Research Note THE MECHANICS OF LEARNING BY DOING: PROBLEM DISCOVERY DURING PROCESS MACHINE USE ERIC VON HIPPEL AND MARGIE TYRE This article is cast in the mold of a relatively new interdisciplinary field called management of technology. Researchers who specialize in this field are interested in understanding the process of techno logical innovation, and in learning how to manage it better. Their primary disciplinary training tends to be in economics or organiza tional behavior, and they often have some engineering or scientific training and experience as well. Most are based in business schools, and their research insights are often influenced and enriched by close contact with R&D engineers and managers who are trying to execute the detailed, day-to-day process of developing new technol ogy-based products and processes. Researchers in management of technology often seek to frame studies that simultaneously address practical issues of concern to technology developers and more general questions ofinterest to aca demic colleagues. The study we report on here follows this patDr . von Hippel is professor of management of technology and Dr. Tyre associate professor of management of technology at the Sloan School of Management, Massa chusetts Institute of Technology'. Editor’s Note.—This article addresses a matter long of interest to historians of technology': how do designers of new technologies relate to the eventual users? The authors proceed from a scholarly context related to, but distinct from, the history of technology. Readers might compare this example of current work in the nascent field of management of technology with the language found in more familiar socio logical writings (e.g., social construction of technology, or actor-network theory), though von Hippel and Tyre argue different issues and use different language than sociological scholars. Theoretical considerations aside, the article stands on its own merits, and the authors’ conclusions are worth pondering on their own terms—in particular their observations on the inherently “unfinished” nature of technological design and the essential rather than accidental relationship between designers and users, both of whom modify the character of new technologies. Unpredictability, in this reading, is not a constraint to be overcome so much as it is a source of tension that fosters creativity.© 1996 by the Society for the History' of Technology. All rights reserved. 0040-165X/96/3702-0004$01.00 312 The Mechanics ofLearning by Doing 313 tern. The issue we explore of practical concern to R&D practitioners arises from the commonplace observation that when new products, processes, and services are introduced for the first time to their in tended use environments, things often go wrong. New software often has unanticipated “bugs,” machines that worked very well in the development laboratory promptly fail when first installed in a fac tory, and so forth.1 This phenomenon is a source of considerable concern and expense to both developers and users, and they want to ameliorate it if possible. The more fundamental issues we seek to explore here are these: why are some problems so difficult, even impossible, to anticipate before a new product, process, or service is deployed in its working environment, and how does actual, real-world use play a role in tech nical problem solving and innovation? We draw on, and hope to contribute to, the rich and complex discussion ongoing among scholars of the history and economics of technical change, who have long explored the generation and diffusion of technological knowl edge and its role in economic growth and development.2 1 Robert L. Hayes and Kim B. Clark, “Exploring the Sources of Productivity Differ ences at the Factory Level,” in The Uneasy Alliance, ed. K. B. Clark, R. L. Hayes, and C. Lorenz (Boston, 1985), pp. 425-58; Dorothy Leonard-Barton, “Implementation as Mutual Adaptation of Technology and Organization,” Research Policy 17 (1988): 251-65; Marcie Tyre and Oscar Hauptman, “Effectiveness of Organizational Re sponse Mechanisms to Technological Change in the Production Process,” Organiza tion Science 3 (1992): 301-21. 2Nathan Rosenberg, “Technological Change in the Machine Tool Industry, 18401910 ,”Journal ofEconomic History 23 (1963): 414-43; Merritt Roe Smith, HarpersFerry Armory and the New Technology: The Challenge of Change (Ithaca, N.Y., 1977); Nathan Rosenberg and Walter G. Vincenti, The Britannia Bridge: The Generation...