8 New Media History PATRICE FLICHY Many histories of information and communication Similarly, these new machines have been expected could be written: the history of institutions and to determine the organization of work (Friedmann, firms, the history of programmes and creative 1978), our leisure activities and, more broadly, our works, the history of techniques, and the history of ways of thinking (McLuhan, 1964) or even society practices and uses which, in turn, can be related to at large (Ellul, 1964). that of work and leisure but also to that of the public By contrast, other researchers such as the histo- sphere. All these histories are related to very differ- rian David Noble (1984) clearly show, through the ent fields of social science, and the information and case of automatically controlled machine tools, that communication sector is far too vast to present all there is no one best way, and that the effect of a these perspectives here. I have chosen to take as the technique cannot be understood without simultane- main theme of this chapter the question of relations ously studying its use and the choices made by its between ICTs and society. With this point of view designers. After the Second World War two alter- we are at the heart of a number of debates: debate natives were explored to automate production: on the effects of communication, which for a long record–playback (automatic analogue) machines time mobilized sociologists of the media; extensive and numerical control machines. The former debate on determinism among historians of tech- recorded the design of a part drawn by a human niques; and debate around the sociotechnical per- operator and then automatically produced copies. spective which has now been adopted by most The numeric machine, by contrast, did not need to sociologists of science and technology. memorize human knowhow; it was capable of pro- We shall focus on three points in particular: the gramming design and production. If the numeric launching and development of ICTs, their uses in a option triumphed, it was not because it was more professional context and, lastly, their uses in a reliable or easier to implement, but because it cor- leisure context. responded to the representations that designers and corporate managers (future buyers) had of auto- mated industrial production. INNOVATION IN ICTS Numerical control was always more than a technology Does Technology Drive History? for cutting metals, especially in the eyes of its MIT designers who knew little about metal cutting: it was a In many histories of computing the transistor, then symbol of the computer age, of mathematical elegance, the microprocessor, are considered to be the deter- of power, order and predictability, of continuous flow, mining elements. Behind this very common theory of remote control, of the automatic factory. Record– lies the idea that technical progress is inevitable and playback, on the other hand, however much it repre- globally linear. Electronic components and basic sented a significant advance on manual methods, technologies such as digitization are seen as deter- retained a vestige of traditional human skills; as such, in mining the form of the technical devices we use. the eyes of the future (and engineers always confuse NEW MEDIA HISTORY 137 the present and the future) it was obsolete. (Noble, providers and especially with the press, 20 per cent 1979: 29–30) of French households adopted the new service (Flichy, 1991).1 Studying Successes and Failures Boundary Objects One conclusion can be drawn from Noble’s study: there is never a single technical solution; as a Behind the questions of coordination of designers rule, several solutions are studied in parallel. The within a company or of partnerships with the outside, historian has to study these different solutions and lies the question of the mobilization of the different analyse both successes and failures (Bijker et al., parties concerned by innovation: R&D engineers, 1987). The RCA videodisk is a good example of marketing specialists, salespeople, repairers, partner failure. Margaret Graham (1986) showed that the companies (manufacturers of components, content television corporation RCA had everything it providers, etc.) but also users. In an interactionist needed to launch its new product successfully one approach, sociologists have used the notion of of the best US research laboratories, positive boundary objects. These are objects situated at the market research, support from the media and, lastly, intersection of several social worlds, which meet the access to a very large programme catalogue. needs of all worlds simultaneously. ‘They are objects Despite all these assets, RCA sold no more than which are both plastic enough to adapt to local needs 500,000 copies of its videodisk player in three and the constraints of the several parties employing years. Another technical system, the VCR, devel- them, yet robust enough to maintain a common iden- oped in Japan, was to dominate the market with a tity’ (Star and Griesemer, 1989: 393). A boundary resulting loss for RCA of $600 million. The com- object is the result of complex interaction between pany that had launched television in the US was the different actors concerned. This is the exact eventually bought out by General Electric and then opposite of the naive idea of innovation spawned by the French company Thomson. Thus, as many ready-made by the inventor’s mind. The history of computer manufacturers were to discover, a com- Macintosh clearly illustrates this point. Two comput- pany can pay very dearly for a failure. The main ers using the principle of graphic windows were pro- lesson that Graham draws from this case is that duced concurrently at Apple: Lisa, a failure, and technical or commercial competencies are not Macintosh, the Californian company’s leading enough if they are not properly coordinated. That machine. Lisa was designed in a rather cumbersome was the underlying cause of the videodisk’s failure. organizational frame, with a strict division of tasks As sociologists of technology have shown: ‘rather between teams. Macintosh, by contrast, was devel- than rational decision-making, it is necessary to talk oped by a small, tightly knit team in which choices of an aggregation of interests that can or cannot be made by individuals were always discussed collec- produced. Innovation is the art of involving a grow- tively. The software developers gave their opinions ing number of allies who are made stronger and on the hardware and vice versa. Moreover, the stronger’ (Akrich et al., 1988: 17; see also Latow, people in charge of building the factory and of 1987). This strategy of alliances was what enabled marketing and finance were also included in these France Télécom to successfully launch its videotex discussions. This continuous negotiation caused the (the minitel) in the early 1980s. A few months project to be amended more than once. A computer before the new medium was launched, this telem- that was originally designed for the general public atic project collided with a virulent combination of was eventually substituted for Lisa as an office media and political opposition. In fact, France machine. The simplicity of its use, imagined from Télécom, which at the time was still the French Post the outset, was one of this computer’s most attractive Office (PTT), wanted to create and run the whole features (Guterl, 1984). The Macintosh is a computer system on its own. It hoped not only to provide the situated on the boundary between hardware and network and terminals but also to install all the software, between the computer specialist and the information to be offered to the public on its own layperson.2 servers. Other European post and telecommunica- tions authorities had opted for the same schema. However, faced with intense opposition France Path Dependency Télécom backed down and decided to move from a closed to an open system in which any firm could While the perspective of boundary objects is very become a service provider. France Télécom simply useful for the study of specific cases, other pers- transported the information and took care of the pectives are needed to analyse more long-term billing (with a share of the revenue paid back to the phenomena. The path-dependency concept devised by supplier). Owing to partnerships with service economist and historian Paul David is particularly 138 TECHNOLOGY DESIGN AND DEVELOPMENT illuminating. Through the paradigmatic example of From the outset ARPANET was a highly decentral- the typewriter keyboard which has not evolved ized network which stopped at the university since its invention, David (1985) builds a model entrance. This technical architecture left a large that compares dynamic growth to a tree. At each degree of leeway to each computing site which branch the actors face a choice. This choice may could organize itself as it wished as regards hard- sometimes be linked to a minor element, but once a ware and software and could create its own local solution has been chosen by a large number of area network. The only constraint was the need to actors it becomes relatively stable. The outcome of be able to connect to an interface. These technical initial choices is relatively unpredictable but a time choices in favour of decentralization were also to be comes when a technique or an industrial mode of found in the organization of work needed to organization is imposed, with a resulting pheno- develop the network. Construction of the network menon of lock-in. was entrusted to a small company closely linked to IBM’s entry into the PC market is an interesting MIT. This company dealt with no technical prob- example of path dependency. Big Blue was aware lems posed by data exchange beyond the interface, that this was an entirely new market and that it considering that they were the universities’ respon- needed to find a specific mode of organization to sibility.