The Welch Medical Library Indexing Project Applies Computer Methods to Library Cataloguing and Research

At a meeting of the Medical Library Association that took place from June 24-27, 1952 physician and librarian Sanford Larkey reported on progress in the Welch Medical Library Indexing Project which had begun in 1949. This project was probably the earliest attempt to apply punched card tabulating in library cataloguing and information retrieval.

"The goal of the project, of which I was a member until its termination in 1953, was to develop computer-derived indexes to the scientific and medical literature. This mechanization of bibliographic information involved the use of IBM tabulating equipment designed for statistical analysis. The Welch project used standard punched-card machines to
prepare subject-heading lists for the Armed Forces Medical Library, the precursor to the National Library of Medicine (E. Garfield, "The preparation of subject-heading lists by automatic punched-card techniques," Journal of Documentation, 10:1-10, 1954)" (Garfield, "Tribute to Calvin N. Mooers, A Pioneer of Information Retrieval", The Scientist, Vol11, #6 (March 17, 1997) 9).

In Larkey's 1952 report there is a very interesting section which he called the "Psychology of Machines", which I quote:

"I think I should say something about 'machines' themselves at this point. Since we are using machines in all the major phases of our work, I should like to describe the machines we are using and just how we are using them. I will discuss the present status of each phase of our work primarily on the basis of the machine operations involved. Another reason for this approach is that we are have found in discussing our program with others, our use of machines seems either to interest or worry people more than any other feature.

"This brings me to what might be called the 'psychology of machines.' The very word 'machines' seems to do things to people. We hear talk of 'electronic robots,' as though they were some sort of 'men from Mars' who could take over all intellectual activities by merely pushing buttons. This sort of talk leads to excessive hopes or to inordinate fears and precludes objective thinking about the possible uses of machines. One should consider machines as practice adjuncts, as we do typewriters, 3 x 5 cards, and visible indexes. Machines are only doing very rapidly what one could do with his own eyes and brain if had all the time in the world to do it and wanted to do it. There is no magic about it.

"There is, however, a more valid psychological aspect to machines. Since machines operate on a strict yes-or-no principle, we must be rigidly exact in presenting a problem. Each step must be in the most precise logical form, since one rarely can stop to correct as one goes along. Each step must be gone over and over in relation to every other one. One has to think not once, but many times. Programming often takes almost as long as the machine operation itself, but the end result is still reached much more quickly than by manual operations.

"These strict limitations of machines have been very useful to us. They not only have tightened up our own thinking processes, but their application has emphasized many semantic inconsistencies in our terminology and classifications. So, perhaps there may be a good psychological side to machines." (pp. 33-34).

Larkey, "The Welch Medical Library Indexing Project", Bull. Med. Libr. Assoc., 41 (19530 32-40.