"Computer Science at the CNRS and in French Universities: A Gradual Institutional Recognition" (Georges-Louis Baron & Pierre Mounier-Kuhn) (original) (raw)

Computer Science at the CNRS and in French Universities: A Gradual Institutional Recognition

GEORGES-LOUIS BARON
PIERRE-E. MOUNIER-KUHN

Abstract

This article describes a case of institutional innovation-the development of a new activity among established scientific disciplines. The official recognition of computer science in the French research and educational organizations met many obstacles, at least until the mid-1970s, although research and teaching had been carried on in this field for 20 years. In the CNRS, informatics emerged as a subdiscipline of applied mathematics and fluid mechanics, then of mathematics, before a specific assessment committee was created in 1976 for computer sciences and automation. In the university system, a subcommittee was established in 1972 and became a full committee in 1982; at that time, computer activities had already penetrated the secondary schools.

Categories and Subject Descriptors: K. 2 [Computing Milieux]: History of Computing—people, systems. K.3.2 [Computing Milieux]: Computers and Education-computer and information science education.

General Terms: Management, Standardization. Additional Terms: CNU, CNRS, University Assessment, Interdisciplinarity, Degree Programs, Theses, DGRST, IRIA, France, Education.

Introduction

A characteristic feature of the history of computer science in France has been its difficulty in becoming accepted as a major field by both the academic world and industry. This article examines how computer science has penetrated fundamental research and higher education, especially in two of the “central” institutions of this community: the Comité National of the Centre National de la Recherche Scientifique (CNRS) and the Conseil National des Universités (CNU).

The Comité National comprises all the CNRS topical evaluation committees; its role is to eval-

[1]uate the scientific quality of its laboratories and researchers, and to define the budgets for laboratories and the assignments of its full-time staff. It was created in 1945 and now consists of 45 sections (or commissions) each with a membership of approximately 20 scientists, and each representing a distinct scientific discipline. It can be compared to the U.S. National Science Foundation committees, or to those of the Science and Engineering Research Council in Great Britain.

The Conseil National des Universités, whose name has changed several times (Comité Consultatif des Universités, Conseil Supérieur des Universités etc.), is the consultative body concerned with the professional careers of teachers in higher education.

There are two reasons why a subject such as computer science should be recognized by these

  1. Author’s addresses: Georges-Louis Baron, 40 rue Bezout, 75014 Paris, France. Pierre-E. Mounier-Kuhn, Centre STS, CNAM, 2 rue Conté, 75003 Paris, France. ↩︎

bodies as an autonomous discipline with its own section:

One of the key problems is to know when computer science became a recognized scientific discipline. There have been major, long-term obstacles to it becoming so considered:

[1]from university prejudice against the industry (reinforced by certain unfortunate experi-ments-Couffignal’s failure in 1951, Bull’s difficulties after 1962, etc). Computer science was confused with numerical analysis for a long time, one of the least prestigious branches of mathematics in the Bourbakian hierarchy (Dieudonné 1974; 1977).

Abstract

Georges-Louis Baron has a PhD in Educational Sciences from the Paris University RenéDescartes. He taught mathematics, then worked at the Ministère de l’Education Nationale, where he was in charge of studying the development of educational innovations using computers. His interest in the history of computing arose in the early 1980s, when he began to study the constitution of “Informatics” in France as a science and as a school discipline. He is a Senior Lecturer at the Institut National de Recherche Pédagogique.

img-0.jpeg

Pierre E. MounierKuhn received his Master’s degree in History at the University of Paris in 1981, while working in the publishing business and in journalism. The same year he joined the Centre National de la Recherche Scientifique (CNRS), where he was in charge of evaluation committees and of the board of administration. This led him to study problems of science policy, research assessment and interdisciplinarity, in the context of the history of CNRS (which celebrated it 50th anniversary in 1989). Mounier-Kuhn is preparing a doctoral thesis on “Informatics in France since 1945: History of Science and Industry Policy.” He is teaching as a research assistant at Conservatoire National des Arts et Métiers (Centre Science, Technologie et Société) and he is currently organizing a conference on the History of Computing, to be held at CNAM in April 1990.

  1. 1{ }^{1} Auguste Comte, French philosopher (1788-1857), Ecole Polytechnique graduate. He developed a theory, “positivism,” according to which mankind, having reached the industrial age, should be ruled by scientific laws (which would replace politics and religions). Comte established a general, linear classification of the sciences, by increasing complexity and decreasing universal value: Mathematics (the basis), astronomy and mechanics, physics, chemistry, biology and, on top, sociology (a word invented by Comte). Comte, who was a professor at the Collège de France, had a great influence on the development of sociology, epistemology, history of civilizations as well as on the organization of policies concerning industry, technology, and science. ↩︎

In short, for a large number of scientific decision makers, computer science could be summed up as merely programming calculations on a computer. At best it was “a technique that could be learned in three hours like glass-blowing,” in the worst case, it was “the dustbin of mathematics.”

Computer Science at the CNRS: a Three-Stage Emergence

In certain laboratories, however, research in computer science developed at the end of the 1950s. Although it was often handicapped by the over-empirical aspect of “rules of thumb” in programming, or by overly ambitious projects in machine translation and machine recognition, it nevertheless produced a large body of solid theoretical work. In the second half of the 1960s, computer science had become a full fledged scientific discipline in France. 2{ }^{2}

When and how did the CNRS recognize computer science as an autonomous research field among its representative instances on the Comité National? Three stages can be distinguished:

  1. The emergence of computer science within general mechanics and applied mathematics (section 3 of the CNRS) between 1945 and 1966,
  2. Its integration into pure mathematics (section 1 of the CNRS) between 1967 and 1975, and
  3. Its institutional recognition in 1976 when a “computer science, control engineering, systems analysis and signal processing” commission was formed.

In the first period, (1945-1966), computer science emerged slowly under the protection of openminded leading scientists. It functioned as a minority activity inside the “General Mechanics and Applied Mathematics” section. In 1956, Jean Kuntzmann and Lucien Malavard, two representatives of the field of electronic calculation, were appointed to the Comité National for a first term, soon followed by others, including René de Possel and François-H. Raymond. In 1964, the Comité

[1]National included 12 computer science representatives or “defenders,” and at that time they wrote a position paper (CNRS Rapport de Conjoncture 1963-1964), which emphasized the existence of computer science as a scientific research discipline quite separate from the activities of the computing centers.

Thus, the problem of the epistemological and institutional status of computer science arose coincidently with the CNRS undertaking the reorganization of the Comité National so as to adapt it to the evolution of science and to the growth of personnel in the various disciplines. This reorganization, including elections and appointments, was to take place during 1967. It was clear that computer science could no longer remain under the same roof as Fluid Mechanics, since these two quite different disciplines stifled each other when grouped together in one under-financed section. What was to be done with computer science? The director general of the CNRS consulted with members of the different commissions who suggested two quite opposite solutions:

  1. The creation of a computer science section to include control engineering and various branches of applied mathematics. The majority of the computer scientists, pure mathematicians as well as the concerned state bodies (the DGRST 3{ }^{3} and the Commissariat General au Plan) recommended this solution; or
  2. The grouping together of computer science and pure mathematics in one section. This was suggested by certain highly influential mathematicians, researchers in mechanics and by physicists. The idea was to carefully select “researchers in numerical analysis” (computer scientists as such were not even mentioned), to protect pure mathematics from “drying up” completely as a result of an eventual total isolation and to respect the principle of the fundamental unity of mathematics.

The choice in favor of the second solution was made by two leading mathematicians, André Lichnerowicz and Pierre Lelong, in March 1965 during a series of meetings at which the following arguments were put forward:

The very term “applied mathematics” is to be rejected, and

[2]

  1. 2{ }^{2} Active teams in this field before 1970 were notably: The Institut Blaise-Pascal, the Picard and Lions laboratories in Paris: the Centre d’Etudes pour la Traduction Automatique and the IMAG in Grenoble: the Centre d’Informatique in Toulouse; the Centre de Recherche en Informatique de Nancy (CRIN). ↩︎
  2. 3{ }^{3} Délégation Générale à la Recherche Scientifique et Technique. ↩︎

Research in “computer science” is “still an uncertain discipline.”

The director general of the CNRS and the Minister of National Education had no choice but to underwrite the creation of the new mathematics section of the Comité National, and the word “Informatique” did not even appear in the official title of the section before 1971.

Computer science, therefore, lived alongside pure mathematics for nine years, from 1967 to 1975. Its representatives formed a strong minority (over a third) of the new “section 1,” and they were supported by two successive section presidents, Jacques L. Lions and Claude F. Picard. This situation offered certain, especially qualitative, advantages. The mathematicians selected the computer scientists on more strict and demanding criteria (than before 1967)-at least from the mathematicians’ point of view. Fruitful contacts developed between computer scientists and pure mathematicians.

However, the differences between the basic concepts and the evaluation criteria of the two communities 4{ }^{4} led to low recruitment levels in computer science. According to the commission session reports that we have examined for the 1969 to 1975 period, this section recruited a yearly average of 19 researchers over six years, six of whom were “computer scientists.” The same order of magnitude and the same proportion (one third) could be found in the previous general mechanics and applied mathematics commission of 10 years earlier. A very low figure, given that the Plan Calcul designed to help France catch up on lost time, was in full swing, and especially when part of this research staff would only stay at the CNRS for a few years.

Beginning in 1971, the CNRS reaction was more effective. Perhaps this was due to the competitive pressure from the university or from the IRIA (Institut de Recherche en Informatique et

[1]en Automatique created in 1966). Opinions on the matter differ. However in 1971 Hubert Curien, the director general of the CNRS, appointed a new scientific director, Wladimir Mercouroff, who was given the special job of developing computer sci-ence-a task he had already been performing for a year at the Ministry of National Education. He was to carry out his program at the CNRS along three lines:

  1. by launching “programmed thematic activities” (ATP)—a method of financing research that by-passed the Comité National-especially in theoretical computer science,
  2. by establishing a separate budget for computer science research, and
  3. by creating a commission in 1975-1976 of the Comité National for “Informatique, Automatique, Analyse des systèmes, Traitement du signal” (Computer science, control engineering, systems analysis, and signal processing) as part of the newly formed “engineering sciences” department. 5{ }^{5}

This new commission has remained in place with an almost unchanged title since 1976, and appears to represent an “adult” phase of computer science research even if it suffers from such usual lapses in French computer science as the under-representation of the discipline of computer architecture, for example. In being given priority in the CNRS funding for staff openings and credits to the engineering sciences, it has helped the CNRS catch up on its backwardness in this field. The number of A rank researchers (senior researchers or research directors) increased from 9 to 50 between 1975 and 1986. In 1986, research in computer science and control engineering cost about 1 billion FF ( 140M140M\140M 140 \mathrm{M} ), of which higher education and CNRS (either its own units or those associated with the CNRS) received 500 M FF ( 70M70M\70M 70 \mathrm{M} ). Of the 3,500 people working in this sector, 2,500 belong to the CNRShigher education body; about 300 are CNRS engineers and researchers in computer science stricto sensu. (Verjus 1986, p. 10).

The tribulations of computer science inside the CNRS clearly reflect the successive directions

[2]

  1. 4{ }^{4} This kind of relationship-both difficult and necessarybetween mathematics and computer sciences was experienced, too, in the United States. The following quotation sounds very familiar to someone who studied the emergence of computer sciences at the CNRS: “The high standards and norms of scholarship of the older sciences have served computer science well by guiding its early developments. At the same time, we have also blundered extensively by assuming either that computer science must behave very much like the older experimental sciences or that it must follow the norms of pure mathematics.” (Juris Hartmanis 1981, “Observations about the Development of Theoretical Computer Science.” Annals of the History of computing, vol. 3 no. 1 p. 50). ↩︎
  2. 5{ }^{5} The circumstances of the creation in 1975 of the CNRS “engineering sciences” department can be compared with the formation, at the US National Science Foundation, of the “RANN” program (Research Applied to National Needs) in 1971. ↩︎

taken by this body. From being a rather tightknit establishment with its own laboratories, the CNRS opened itself to the universities in 1966 by developing an arrangement of associated laboratories. Later in the 1970s, when it was confronted by budgetary problems, it tried hard to demonstrate its socioeconomic usefulness by extending its interests to the world of industry and creating the “engineering sciences” sector which included the new Computer Science section.

Computer Science at the Universities: A Growing Awareness

At the end of the 1950s, computer science as such did not exist, and only by emerging little by little from the framework of other disciplines, (principally mathematics) did it achieve any legitimacy. It is possible to analyze how the university as an institution has become aware of this process, first by examining the teachers’ point of view and then the students’.

Recruiting Teachers for Higher Education

Computer science, which at first did not exist as such on the Higher Universities Committee, became part of it in 1972, as the second subsection of the 18th section (applied mathematics). In 1982, a 24th section, entitled “Informatique” was created with two subsections, “Informatique Theorique” and “Informatique des Organizations Techniques.” Computer science can also be found as industrial computing: the second subsection of the second section (components, signals, and systems) is named: “Automatique, Informatique Industrielle, Signaux, Robotique.” As is the norm, institutional recognition in fact regularized an existing state of affairs, and university teaching in computer science had in fact been provided long before 1972.

The First University Teaching Programs

Courses in numerical analysis and applied mathematics had been established quite early in several universities. At the Institut Polytechnique de Grenoble, under Jean Kuntzmann’s leadership, courses in applied analysis had been organized as from 1948. In the 1960s, engineering graduate schools developed a “computing” (calcul) option which changed later into computer science options. The Calculations center of the

Ecole Supérieure d’Electricité was inaugurated in 1962 and in 1966 a computer science section was formed for third-year students. Under the instigation of Professor De Possel, who held the Chair of Numerical Analysis at the Faculty of Sciences in Paris, a university institute, the Institut de Programmation was founded officially in November 1963, although it had been in the process of being established since 1961, followed by the creation in 1962 of a Diplôme d’Etudes Supérieurs Techniques (a graduate diploma) with a specialty in programming.

Created with professional training in mind, this institute accepted students without a high-school diploma (baccalauréat), but took the then bold step of making them take a probationary examination after one month of courses. Diplomas at three different levels were awarded: operator-programmer (until 1971), advanced programmer, and then from 1965, information processing expert. The institute also organized the Masters’ Degrees in Computer Science at the Faculty of Sciences in Paris, and later at the University of Paris VI. As Figure 1 shows, a large number of students studied there. In 1972, the first level of training disappeared, while there was a high intake of students until around 1970. The subsequent reduction in numbers is not of great significance insofar as the Masters programs were established and the Paris science faculty was split up into two universities located on the same campus (Paris VI and VII).

The Fouchet reform of 1966 instituted the Masters Degree in Computer Science (Maîtrise d’Informatique), which rapidly increased from 75
img-1.jpeg

Figure 1. Number of students at the Institut de Programmation de Paris; (一), information processing expert, ( ∼\boldsymbol{\sim} ), operator programmers, ( A\boldsymbol{\mathcal { A }} ), advanced programmers.

degrees awarded in 1968 to 174 in 1969 (M.E.N. 1971). In 1972 eight universities awarded Masters in computer science ( 600 degrees). Five universities were entitled to award the recently created Masters in Computer Science degree for Management (MIAGE) degrees, for which 485 students were enrolled. As a comparison, 10 years later 16 universities awarded Masters, and 1241 degrees were awarded ( 35%35 \% to women), including 25 Masters in Science and Technology, and 537 MIAGEs, according to the 1983/1984 statistics issued by the Ministry of Education.

Recognizing that there was a period of rapid growth prior to 1970, trends were much slower over the next 12 years up to 1982, and this relatively low rate of growth demands an explanation. Primarily the numbers given only represent those students in university programs who were awarded a Masters degree classed as “Computer Science” by the Statistics Department of the Ministry of National Education. Other such programs were offered in engineering schools, Instituts Universitaires de Technologie (I.U.T.), 6{ }^{6} and in other specialized institutes. Computer science appeared in a growing number of programs for noncomputer specialists, for example as minor certificates within other scientific Masters degrees. Yet the growth of the universities’ offer of training in computer science was limited because it was a new discipline where qualified professors were relatively rare.

The Development of University Research

The index of all the theses defended at French universities shows that research in computer science began in the 1960s in several centersmainly Grenoble, Nancy, Paris, and Toulouse.

Towards the middle of the decade, the first thèses d’Etat (postdoctoral thesis) appeared. In 1964, Jean Claude Boussard defended a thesis in applied science on the compilation of the ALGOL language at Grenoble, and in 1965 Claude Pair defended his postdoctoral thesis on the concept of the stack. These two works entailed a conceptual shift since they were no longer associated with numerical analysis but rather with algebra and

[1]recent theories in linguistics. Other theses include those of Louis Bolliet, Grenoble 1967. The first postdoctoral thesis listed under “Informatique” was that of J. Hebenstreit in 1969.

Toulouse holds a rather special place since a whole series of research projects were dedicated to the design and construction of a calculator named the CAT (Calculatrice Arithmétique de Toulouse). A first thesis related to this research was defended in 1961, and then at least two doctoral theses were defended in 1964, two in 1965, three in 1967, and at least one in 1968. The calculator ran efficiently for a dozen years.

Until 1969, the theses are indexed under “mathematics,” “applied mathematics,” “electronics,” “information processing,” or “control engineering” depending on the university. After 1970, doctoral theses with the word “Informatique” (Computer Science) in their titles begin to increase in number, no doubt because often the first students who began to study for a Masters in Computer Science degree in 1966 continued into doctoral studies in the same field. From 1972 on, the official index of doctoral theses includes a heading “Informatique.” A certain amount of diversification took place at the same time and doctoral thesis titles such as “practical computing” or “geological computer science” were presented. As for the universities that awarded them, three names, not surprisingly, dominate the field up until 1971-Grenoble, Paris, and Toulouse. Later on, Rennes, Lyon, Aix-en-Provence and others enter the scene (16 universities in 1984). Quantitatively speaking, there was at first a fairly rapid growth (approximately 20 theses defended in 1970, and 60 in 1973), which then slowed (approximately 130 in 1983). The growth rate of doctoral theses between 1973 and 1983 was of the same order as that found for the Masters degrees.

As for the contents of computer science teaching, programming has always played a major role. Ideas on the subject began to be established at the end of the 1960s in the surge of activity that followed the creation of ALGOL 60 up to ALGOL 68, and Pascal in the work of E. Dijkstra, D. Knuth, N. Wirth and, in France, with researchers such as J. Arsac, L. Bolliet, M. Nivat, and C. Pair.

As the CNRS supported the development of programmed thematic activities in computer science from 1972, new ideas gradually made themselves felt in university teaching. These included establishing a rigorous analysis of the problem to be solved before beginning to program it, dis-

  1. 6{ }^{6} The Instituts Universitaires de Technologie (IUT) were created by the Fouchet reform in 1966. 20 computer departments were established in the first six years; they trained 1200 students each year. This innovation had important consequences for the constitution of computer science as a recognized discipline in the higher education system. ↩︎

tinguishing between a program written in a particular programming language and an algorithm, so that the former could be considered as a translation of the latter into the language of implementation. Interest also began to be shown in proofs of programs. After a first period during which teaching programming meant above all teaching a specific language, these new ideas were applied to university computer science training as from the middle of the 1970s, and in papers given at scientific conferences in which the terms “programming,” “methodology,” and “algorithmic methods” are often included. For example, in 1976, the congress organized by the AFCET (Association Française pour la Cybernétique Économique et Technique) (AFCET 1977) included an “Algorithmic procedures and programming” workshop in which the majority of contributions had to do with teaching algorithms and programming, even though the meaning of the two terms was not unanimously agreed upon. Among the contributions, M. Lucas and P. C. School wrote:

During recent years, teaching of programming has considerably progressed as a result of methodological results obtained in the general field of programming . . . various approaches for this teaching have been put forward recently . . . all characterized by a distinction between algorithmic procedures and programming. Although the meaning of these terms varies from one author to another, they share the common feature of integrating “theoretical bases” (concepts and methods) into teaching.

This, then, is the period when analytical methods were devised, tested, and taught. 7{ }^{7} The approaches varied, according to the university center where they were created, but they shared the common aim of encouraging the teaching of programming. This direction in teaching methodologies for programming continued and the second half of the 1970s saw the building up of a consensus around which concepts should be taught in programming and which methods should be developed.

Conclusion

This short survey calls for a series of remarks which express the authors’ differing opinions. On

[1]the one hand, this slice of history can be summed up as a series of delays:

On the other hand, it is important to note that once the movement had begun in France, it was supported by a real political engagement that was never abandoned. It should be emphasized that the CNRS and the universities form a solid grouping and that computer science was taken into consideration by these institutions in a co-

  1. 7{ }^{7} One of the very first to be formalized was the deductive method, presented in 1973 by C. Pair. ↩︎

ordinated way, especially when the Government authorities began to be interested in the development of computer science towards the middle of the 1960s.

It is even possible to assert that the recognition of computer science by higher education institutions was a fairly rapid occurence, given the particular conditions existing in France and the manner in which public administrations work. This can be linked to a willingness on the part of the politicians, expressed by an interest in training in computer science: the Masters in Computer Science degrees were created in 1966, the first doctoral theses listed under “Informatique” were defended in 1969, and a CNU Computer Science subsection was set up in 1972. In the 1970s, there was a steadily growing interest in teaching programming and computer science began to be gradually introduced at the secondary school level (Baron 1989).

Acknowledgements

We would like to express our gratitude to all the persons who granted us interviews and gave us criticisms and advices, particularly Jacques Arsac, Claude Benzaken, André Berroir, Jean Coulomb, André Danzin, Michel Debré, François Genuys, Paul Germain, Jacques Hebenstreit, Viviane Isambert-Jamati, Pierre Jacquinot, Georges Jobert, Daniel Kayser, Jean Kuntzmann, Jean-M. Laborde, Jean Lagasse, Gabriel Langouet, Robert Latteš, André Lichnerowicz, Jacques-L. Lions, Lucien Malavard, Pascal Maroni, Wladimir Mercouroff, Maurice Nivat, Claude Pair, Jacques Pitrat, Pierre Poulain, Antoine Prost, François-H. Raymond, Jean-L. Rigal, Jean-Jacques Salomon, Geneviève Schmeder, Marcel-P. Schutzenberger, and Jacques Vélu.

References

Abir-Ham Pnina, G. 1988. “The assessment of interdisciplinary research in the 1930s: the Rockefeller Foundation and physico-chemical morphology” Minerva, 1.
Chatelin, P. (éd.) 1988. Actes du colloque sur l’histoire de l’informatique en France, Grenoble, INPG.
AFCET. 1977. Panorama de la nouveauté informatique en France, Paris, AFCET.
Arsac, J. 1970. La science informatique, Paris, Dunod.
Baron, G. L. 1985. Compétences secondes et trajectoires d’enseignants du second degré, le cas de l’informatique, Paris, DEA thesis, Université René-Descartes.

Baron, G. L. 1989. L’informatique, discipline scolaire? Le cas des lycées, Paris, P.U.F.
Ben-David, J. 1968. La recherche fondamentale et les universités, Paris, OECE.
Book, R. V., S. A. Greibach, J. Hartmanis, and S. C. Kleene. 1981. “The development of Theoretical Computer Science.” Annals Hist. Comput, 3, pp. 1360.

Bourdieu, P. 1976. “Le Champ scientifique.” Actes de la Recherche en Sciences sociales, 4.
Bourdieu, P. 1984. Homo Academicus, Paris, Editions de Minuit.
Breton, P. November 1985. Les nouvelles technologies de l’information, études sur la naissance et le développement des technologies, Strasbourg, Doctorate thesis GERSULP, Université Louis-Pasteur.
Cahiers pour l’Histoire du CNRS, Paris, Editions du CNRS, 1989.
Campbell-Kelly, M. April 1982. “The Development of Computer Programming in Britain 1945 to 1955.” Annals of the History of Computing, 4.
Clark, T. N. 1971. “Le patron et son cercle: clef de l’université française.” Revue française de sociologie, 12.
CNRS. 1959-1985. Rapports de Conjoncture.
CNRS. 1947-1988. Rapports d’Activité.
Crozier, M. 1977. L’acteur et le système, Paris, Seuil.
Curien, H. September 1973. “L’avenir du CNRS.” La Recherche.
Department for Scientific and Industrial Research. 1946-1964. Annual Report, HMSO, London, MSO.
Dieudonné, J. 1977. Panorama des Mathématiques Pures: le choix bourbachique, Paris, Bordas.
Dieudonné, J. October 1974. “Orientation générale des Mathématiques pures en 1973.” Gazette des Mathématiciens.
Druesne, G. 1975. Le CNRS, Paris, Masson.
Durckheim, E. 1937. Les règles de la méthode sociologique, Paris, PUF.
ESRC-NSF. 1986. Rise and Fall of a Priority Field, Strasbourg, ESRC, Proceedings Symposium 22-24 Sept. 1985.
Gomis, A., J. Cottam, J. Wesselius, B. Lecuyer, G. Lemaire, and C. Barthelemy. 1971. “Eléments d’organisation de la recherche fondamentale en France.” CNRS-EPHE, Groupe d’Etudes et de Recherche sur la Science.
Groupe d’Etude sur le fonctionnement du Comité National de la Recherche Scientifique. 1973. Rapport Antoine Bernard. Paris, CNRS.
Greussay, P. 1989. “De la Musique à l’Intelligence Artificielle” Journal du CNRS, 6, p. 4.
Hebenstreit, J. 1985. “Informatique et enseignement.” La Vie des Sciences, comptes rendus de l’Académie des Sciences, série générale, tome 1, Paris, GauthierVillars.
Kuhn, T. 1962. The Structure of Scientific Revolutions, Chicago, University of Chicago Press.
Lemaine, G., R. Mac Leod, M. Mulkay, and P. Weingart. 1976. Perspectives on the Emergence of Scientific Disciplines, Paris, Mouton-Aldine.

M.E.N. 1971. Rapports sur les formations supérieures, Paris, Ministère de l’Education Nationale, La Documentation Française.
CCRST, 1976. Rapport sur la recherche universitaire, Paris, Ministère de l’industrie et de la recherche.
Mounier-Kuhn, P. E. 1987. Le Comité national et l’émergence de nouvelles disciplines au CNRS: le cas de l’informatique 1946-1976, Paris, DEA thesis, Centre Science, Technologie et Société, CNAM.
Mounier-Kuhn, P. E. 1988. “La résistible émergence de l’informatique au CNRS.” Paris, Ministère de l’Education nationale, Informatiques, 4.
Mounier-Kuhn, P. E. 1990. “La lente émergence d’une discipline: la recherche en informatique au CNRS 1946-1976.” Paris, Cahiers pour l’Histoire du CNRS, Editions du CNRS, 6.
NSF. 1959-1971. Annual Report, Washington D.C., National Science Foundation, Washington DC.
Papon, P. 1978. Le pouvoir et la science en France, Paris, Le Centurion éditions.

Pitrat, J. October 1985. “La naissance de l’intelligence artificielle” La Recherche, 170.
Ralston, A. 1976. Encyclopedia of Computer Science, New York, Van Nostrand & Reinhold Company.
Salomon, J. J. 1988a. L’écrivain public et l’ordinateur. Paris, Hachette.
Salomon, J. J. 1988b. “Technology and Democracy.” In From Research Politics to Social Intelligence. New York, MacMillan Press.
Shils, E. 1968. Criteria for Scientific Development. Public Policy and National Goals, Cambridge, MIT Press.
Université de Paris VI. 1973. Institut de programmation 1963-1973.
Verjus, J. P. 1986. Rapport sur la Recherche et la Formation supérieure en Informatique, Paris, Documentation Française.
Weinberg, A. 1967. Reflexions on Big Science, Oxford, Pergamon.
Wilkes, M. V. 1985. Memoirs of a Computer Pioneer, Cambridge, MA, MIT Press.