined the university start-up, Shane (2004, translation, p.83) has noted seven
characteristics of the technology which leads to founding of a university start-up: the technology is
1) radical, 2) implicit, 3) at an early stage, 4) intended for general use, 5) highly valuable to
customers, 6) representative of a major technological advance, and 7) strongly protected by
intellectual property rights. According to Shane, if the technology has these characteristics, a
start-up is likely to be founded to exploit the technology.
Shane (2004) has argued that existing companies are at an advantage over university start-ups in
commercialization of universities' technology. He has noted seven characteristics observed in many
technologies invented at university: the technology is 1) incremental, 2) explicit, 3) at a late stage,
4) intended for specific use, 5) moderately valuable to customers, 6) representative of a minor
technological advance, and 7) weakly protected by intellectual property rights. According to Shane,
where these characteristics apply, the technology is likely to be licensed to existing companies.
In this respect, Shane (2004) has cited "cannibalizing existing assets" (Utterback, 1994) to
explain why radical innovations are avoided by existing companies. In short, technology of the
university start-up is radical, while that of the existing company is incremental.
On the other hand, Christensen (1997, 2003) has further developed the framework for analyzing
radical versus incremental innovations suggested by Utterback (1994). Christensen (2003) has
refined the concept "disruptive innovation", which is a counter-concept of "continuous innovation",
into "low-end disruptive innovation" and "new-market disruptive innovation", describing the former
as an innovation that "address(es) overserved customers with a lower-cost business model", and the
latter as an innovation that "compete(s) against nonconsumption" (Christensen, 2003, translation, pp.
55-63).
-4-
Referring to "compete against nonconsumption", Nishizawa (2012) has expressed as "twofold
start-up risks" (p. 27), from the viewpoint that the university start-up involves not only a risk
associated with new technology but also that of creating a new market. A university start-up
entrepreneur, in a narrow sense according to Shane (2004), is defined as a person who achieves the
"new-market type disruptive innovation" which involves the "twofold start-up risks". In his case
study about DEC, which is a start-up of Massachusetts Institute of Technology (MIT) and will be
discussed in this paper, Nishizawa (2012) has shown how these "twofold start-up risks" accompany
the university start-up. Remarks related to the "twofold start-up risks" include "start-up from minus
two stage" by Shane (2004, translation, p. 174) and "start-up involving a valley of death" by Kanai
(2010, p. 8). To sum up the arguments above, the university start-up is the achiever of the
"new-market type disruptive innovation" which involves problems different from those faced in
technology transfer to existing companies or founding of an ordinary company (Christensen, 2003),
and these problems with "start-up from minus two stage" or "start-up involving a valley of death"
are expressed as the "twofold start-up risks".
Assuming that the university start-up is the achiever of the "new-market type disruptive
innovation" and necessarily involves the "twofold start-up risks" in terms of technology and market,
then how does a university start-up entrepreneur deal with the "twofold start-up risks"?
Christensen (2003, translation, pp. 99-100; pp. 128-132) has taken note of the transistor which
was the "disruptive technology" of the vacuum tube, and taken Sony as an example of the achiever
of the "new-market type disruptive innovation" to compare with the case of the vacuum tube and
Radio Corporation of America (RCA) which as an existing company achieved the innovation of the
vacuum tube. According to Christensen (2003), Sony has "succeeded in building 12 new-market
type disruptive innovations" in the period from 1950 to 1982 (p. 99), and among others,
productization of the transistor was the "new-market type disruptive innovation" which directly
contributed to founding of Sony. RCA, a major company at that time, targeted the existing market
of the vacuum tube (e.g., desktop radios and stationary televisions), and was struggling to apply the
-5-
transistor technology. In contrast, Sony chose to "compete against nonconsumption" in the new
market for portable electronics, namely, the pocket radio (1995) or the portable black-and-white
television (1959), eventually creating a new sales channel which eliminated the need for
maintenance as required for the vacuum tube, and establishing a new value network.
Christensen (2003) has remarked that only about five engineers, including Akio Morita,
participated in these "new-market type disruptive innovations". Christensen has argued that careful
observation and conversation with customers, through a conceptual lens of "jobs-to-be-done"
(Christensen, 2003, translation, p. 99) which narrows down jobs to one which no products have ever
been able to do, are breakthrough for the "new-market type disruptive innovation" as well as a
method to "compete against nonconsumption".
On the basis of this argument, a person who has the "ability for identifying business
opportunities", despite the "twofold start-up risks" in terms of technology and market, is appropriate
as an entrepreneur who achieves the "new-market type disruptive innovation". This view agrees
with the recent trend in the entrepreneur study (cf. Venkataraman, Sarasvathy, et al., 2012; Shane,
2012).
2.2 Problems in commercialization of new technology in Japan
This paper deals with the case of commercialization of the transistor computer around 1957,
when Japan and the US were presumably at similar technological levels. The following two
questions remain unanswered: One is why is it that, in the computer field, founding of a company
by researchers, namely, the university start-up, was undertaken in the US, while in Japan,
technology transfer to existing companies was promoted? The other is why is it that, university
researchers were also able to achieve the "new-market type disruptive innovation" in the US, while
in Japan, the achiever of the "new-market type disruptive innovation" emerged only from the
private sector, like Sony's founder Akio Morita, but none emerged from university researchers for a
long time?
-6-
The discussion so far has brought to light a problem for Japan: entrepreneurial human resources,
including Akio Morita, who understand technology as well as management, individually bear the
"twofold start-up risks". Christensen (2003) has pointed out that, since 1980 when Akio Morita
undertook political activities, Sony has created no "new-market type disruptive innovation". In the
case of Japan, entrepreneurs having the lens of "job-to-be-done" may have been burdened with the
"twofold start-up risks" behind the "continuous innovation" during the economic growth from 1950
to 1970.
In this regard, Nishizawa (2012) has noted a dissociation between micro and macro in creation
of start-ups in Japan, and introduced the concept of eco-system, as a mesolevel-framework serving
as a node of the micro and macro, into the discussion of the university start-up, and theoretically
explained the problem of the "twofold start-up risks". The eco-system is a concept not for analyzing
individual companies but concerned with a cluster or a relationship with a core company (e.g.,
Iansiti and Levien, 2004). The discussion of "triple helix" by Etzkowitz (2008), who has focused on
industry-government-academia cooperation, can be regarded as the first example of the eco-system
featured in the entrepreneur study. Etzkowitz (2002) has provided a historical account of why many
university start-ups were created by MIT. Accumulation of these studies shows that the
entrepreneurial university is noteworthy as a framework for realizing the "new-market type
disruptive innovation" beyond the "twofold start-up risks" which are imposed on individual
entrepreneurs.
3. Analysis framework and study subject
As a problem facing Japanese university start-ups, Kanai (2010, pp. 273-275) has described a
problem of entrepreneurs in the context of Japanese culture. However, it was 1955 that the Nobel
laureate William B. Shockley Jr. founded a company, which became the origin of Silicon Valley,
and in commercialization of the transistor developed by Shockley, as mentioned above, Sony had
already reached a similar technological level. Therefore, in view of the current situation where the
-7-
achievers of the "new-market type disruptive innovation" are being exhausted in Japan, it would
have a certain significance to trace the roots of the university start-up in the US and clarify the
historical development leading to creation of the commercial computer in Japan which introduced
the transistor around the same time as the US.
To make a Japan-US comparison, we have selected the cases from around the same time. These
cases are events around 1957, when the transistor was gaining in popularity, and Japan and the US
were at similar levels of the technology required for computer development (Special Committee of
the History of Computing, Information Processing Society of Japan, 1998). We have therefore
determined that these examples would make it easy to clarify the differences between the countries
in commercialization of new technology and its historical development.
This paper will make a Japan-US comparison on commercialization of the transistor computer
at the dawn of the computer age, on the basis of the following three subjects: 1) when and how was
the transistor computer commercialized in the US?; 2) how was the transistor computer
commercialized in Japan around the same time?; and 3) what difference is found between Japan and
the US in the commercialization process of the transistor computer? By clarifying these questions,
we will extract and examine the problems that impede Japanese university start-ups from starting a
business.
4. Cases under study
This case study has a focus on Ken Olsen, the founder of DEC, as the US's case, and Shigeru
Takahashi of ETL as Japan's case.
The following table briefly describes their biographies.
-8-
Table 1 Biographies of key persons in cases under study
Ken Olsen Ken Olsen, the founder of DEC, was born February 20, 1926, as a
second-generation northern European immigrant. In 1952, he earned his master's
degree in electrical engineering at MIT. At MIT Lincoln Laboratory, Olsen was engaged
in the development of TX-0, a successor unit to Whirlwind Computer. His
accomplishments include invention of the magnetic core memory and development of
the minicomputer. Ken Olsen passed away at the age of 84, on February 6, 2011.
Shigeru Takahashi Takahashi was born November 22, 1921. After graduating from Keio University, he
worked as president of Tokyo University of Technology, and later as vice-president of
Information Processing Society of Japan. He received the first Doctor of Engineering
from the Keio University. In 1944, Takahashi joined Electronics Test Laboratory, and
transferred to Hitachi, Ltd. in 1962. In 1980, he became a professor at the University of
Tsukuba. Shigeru Takahashi passed away at the age of 84, on January 4, 2005.
Now, let us take a look at the history of the transistor technology, which, as mentioned above in
Sony's case, provides the setting for this study. In 1947, the transistor was jointly invented by three
members at Bell Laboratory. Shockley, one of the members, aimed at commercialization of the
transistor and established Shockley Semiconductor Laboratory in Mountain View, California, in
1955. This eventually became the origin of Silicon Valley. In 1956, the three members including
Shockley of Bell Laboratory won the Nobel Prize in physics. In 1957, as a silicon-based
semiconductor research was canceled, eight researches left Shockley Semiconductor Laboratory,
and founded Fairchild Semiconductor International Inc. Among these "traitorous 8" were Robert
Norton Noyce and Gordon E. Moore, who established Intel in 1970.
5. Case study
5.1. Ken Olsen's case
5.1.1 Prototype unit TX-0 and commercial unit PDP-1
-9-
In 1957, Ken Olsen, then a researcher with MIT Lincoln Laboratory, founded a company to
commercialize a prototype unit (TX-0), which was the result of the national military project, Project
Whirlwind. The company founded at that time is DEC.
When mention is made of the transistor, Olsen was responsible for implanting about 3,600
transistors into Whirlwind Computer (which started operation in 1951) produced with about 5,000
vacuum tubes. TX-0 (Transistor Experimental Computer; started operation in 1956) which Olsen
had worked on became the earliest-stage transistor computer and the world's most advanced
computer of the time (Ceruzzi, 2004, translation, p. 158).
Although the original purpose of the military research commissioned to MIT was computer
development for controlling flight simulators and air-defense missiles, setback of a project
succeeding TX-0 prompted the project members to found a company. After starting the company in
1957, Olsen, with his expertise in application of the transistor to the computer, adopted an
architecture which took full advantage of all the functions that the transistor can provide, and in
1961, completed a real-time computer PDP-1.
5.1.2 Military research at MIT and commercialization of its outcome
MIT became a pioneer of the "entrepreneur university" which aggressively commercializes
research outcome (Etzkowitz, 2002). Vice-president Vannevar Bush was a central figure in this
movement, and is known to have had an important governmental position concerned with military
research during World War II (Ueyama, 2009; Etzkowitz, 2002; Nishizawa, 2012). MIT Lincoln
Laboratory, where Olsen was active, had been invited to MIT by Bush for the purpose of the
national defense project, and established on the premises of Lexington Air Force Base near Boston.
Thus, it is reasonable to assume that the computer back then was regarded as a weapon just as
nuclear power generation was. That is, Whirlwind Computer was developed at MIT as the computer
for controlling flight simulators or air-defense missiles on a bountiful defense budget during World
War II and the Cold War.
-10-
In the prelude to this event was Karl Compton, who was president of MIT from 1930 to 1948.
In order to restore Boston's economy, which, after prosperity of its textile industry, was beginning
to decline in the face of rise of the auto industry in the Middle West of the US, Compton introduced
a framework for industry-academia cooperation into MIT, in which MIT was to play a role different
from that played by "research universities", pioneered by Berlin University and including Harvard
University, Johns Hopkins University, and Chicago University (Nishizawa, 2012). One of the
features of the cooperative framework was a side job provision. When the Humboldt revolution,
which is a "fusion of study and education" at Berlin University in the late 19th century, is regarded
as the "first university revolution", then, this MIT's approach pursued by Compton can be called the
"second university revolution", an attempt to lead the research activities to promotion of economic
and social growth. In this way, MIT became the world's first "entrepreneur university" which
assumes three roles of education, research, and industry-university cooperation at the same time
(Etzkowitz, 2002; 2008).
5.1.3 Creation of American Research and Development Corporation: the world's first venture
capital
Under the initiative of Bush, MIT was enjoying the US federal government's support for its
military technology research, ample R&D funds, as well as a high degree of freedom during World
War II. However, no fund was provided for founding a start-up in which researchers had a personal
stake, which inevitably led to founding of companies in other regions based on MITs technological
seeds; Boston after the rapid economic growth had turned into a mature market with high labor cost.
There was an earnest proposal from Boston's financial circles for establishment of a venture
capital as a special monetary institution which provides an initial fund for MIT's technological
seeds as well as managerial support for helping the start-up out of debt at an early stage. In 1946,
the world's first venture capital, American Research and Development Corporation (ARDC), was
established by MIT's president Karl Compton and Georges F. Doriot, a professor of Harvard
-11-
Business School. (Nishizawa, 2012).
5.14 Georges Doriot and Jay Forrester
The preceding events were setting up the stage for MIT, in terms of technology and
management, to create start-ups. In 1957, Olsen received a 70-thousand dollar investment for
founding DEC from ARDC headed by Doriot. By the time DEC went public in 1966, the market
value of the company had reached two million dollars. Thus, DEC is regarded as the world's first
successful university start-up, and ARDC as the world's first successful venture capital (Rafkin, et
al., 1988; Nishizawa, 2012).
Another factor which significantly contributed to the successful start-up of DEC is that the
invention of the magnetic core memory, which is a memory device of the computer developed in
the national project, Project Whirlwind, was jointly patented by the project leader Jay Forrester and
Ken Olsen (Patterson and Hennessy, 2007). In those days, computer development required huge
amounts of money and was shied away by inventors because of the difficulty of fund recovery. As a
condition for making an investment, Doriot allegedly requested change of the company name from
Digital Computer Corporation to Digital Equipment Corporation (DEC) (Rafkin et al., 1988). Olsen
adhered to Doriot's advice, and for several years since the company's founding in 1957, DEC was
dedicated to manufacture and sales of digital equipment, such as the core memory and the logic
circuit, for the company name's sake. Only after achieving a sale of digital equipment that can cover
the development cost of a real-time computer, Olsen finally embarked on development of PDP-1 in
1960.
5.1.5 Founder Olsen and launch of university start-up DEC
For better or worse, Olsen was a technology-oriented, unique entrepreneur. It is known that, by
the time of initial public offering in 1966, all the founding members of DEC except for the directors
sent from ARDC had left (Schein, 1985). After conclusion of Project Whirlwind, Forrester, Olsen's
-12-
superior at MIT Lincoln Laboratory, assumed a post as a professor of system dynamics at MIT
Sloan School of Management, and acted as DEC's external board member. Later Forrester got into
bad terms with Olsen and stepped down when the company went public (Rafkin et al., 1988).
What is remarkable is that Olsen had had no business experience before he founded the
company. Olsen was a researcher who produced the prototype unit through the national project for
military technology development and acquired the patent which became the core of his business.
Doriot, on the other hand, compensated for Olsen's lack of business experience by raising funds and
sending directors to DEC's board meeting, providing a ground for Olsen to have experience in
managerial work. In short, Doriot established a framework for supporting entrepreneurs through
ARDC.
Computers back then were expensive. IBM therefore adopted a lease-based business model
(Ceruzzi, 2006). It was a common practice in the computer business before the appearance of DEC,
that software run on a leased computer was provided to the users for free, meaning that the software
development cost was included in the lease fee. Hence IBM users were prohibited from modifying
or changing the machine or software, as in that case the computer falls outside the scope of support.
In contrast, DEC introduced a business style of selling out the computers. It disclosed products
specifications and provided user-oriented manuals to encourage the users to modify or change the
machine or the software. For this reason, PDP-8, a successor unit to PDP-1, became a representative
minicomputer and found ardent support from hackers who were students of MIT Artificial
Intelligence Laboratory (Levy, 1984). Thus, DEC is regarded as the origin of "open innovation"
(Chesbrough, 2003).
5.1.6 MIT connection and matrix organization
Olsen adopted the MIT's rules when he set DEC's company rules (Rafkin et al., 1988). For
example, a sabbatical system as with university faculty was introduced. Gordon Bell, who led the
VAX system as the number two of DEC, was away from the company from 1966 to 1972 to be
-13-
engaged in research of computer architecture with Allen Newell at Carnegie Melon University (Bell
and Newell, 1971).
DEC was a typical university start-up also in the sense that university's research culture which
values technology as in MIT was introduced as it is (Schein, 1985). University researchers present
their own argument obtained from research and discuss matters. This was also the case at DEC;
under the philosophy of "do what you proposed by yourself ", there was a constant discussion
toward productization (Kunda, 2006).
From the viewpoint of engineering, donation of PDP-1 from DEC to MIT, and adoption of the
interactive computer and user-friendly open architecture made the connection between MIT and
DEC even stronger. Eventually, researchers of Lincoln Laboratory started to crowd into DEC's
Maynard Mill, which looked just like a university laboratory (Rafkin et al, 1988, translation, p. 45).
This connection was the only way a start-up like DEC could compete against the large company
IBM (Ceruzzi, 2004).
After Forrester had left DEC at the time of the company's initial public offering in 1966, Olsen
invited Edgar H. Schein, a professor of organizational psychology, as a consultant from MIT Sloan
School.
This has some bearing on the fact that, around 1965, before going public DEC had introduced a
"matrix organization" as a system for managing business start-up by researchers (Kanter, 1983;
Rafkin et al., 1988). While a manager having overall authority for his/her own machine acted as an
entrepreneur who managed from research and development through marketing to service, there was
an incessant chaos due to such contradiction, as is often said, "1 man 2 bosses".
Schein made several attempts. Among them was a "woods meeting" in which future course
(strategy) of DEC was discussed in groups. This meeting was held monthly outside the company,
and the manager having overall authority for the products was freed from daily tasks. The role
Schein played there was a role known as that of a "process consultant" (Schein, 1969). It was a role
not to advise on business issues including the computer, but to promote the discussion process so as
-14-
to evoke individuality amid the chaos of the matrix organization (Rafkin et al., 1988).
Schein's consulting activity for DEC lasted for over 20 years. During this period,
technology-related exchange with MIT became a source of technological innovations at DEC. In
particular, partly because the DEC's machine provided a foundation for the development of the
Internet or UNIX, Olsen's founding of DEC and the machine developed by the company are today
highly evaluated in the computer history (Ceruzzi, 2004).
5.2 Shigeru Takahashi's case
5.2.1 Split of Electronics Test Laboratory
In 1946, the world's first general-purpose computer ENIAC using vacuum tubes was developed
at the University of Pennsylvania with military funds intended for vast ballistic calculations. This
prompted the University of Tokyo (TAC), Osaka University, and Fujifilm Corporation (FUJIC) in
Japan to produce vacuum tube machines (Endo, 2010).
In 1948, the former Postal and Telecommunication Ministry's Electronics Test Laboratory,
which had been engaged in the development of military radars before World War II, was split by
General Headquarters (GHQ) into two bases: one is Denden-kosha's (Nippon Telegraph and
Telephone Public Corporation's) Electrical Communication Laboratory which succeeded the
communication engineering division of Electronics Test Laboratory, and the other is Electronics
Test Laboratory (ETL) of Agency of Industrial Science and Technology, which was virtually newly
established under the Ministry of International Trade and Industry (MITI). At these two bases,
research and development for the next-generation electronic computer were competitively pursued.
Following the commencement of a power distribution business in 1890, ETL was established in
1891 for inspection of telephone and telegraph equipment under the Postal and Telecommunication
Ministry. With the progress of the electrical industry, the laboratory took on research on electrical
power and its applications in 1903. Until the early Showa era, so-called heavy current engineering
(electric power field) such as power generation, transmission, and distribution accounted for a
-15-
larger part of research activities. Since then, the light current engineering (communication field) has
also become popular as a research subject, and by 1941 when Japan plunged into World War II,
ETL had established a status as the national research center for military technology in the field of
electricity and electronics, and grown to be a large laboratory with more than 2,000 staff members
(Hiroshige, 1973; Imaoka, 1989).
The mission of Bell Laboratory in the US at that time was development and research. A
structure for placing orders for equipment exclusively with Western Electric, which is a
manufacturing division of its parent company AT&T, in other words, a structure of "AT&T as the
telephone company, Bell Laboratory for basic research, and Western Electric as the telephone
equipment manufacturer" was established. Japan followed this example and created a structure
constituted of Denden-kosha, Electrical Communication Laboratory, and affiliated domestic
manufacturers. It was a role-sharing system where Electrical Communication Laboratory provided
purchase specifications and the manufacturers designed and manufactured the products accordingly.
The name "Denden-family" was derived from this cooperative relationship (Toda/Matsunaga,
2003).
5.2.2 Transistor and parametron
In 1947, the transistor was invented at Bell Laboratory. It took a while for the transistor to find
new applications and surpass the vacuum tube in terms of cost as well as reliability. Meanwhile, in
1954, the parametron was invented by Eiichi Goto, who was then a graduate student at Hidetoshi
Takahashi's laboratory in Faculty of Science of the University of Tokyo, as a logic circuit element
utilizing the excitation phenomenon (Takahashi, 1998; Endo, 2010). Zenichi Kiyasu of
Denden-kosha's Electrical Communication Laboratory, taking notice of the parametron as a
computer element, encouraged the affiliated manufacturers and brought MUSASHINO-1 to
completion in March 1957. In line with Kiyasu's intension, Electrical Communication Laboratory
decided to focus exclusively on the parametron as a computer element (Endo, 2010, p. 93).
-16-
At ETL, on the other hand, the transistor was intensively researched according to the intention
of Hiroshi Wada. Wada had been at MIT as a delegate of the MITI for one year after the conclusion
of the San Francisco Peace Treaty. During the visit, he saw real Whirlwind Computer (Takahashi, et
al, 2010). After returning to Japan, seeing Japan's future in the electronics such as the transistor and
the computer, Wada set up an electronics division in ETL in August 1954 and became its manager.
At that time Kiyasu suggested to Wada to adopt the parametron. Since Shigeru Takahashi who was
in charge of the relevant experiments reported that "(the parametron type) operates at low speed and
consumes a large amount of power", ETL decided to develop the transistor computer (Imaoka,
1989; Takahashi, 1998).
To sum up, in Japan, the two laboratories having the same roots in the Postal and
Telecommunication Ministry's Electronics Test Laboratory ― Denden-kosha's Electrical
Communication Laboratory and ETL ― competitively conducted research and development based
on the parametron and the transistor, respectively. In the mid-1950s, there was a movement of
commercialization of the computer, while weighing these two technologies against each other, by
large electronics manufacturers backed by Denden-kosha and the MITI.
5.2.3 MITI's computer policy
A 1972 report titled "Corporation Japan: close tie between government and industry" by the
United States Department of Commerce includes a case study on the Japanese computer industry,
which was written under the instruction of James Abegglen (Kaplan, 1972). According to this report,
the computer was first imported from the US in 1954, and five large manufacturers, NEC, Fujitsu,
Hitachi, Matsushita, and Toshiba, started licensed production of the transistor. These manufacturers,
fearing that Japan's computer market would be dominated by foreign manufacturers, made a request
to the MITI to speed up the development of domestic technology (Kaplan, 1972, translation, p.
135).
In response, the MITI set up the Electronic computer development committee in the Radio
-17-
Engineering & Electronics Association in April 1955, and in the following year, started research on
foreign computers on a budget of 800 thousand yen. In May 1957, nine million yen was allocated
for an attempt to produce a computer superior in performance to the middle-sized machine IBM650
which was popular then, with cooperation of the large domestic manufacturers. Hiroshi Wada, the
manager of the electronics division of ETL of the Agency of Industrial Science and Technology
under the MITI took the initiative in this series of moves (Takahashi, 2003). In June 1957, to
promote the electronics industry including the computer, the MITI drafted the Act on temporary
measures for promoting the electronics industry. Hiroshi Wada had a strong interest in this drafting,
too, and instructed the ministry's officials in drafting the outline of the bill, and responded to Diet
questions. In August of the same year, the electronics industry section was established in the heavy
industry division of the MITI (Takahashi, 1998, 2003).
Since then, the MITI has held a leading position and put a great effort into protection and
cultivation of Japan's computer industry. Examples include: negotiations with IBM for a patent
agreement which was concluded on October 29, 1960; establishment of Japan Electronic Computer
Company which was a rental company and fifty-fifty joint venture of the government and private
sectors; and creation of large-scale industrial technology research and development system
(large-scale project) (Kaplan, 1972).
5.2.4 Computer development at ETL
The electronics division set up in ETL in August 1954 was constituted of 50 staff members and
three sections of the circuit, component, and electronic measurement. The manager Wada set a
strategic target of developing an electronic computer, which included many unknown elements then,
and a computer using transistors. Research and development were vigorously promoted, circuit
theory and circuit technology in the circuit section, and semiconductor components such as the
transistor and the printed circuit board in the component section (Narisada, 1995). Unlike the US,
research and development of the transistor and the circuit design of the transistor computer were
-18-
conducted side by side in the component section and the circuit section, respectively, within ETL's
electronics division which had been directed by Wada since 1954.
As of 1954, before the transistor computer, ETL had already completed a relay type computer
ETL-mark I (Electronics Test Laboratory-mark I) in 1952, and an improved relay-type computer
ETL-mark II was subsequently under development. In November 1954, Shigeru Takahashi, who
became the leader of the newly established circuit section of the electronics division, Hiroji Nishino,
and Isokazu Matsuzaki, et al. started development of prototype unit ETL-mark III, which was not a
vacuum tube computer, but instead, used the next-generation "point-contact transistor" as its major
element.
ETL-mark III utilized T-1698, a point-contact transistor produced by Tokyo
Telecommunications Engineering Company (later Sony). Hiroji Nishino has recalled that, Sony's
former president Masaru Ibuka would often visit a laboratory of ETL in Nagata-cho with the
point-contact transistor T-1698 in his hand which had just been completed that day (Ukai, et al.,
2011). This T-1698 was the only one domestic high-speed transistor available at that time, cost as
much as 4,000 yen apiece, and still had problems in reliability (Takahashi, 2003). Active research
was conducted by Makoto Kikuchi, Yasuo Tarui, Seiichi Denda, et al., as a joint effort of the public
and private sectors, toward domestic production of the transistor including its manufacturing
technology (Aida, 1991).
As a result, the design work was finished in March 1956, manufacturing work in April, and
operation was started in July. Having started off in November 1954 with a total estimate of 2,840
thousand yen, ETL-mark III was completed in as short a period as one and a half years. Shigeru
Takahashi has attributed the early completion to the facts that all the production work except for the
components was performed within the laboratory and that reasonable specifications were adopted
(Takahashi, 2003).
5.2.5 Creation of ETL-mark IV
-19-
ETL-mark III was the world's first stored-program transistor computer, which started operation
almost at the same time as TX-0 developed by Ken Olsen. As the point-contact transistor was used,
ETL-mark III operated at high speed, but often broke down (Nishino, 1985). In view of this,
development of ETL-mark IV using the junction type transistor was started in October 1956, in the
circuit section of the electronics division, to enhance the reliability even at the cost of speed. The
members were Shigeru Takahashi, Hiroji Nishino, Isokazu Matsuzaki, and Hiroshi Yoneda. At this
stage, domestic production of the germanium alloy junction type transistor had already been started
and become far more stable than the point-contact transistor (Takahashi, 2003).
ETL-mark IV used about 470 units of HJ-23 manufactured by Hitachi, and the total budget was
about five million yen. The binary-coded decimal notation (BCD) was adopted, and ETL intended
to commercialize ETL-mark IV by technology transfer. In November 1957, ETL-mark IV was
completed only after 13 months from the kickoff. Compared with ETL-mark III, ETL-mark IV was
much more stable and capable of performing calculations for obtaining large prime numbers
without being shut down for more than 100 hours in 5 days (Takahashi, 1976).
Many Japanese manufacturers wished to introduce the technology of ETL-mark IV, and there
were offers from four companies, NEC, Hitachi, Matsushita Communication Industrial, and
Hokushin Electric Works. Although Japanese manufacturers including Fujitsu, Hitachi, and NEC
had commercialized the parametron-type electronic computer developed by Denden-kosha and
Electrical Communication Laboratory, partly because five large manufacturers, NEC, Fujitsu,
Hitachi, Matsushita, and Toshiba had already started the licensed production of the transistor, by
1957, Japanese manufacturers began full-scale commercialization of the computer using the
junction type transistor.
5.2.6 Cooperative tie between Hitachi and ETL
In March 1962, Shigeru Takahashi left ETL and joined Hitachi. Before that, Takahashi had been
maintaining close contact with the company through technical guidance he provided on Hitachi's
-20-
electronic computer (Imaoka, 1989). In the case of HITAC301, which is a commercial machine of
ETL-mark IV, it is known that Yasukichi Hatano, Shigeru Takeuchi, et al. of Hitachi had been
doing designing work in Takahashi's laboratory in ETL since around June 1958 (Takahashi, 1996).
Prior to this, Hitachi had completed parametron-type electronic computer HIPAC MK-1 in
December 1957 to respond to an internal need for a computer to be used by Hitachi Cable (an
electric cable factory then) for designing the tension of power lines for Electric Power
Development's Tadami-kansen line. In March 1958, a computer designing section of the radio
division was established in the Tozuka factory. Thus, the transistor electronic computer HITAC301
was completed in April 1959 under the direction of Shigeru Takahashi of ETL after about one year
from the start of development.
While DEC's PDP-1 (completed in 1961) has the roots in Whirlwind Computer designed for
controlling flight simulators and air-defense missiles and are characterized by its capability of
real-time processing, the system which became the origin of on-line, real-time processing in Japan
was a prototype of former Japan National Railways' seat reservation system, Magnetic-electronic
Automatic Reservation System (MARS), which is known today familiarly as "Green window".
At Japan National Railways' technology laboratory, since 1955, members including Mamoru
Hosaka and Yutaka Ohno spent a lot of time in study meetings and discussion aiming at
introduction of the computer for automatic control. Hosaka et al. were keenly aware of the need for
a seat reservation system, and had a technological vision of vacant seat pattern search, quick file
updating, and independent parallel processing of input and output, which are essential for the
system. In 1957, they set up a research committee with Hidetoshi Takahashi of the University of
Tokyo as the chairman, and external members from companies, etc., and officially started to
consider computerization of the seat reservation system. As a result, in 1958, MARS1 was
contracted to Hitachi and developed at the Tozuka factory mainly by Yoshihiko Tani.
In 1959, MARS1 was installed at Tokyo station and connected with reservation devices
provided at ten ticket windows. In February 1960, this Japan's first domestic on-line system has
-21-
started handling reservations for 15 days of about 2,100 seats for Tsubame and Hato trains, and in
the following year, MARS1 was also installed at Nagoya and Osaka stations to extend its service
area. The operation rate in the first ten months was 99.86%, which later grew to 99.95% or more
(Uraki/Odaka/Kawabe, 1985).
Shigeru Takahashi worked at Hitachi for about 18 years from April 1962 to April 1980. After
that, he became a professor at the University of Tsukuba. Here is a brief look back on his career: He
was originally in the position, as a researcher with ETL, to provide guidance on starting a computer
business in Hitachi. Then, at the stage where launch of the business came in sight, he moved to
Hitachi and took charge of design and development of major products such as HITAC 8000 series,
DIPS, and early M-series. Records show that, throughout his career at Hitachi, Takahashi played an
active role as a person responsible for product design in the company's computer operation
(Takahashi, 1996).
6. Findings
6.1 Findings from Ken Olsen's case
This chapter summarizes the findings from the case study. First, there are three findings from
Ken Olsen's founding of DEC, namely, the first successful university start-up, aiming at
commercialization of the transistor computer in the US, which is the first subject of this paper.
First, Ken Olsen of MIT Lincoln Laboratory, who developed the prototype unit TX-0 as a result
of the national military project, founded DEC and played a pivotal role in commercialization of the
transistor computer and setting up the accompanying business.
Secondly, founding of DEC was supported by many university researchers in terms of funds,
technology, and management. More specifically, in terms of funds, DEC obtained an investment
from ARDC headed by Professor George Doriot of Harvard Business School, which became the
world's first successful venture capital. In terms of technology, Olsen and Jay Forrester, who was
Olsen's superior at MIT Lincoln Laboratory and later became a professor at MIT Sloan School,
-22-
obtained a joint patent for the magnetic core memory. In addition, Olsen received Forrester's
cooperation for joining as the director of DEC. In terms of management, process consultation was
provided by Edger Shein, also professor at MIT Sloan School, to deal with problems accompanying
introduction of the matrix organization.
Thirdly, after DEC's successful public offering in 1966, the university start-up and the venture
capital obtained acknowledgement, and each pioneered the eco-system for supporting
commercialization of new technology. These were transferred from MIT to Stanford, from Boston
to San Francisco, constituting the core of the industry accumulation in Silicon Valley.
6.2 Findings from Shigeru Takahashi's case
There are three findings from the second subject, commercialization of the transistor computer
in Japan, namely, the case of technology transfer from ETL to Hitachi by Shigeru Takahashi.
First, development of the transistor computer was jointly conducted between ETL and the large
companies subsidized by the MITI, under the direction of Shigeru Takahashi, et al. who was the
researcher in charge at ETL under the MITI. In Japan, researchers who developed the prototype unit
did not found a company to directly get involved with commercialization, but instead only provided
guidance on technology transfer to a large company. Then, by moving to the large company from
ETL, leading researchers promoted the technology transfer.
Secondly, Japan was technologically far behind the US as to peripheral equipment such as the
memory device and the input/output device, even though the countries had equivalent level of
transistor technology. For this reason, there was no other choice for Japan but to introduce patents
related to manufacturing technology by technology-sharing agreement with IBM, RCA, etc. This
case of joint development set a precedent for the system of cooperation between the government
and private sectors led by the MITI.
Thirdly, it was difficult for Japanese universities at that time to conduct research and
development of the computer for military purposes such as ballistic calculation supported by
-23-
military funds. Under such circumstances, the measures drafted by key members of ETL under the
MITI were implemented by existing large electronics manufacturers, who realized
commercialization of the transistor computer.
6.3 Findings from the Japan-US comparison
On the basis of the Japan-US comparison, this section presents findings about differences
between Japan and the US in commercialization process of the transistor computer, which is the
third subject.
Ken Olsen and Shigeru Takahashi were researchers representative of the US and Japan,
respectively, who implemented integration of the transistor into the computer. Judging from
perfection levels of their machines, Olsen and Takahashi are assumed to have been at similar levels
of technological capability including skills and knowledge around 1957. In the US, Ken Olsen
founded DEC with an investment from the world's first venture capital ARDC. In Japan, on the
other hand, Shigeru Takahashi as a researcher with ETL gave technical guidance to the large
electronics manufacturer, and then joined Hitachi to support its computer division.
-24-
Table 2 Japan-US comparison on commercialization of transistor computer
Japan United States
Shigeru Takahashi Key person Ken Olsen
Electronics Test Laboratory Laboratory MIT Lincoln Laboratory
ETL-mark IV Prototype TX-0
Technology transfer to large companies
Method of commercialization of
new technology University start-up
HITAC301 Commercial unit PDP-1
Hitachi Technology transfer to DEC
One business division of a large company Business form Company specialized in
computer Protective measures by the
MITI Characteristics Commercialization of military
technology International competitiveness
of domestic computers (up to the 1980s)
Result Creation of the world's first
successful university start-up and venture capital
To sum up the findings above, researchers in the US and Japan, who had equivalent levels of the
ability for identifying business opportunities regarding technology (technological ability) of
"introducing the transistor to the computer", pursued commercialization in different paths, namely,
through "university start-up" and "technology transfer to large companies", respectively.
Accordingly, we infer that, unlike Ken Olsen, it is not necessary for Shigeru Takahashi to possess
the "ability for identifying business opportunities" regarding the market which is essential to access
a "fund source" such as a venture capital. Table 2 lists characteristics of commercialization of the
transistor computer in the Japan-US comparison.
7. Observation
This comparison study covers the dawn of the computer age about half a century ago. Our
comparison has shown that Japan and the US adopted different policies for commercialization of
-25-
new technology (transistor computer). Commercialization of new technology was realized through
founding of companies by university researchers backed by venture capitals in the US, and through
technology transfer to large companies in Japan, and both countries are evaluated as successful.
Following these successful cases, efforts were made toward commercialization of new technology
in a manner suitable to the situation of each country.
This resulted in two important differences between Japan and the US in commercialization of
new technology: First, in the US, universities became the starting points of commercialization of
new technology, while in Japan, ETL and large companies served as the starting points, which was
established as their social roles. Specifically, in the US, MIT, Stanford, Carnegie Melon
Universities, etc. became the bases and led the computer study with military research funds
(Etzkowitz, 2002; 2008). In Japan, the cooperative structure of the MITI, ETL, and large companies
was established through national projects, while keeping out universities for fear of involvement in
military research. Of the social roles of universities, which were under the jurisdiction of the
Ministry of Education, Science, Sports and Culture, the educational aspect with the primary
objective of supplying human resources to large companies assumed more importance than
advanced research (Takahashi, 1996).
Secondly, in the US, young researchers who had experienced large-scale research projects
became creators of university start-ups, while in Japan, almost no university start-up was created by
young researchers who had experienced large-scale research projects. In this respect, since the
initial public offering of DEC (1966), many young university researchers have started a business
with the support of venture capitals (Shane, 2004). In Japan, by contrast, until 1999, the university
start-up was virtually not allowed by the code of faculty members of national universities (Ogura,
2010).
Now, in light of the differences between Japan and the US in commercialization of new
technology, we will consider the issue: why are Japanese university start-ups not successful in
starting a business? First of all, the framework for supporting commercialization of new technology
-26-
discussed in this paper has already been explained by the concept of eco-system in previous studies
(Iansiti and Levien, 2004; Sugiyama/Takao, 2011; Nishizawa, 2012). Our comparison study has
shown that the eco-system for promoting commercialization of new technology was originated from
commercialization of the transistor computer about 50 years ago. On this basis, the difference
between the eco-systems of Japan and the US derived from the findings can be put in this way: an
"eco-system centered on the university start-up" in the US, and an "eco-system centered on
technology transfer to large companies" in Japan. In other words, the former is an
"entrepreneur-centered eco-system", while the latter is an "entrepreneur-absent eco-system".
In the US, cutting-edge technologies were developed at universities for military purposes, and
these developers were supported by venture capitals and encouraged to found a company, which led
to creation of university start-up entrepreneurs who has the ability for identifying business
opportunities and creating a business concept (Etzkowitz, 2008). In Japan, as the protective
measures were implemented for fostering the domestic computer industry, projects were
orchestrated to transfer technology from ETL to large companies based on the basic technology
which had already been successfully developed in the US (Takahashi, 1996). In this respect, as long
as Japan aims to catch up with the US, Japan faces little market risk, although with some
technological risk, thanks to the US market already in existence.
Technologies utilized for catching up with the US or technology transfer to large companies are
incremental technologies for which there is no need to worry about cannibalization or create a
business concept (Shane, 2004). Thus, in Japan under the protective policy, university start-up
entrepreneurs who bear the "twofold start-up risks" (Nishizawa, 2012) or achieve the "new-market
type disruptive innovation" (Christensen, 2003) were not necessary for a long time. That is, it was
not necessary for Japan, at least until the 1990s, to establish a system for cultivating university
start-up entrepreneurs.
8. Conclusion and future outlook
-27-
This paper has dealt with the case of commercialization of the transistor computer and
compared the eco-systems of Japan and the US for promoting the commercialization of new
technology.
As a result, two differences have been found between the eco-systems of the US and Japan: one
is whether or not a "framework for cultivating the ability as an entrepreneur who identifies business
opportunities and creates a business concept" is in place, and the other is whether or not "funds and
experimental fields for exploiting the business opportunity" is secured. The first difference, the
position of entrepreneurs in the eco-system, is based on the assumption that the US is
"entrepreneur-centric" while Japan is "entrepreneur-absent". As for the second difference, funds
and experimental fields, the US promoted "development of advanced technology for military
purposes", while Japan took the approach of "protecting the domestic computer industry". In this
regard, it was difficult for university researchers in Japan, one of the defeated countries of World
War II, to found a company, as commercialization of computer-related new technology could have
been regarded as commercialization of military technology (Hiroshige, 1973). Thus, Japan achieved
commercialization of new technology through cooperation between ETL and large companies
without involving the universities.
It is not the point of our argument, however, that the US's eco-system for promoting
commercialization of new technology should be introduced to Japan as a model to follow. In today's
Japan, it would be difficult to secure "funds and experimental fields", similar to those of the US, for
the purpose of military technology development.
When the transistor appeared about 60 years ago, application of mounting the transistor to the
computer was termed as "real-time processing" (Takahashi, 1996). In the US, the application was
control of flight simulators or air-defense missiles, while in Japan, it was the seat reservation system
of Japan National Railways. How can one secure "funds and experimental fields" in
commercialization of new technology? The answer should not be to follow the US as a model but to
develop an approach while building a social consensus in a manner suitable to Japan's situation.
-28-
For example, if in the context of "commercialization of disaster-prevention technology" instead
of "protection of the domestic industry", it would be relatively easier to obtain a social consensus. It
is known that the cleaner robot of iRobot Corporation, which is an MIT's start-up, is a spin-off from
research and development on military robots (4). Research and development of rescue robots for
Fukushima nuclear power plant which operate under extremely stringent conditions would fulfill
the purpose of protecting the workers from exposure to radiation as well as provide a fundamental
technology for rescue robots for counterterrorism, traffic accidents, or construction sites (5). For
commercialization of such robots, not only internal start-ups of large companies, but, from the
viewpoint of building a social consensus, university start-ups by university researchers should also
be involved. This is because action of designing an artifact inevitably involves the problem of value
which gives a social meaning to the artifact (Sakakibara, 2010; Sugiyama/Takao, 2011), and
requires "funds and experimental fields". In order to realize the "new-market type disruptive
innovation", it is necessary to develop a commercialization path also from universities which have
logics different from those of companies. When the technology or the eco-system is regarded as an
"artifact", a counter-concept of a "natural object" (Simon, 1996), it is essential to newly establish a
corresponding instrument (eco-system) in order to create a new value.
The limits of this study are that only the comparative cases of commercialization of the
transistor computer half a century ago have been described, and the generality of our findings
requires careful assessment. In the computer history, records show that, since appearance of the
transistor, Japan and the US have taken various measures for commercialization of new technology
(Ceruzzi, 2003; Takahashi, 1996). As mentioned above, commercialization of computer-related new
technology was initially positioned as the "cooperative project between the government and
universities" in the US, and as the "cooperative project between the government and large
companies" in Japan. Later, in the 1980s, the US launched a consortium including large companies
as a means to compete against Japan's "fifth-generation computer project" (Nishizawa, 2012). On
the other hand, since 2000, Japan has been officially addressing promotion of university start-ups
-29-
(Kanai, 2010).
However, the system for cultivating human resources of computer scientists, which was
established on the basis of successful experience at the industrial level, is path-dependent (Arthur,
1994), rigid, and difficult to change (cf. Kagono, 1998). In the US, excellent computer researchers
gathered mainly at three universities: MIT, Stanford University, and Carnegie Melon University
which were engaged in military research (Ceruzzi, 2003). Then, researchers at these universities
personally founded companies, further promoting accumulation of university start-ups (Etzkowitz,
2008). In Japan, by contrast, ETL exclusively gathered excellent students, and the typical career
path of the computer scientist is first experiencing national projects and being engaged for about a
decade in research, then being sent to a university to foster young people (Furukawa, 2010). This
system remains almost the same today, and the former ETL is now National Institute of Advanced
Industrial Science and Technology (AIST).
In order to further consider the issue of "why Japanese university start-ups are not successful in
starting a business", it is necessary to focus on former researchers with ETL in light of the computer
history. Kazuo Taki, a pioneer of university start-ups in the computer field in Japan, founded a
system LSI designing company and was a leading young researcher in the "fifth-generation
computer project" orchestrated by former researchers with ETL (Taki, 1993; Takase, 2012). The
parallel architecture of the fifth-generation computer he worked on has been passed down as the
basic technology of today's grid computer and parallel computer, and known as a world-class
research accomplishment. Taki has remarked that "today, more than 200 researchers who were
involved in the fifth-generation computer project are active at universities, but I have never heard of
anyone who addressed creation of a university start-up." (6) The problem of Japan's eco-system
including cultivation of entrepreneurs or researchers in the computer field will be clarified by
further studies focusing on moves of former researchers with ETL and the recent efforts of creating
start-ups by the AIST.
-30-
Endnotes
(1) The definition of the eco-system according to Iansiti, who was the first to officially introduce the concept,
was: " a network relation among companies compared to a biological eco-system (ecological system),
and a concept representing a state of mutual coexistence for the purpose of common prosperity and
survival of loosely-connected network participants" (Iansiti and Levien, 2004). However, for the
eco-system in the context of commercialization of new technology to be discussed in this paper, the
definition by Sugiyama and Takao (2011) as follows is more compatible: "an aggregation of agents who
contribute to realization of a new value system by development/production of artifacts". Therefore,
especially in order to stress the necessity of evaluation of robustness of the environment in which
university start-up entrepreneurs are to live, this paper has adopted the latter definition.
(2) See Venture review Vol. 13 (pp. 59-68).
(3) The reason why Ken Olsen and Shigeru Takahashi have been selected as representative cases is that they
were responsible persons for introduction of the transistor to the computer. Progress of the computer was
realized mainly through start-ups by university researchers (Ceruzzi, 2003), and one of the technologies
which created most important new-market disruptive innovation in the computer history is said to be the
transistor (Christensen, 2003). In this respect, this paper recognizes the transistor as the "basic technology
which (consequently) created the framework of the university start-up".
There are three reasons why DEC has been selected as a representative case of the university start-up: 1)
it was founded by a university researcher, 2) supported (funded) by a venture capital, and 3) went public.
There are other cases including Hewlett-Packard supported by Professor Frederic Terman of Stanford
University, but DEC is positioned as the origin of the university start-up in that the company prompted
creation of the venture capital (Robert, 1991; Etzkowitz, 2002; Shane, 2004).
As described later (5.1.2), it was MIT's vice-president Vannevar Bush who promoted military research at
university, and Terman of Stanford University was Bush's student. Like Bush, Terman is also known to
have promoted military research. Hence it is believed that Bush and Terman have established the
-31-
foundation of MIT and Stanford University which are today called entrepreneurial universities (Robert,
1991; Etzkowitz, 2002; Shane, 2004; Ueyama, 2009; Nishizawa, 2012).
ETL has been selected as the subject for comparison with DEC, because the technological condition, the
transistor computer, was the same, and technology transfer from ETL to large companies was a means for
Japan to realize commercialization of new technology. MIT Lincoln Laboratory in the US and ETL
before World War II in Japan both have a history of being involved in military technology. Both
institutes gathered excellent students from across the country or universities. Their career path is to have
a certain degree of experience at the laboratory and then assume a post at university for fostering young
people, and the both institutes had a function of cultivating human resources of researchers. On the basis
of these circumstances, it should be noted that the focus of this study is not on the university-based
business start-up, but on the eco-system for promoting commercialization of new technology.
(4) For details, see iRobot Corporation's website:
http://www.irobot-jp.com/irobot/ (Date of access: February 25, 2012)
(5) For details, see the website of International Rescue System:
http://www.rescuesystem.org/ (Date of access: February 25, 2012)
(6) From an interview dated July 26, 2011
References
Aiso, Hideo, et al. (Ed.) (1985) This is how the domestic computer was created: history of development from
1960 to 1985 and outlook for the new generation, Kyoritsu Press
Aida, Yutaka (1991) Memoir of NHK and electronics-oriented Japan (the first volume), NHK Publishing
Arthur, W. Brian (1994) Increasing returns and path dependence in the economy, University of Michigan
Press
Bell, C. Gordon and Allen Newell (1971) Computer Structures: readings and examples, McGraw-Hill
Ceruzzi, Paul E. (2003) A History of Modern Computing, 2nd edition, MIT Press
Chesbrough, Henry W. (2003) Open Innovation: The new imperative for creating and profiting from
-32-
technology, Harvard Business School Press
Christensen, Clayton. M. (1997) The Innovator's Dilemma: When new technology cause great firm to fall,
Harvard Business School Press
Christensen, Clayton M. and Raynor, Michael E. (2003) The Innovator's Solution: Creating and sustaining
successful growth, Harvard Business School Press
Endo, Satoshi (2010) It is the Japanese who created the computer! Ascii Media Works
Etzkowitz, Henry (2002) MIT and the Rise of Entrepreneurial Science, Routledge
Etzkowitz, Henry (2008) The Triple Helix: University-industry-government innovation in action, Routledge
Furukawa, Koichi (Ed.) (2010) Kazuhiro Fuchi – personality and computer science, Kindaikagaku-sha
Gupta, Udayan (2004) The First Venture Capitalist: Georges Doriot on leadership, capital, and business
organization, Gondolier
Hiroshige, Tetsu (1973) Social history of science (the first volume) war and science, Chuokoron-sha
(reprinted edition, Iwanami Gendai Bunko, 2002)
Iansiti, Marco and Roy Levien (2004) The keystone advantage: what the new dynamics of business
ecosystems mean for strategy, innovation, and sustainability, Harvard Business School Press
Imai, Kazuhiko (1989) My ambitions, the fifth-generation computer – Kazuhiro Fuchi and ICOT's
technology-champions, Hankyu Communications
Kagono, Tadao (1988) Organizational epistemology, Chikura Publishing
Kagono, Tadao (2009) Communal society and university start-up, Journal of the Japan Academic Society for
Ventures and Entrepreneurs, Vol.13, pp. 3-9
Kanai, Kazuyori (2010) Comparison of university start-ups between Japan and South Korea, Chuokeizai-sha
Kanter, Rosabeth Moss (1983) The change masters: innovation for productivity in the American corporation,
Simon and Schuster
Kaplan, Eugene J (1972) Japan, the government-business relationship: a guide for the American
businessman, U.S. Dept. of Commerce, Bureau of International Commerce
Kirihata, Tetsuya (2010) Japanese university start-ups: industry-government-academia innovation at a
-33-
turning point, Kyoto University Press
Kunda, Gideon (2006) Engineering Culture: Control and commitment in a high-tech corporation, Revised,
Temple University Publishing
Levy, Steven (1984) Hackers: Heroes of the computer revolution, Doubleday
Narisada, Kaoru (1995) Establishment of physics division and electronics division, 100-year history of
Electrotechnical Laboratory, pp. 740-746, Ohm-sha
Nishizawa, Akio (2012) NTBFs and start-up companies (Nishizawa, Akio, Kutsuna, Kenji, Hibara,
Nobuhiko, Saburi, Masataka, Wakabayashi, Naoki, and Kanai, Kazuyori, Local eco-system for creating
high-tech industry, Yuhikaku. pp. 15-37)
Ogura, Miyako (2011) A 2010 research on university start-ups ― current situation of founding of start-ups
and awareness about industry-university cooperation for supporting start-ups, based on a questionnaire
survey on universities, etc. ―, National Institute of Science and Technology Policy, Ministry of
Education, Culture, Sports, Science and Technology, research documentation, No.200
Patterson, David A. and John L. Hennessy (2003) Computer Architecture: A quantitative approach, 3rd ed.,
international student ed., Morgan Kaufmann Publishers
Rifkin, Glenn and George Harrar (1988) The Ultimate Entrepreneur, Contemporary Books
Roberts, Edward B (1991) Entrepreneurs in high technology: lessons from MIT and beyond, Oxford
University Press
Sakakibara, Kiyonori (2010) Study on "Artifact and its value", Organization science, Vol. 44, No. 1, pp.
26-33
Schein, Edger H. (1969) Process Consultation: Its role in organization development, Addison-Wesley
Schein, Edger H. (1985) Organizational Culture and Leadership: A dynamic view, San Francisco:
Jossey-Bass Publishers
Shane, Scott (2004) Academic Entrepreneurship: University spinoffs and wealth creation, Edward Elgar
Publishing
Shane, Scott (2012) Reflections on the 2010 AMR Decade Award: Delivering on the promise of
-34-
entrepreneurship as a field of research, Academy of Management Review, Vol. 37, No. 1, pp. 10-20
Simon, Herbert A. (1947) Administrative Behavior: A study of decision-making processes in administrative
organization, Macmillan
Simon, Herbert A. (1996) The Science of the Artificial, 3rd edition, MIT Press
Special Committee of the History of Computing, Information Processing Society of Japan (Ed.) (1998)
History of Japanese computer development, Ohm-Sha
Sugiyama, Yasuo and Takao, Yoshiaki (2011) Boundaries of eco-system and its dynamism, Organization
science, Vol. 45, No. 1, pp. 4-16
Takahashi, Shigeru (1976) Transistor computers (ETL Mark III to VI), Information processing, Vol. 17, No.
2, pp. 133-141
Takahashi, Shigeru (1996) Computer chronicle, Ohm-sha
Takahashi, Shigeru (1998) Origin of the computer in Japan (Special Committee of the History of Computing,
Information Processing Society of Japan (Ed.) History of Japanese computer development, Ohm-sha, pp.
27-39)
Takahashi, Shigeru (2003) Ministry of International Trade and Industry and Japanese computer
manufacturers, Information processing, Vol. 44, No. 10, pp. 1069-1077
Takahashi, Shigeru, Hiroo, Mitsuru, Fuchi, Kazuhiro, Yamada, Akihiko, and Wada, Eiichi (2010) Oral
history/Interview with Hiroshi Wada, Information processing, Vol. 51, No. 7, pp. 870-877
Takase, Susumu (2012) Development of academic entrepreneur's expertise: a case of Taki Kazuo's
pioneering university start-up in Japan, The Rokkodai ronshu. Keieigaku-hen, Vol. 59, No. 1, pp. 1-22
Taki, Kazuo (Ed.) (1993) Parallel processing of the fifth-generation computer: path to general-purpose
parallel processing; language, OS, and programming, Kyoritsu Press
Toda, Iwao and Matsunaga, Toshio (2003) Computer development by Denden-kosha, Information
processing, Vol. 44, No. 6, pp. 631-639
Ueyama, Takahiro (2010) Beyond the academic capitalism: American universities and current scientific
research, NTT Publishing
-35-
Ukai, Naoya, Uraki, Tsuneo, Matsunaga, Toshio, and Yamada, Akihiko (2011) Oral history/interview with
Hiroji Nishino, Information processing, Vol. 52, No. 12, pp. 1554-1560
Uraki, Tsuneo, Odaka, Toshihiko, and Kawabe, Shun (1985) Hitachi, Ltd., (Aiso, Hideo, et al. (Ed.) This is
how the domestic computer was created: history of development from 1960 to 1985 and outlook for the
new generation, Kyoritsu Press. pp. 95-122)
Utterback, James M. (1994) Mastering the Dynamics of Innovation: How companies can seize opportunities
in the face of technological change, Harvard Business School Press
Venkataraman, Sankaran, Saras D. Sarasvathy, Nicholas Dew, and William R. Forster (2012) Reflections on
the 2010 AMR Decade Award: Whither the promise? Moving forward with entrepreneurship as a science
of the artificial, Academy of Management Review, Vol. 37, No. 1, pp. 21-33