How DENDRAL was conceived and born

ABSTRACT

As agreed with your organizers, this will be a somewhat personal history. They have given me permission to recall how I came to work with Ed Feigenbaum on DENDRAL, an exemplar of expert systems and of modeling problem-solving behavior. My recollections are based on a modest effort of historiography, but not a definitive survey of and search for all relevant documents. On the other hand, they will give more of the flow of ideas and events as they happened than is customary in published papers in scientific journals — accounts so dry that Medawar lugubriously calls them fraudulent {43}; cf. Merton & Zuckerman {44, 45, 61}. These authors point out that the standard scientific publication is narrowmindedly devoted to the context of justification. The DENDRAL effort (along with much of medical informatics) is dedicated to discovery: should we use a different standard for its history?

I hope it will be eventually possible to divert my colleagues from the more important work they do from day to day, and join me in a larger effort at historical research and informed consensus. My account is inevitably incomplete, especially about what others were thinking at a given moment. Built into the phenomenon of history, as soon as enough time has passed to enable some detached judgment, the evidence becomes frail, and we become vulnerable to the myths we create. Understanding all of these limitations, I will no longer qualify every remark: it should be implicit that each is “to the best of my recollection / or / as best as can be inferred from the fragmentary documentary record”.

I will assume you are generally familiar with DENDRAL, and will concentrate mainly on material not found in the published papers, especially as there is a comprehensive synopsis {41}.

As computer science is not my primary profession, my relationship to it has been more episodic; and I can more readily isolate how I came to take some part in it, at Stanford from 1962 - 1978, mainly in very close collaboration with Ed Feigenbaum, Bruce Buchanan, and a host of others. My central scientific commitments have been to molecular genetics, starting when I was a 20-year old medical student in 1945 {38}. At Columbia and then at Yale, I worked on the genetics of bacteria, a specialty which converged with the role of DNA as genetic information. My first academic appointment was at the University of Wisconsin from 1947 - 1958; then I went to Stanford in 1959 to take part in the reconstruction of its School of Medicine (formerly in San Francisco) at the Palo Alto campus. My role was to found a new Department of Genetics; I had no plan to be working with computers. In fact, I met Ed Feigenbaum in 1963. Then, promptly after he moved from Berkeley to Stanford faculty in early 1965, we initiated the collaboration that became the DENDRAL project.

References

1. Lederberg, g. 'Dendral- 64 A System for Computer Construction, Enumeration and Notation of Organic Molecules as Free Structures and Cyclic Graphs". Interim Report to the National Aeronautics and Space Administration, December 15, 1964. (Available from either author of this memo).
2. Lederberg, J. "Topological Mapping of Organic Molecules". Proc. Nat. Acad. Sci. 53:134-139, 1965. (Available from either author of this memo).

Index Terms

• Applied computing

  • Life and medical sciences

    • Computational biology

    • Genetics

    • Systems biology

• Human-centered computing

  • Human computer interaction (HCI)

• Social and professional topics

  • Professional topics

    • History of computing