Books and Computers (from Billings to MEDLARS), 1964
Page  1

BOOKS and COMPUTERS (From BILLINGS to MEDLARS) By MARTIN M. CUMMINGS, M.D.* Reprints from Academy of Medicine Bulletin Vol. 11 No. 1 Mar. 1965

Page  2 BOOKS and COMPUTERS (From BILLINGS to MEDLARS) By MARTIN M. CUMMINGS, M.D. * Director of National Library of Medicine, Bethesda, Maryland. Man began to draw pictures of his environment about 25,000 years ago. Pictographic writing and formal language development evolved about 3,000 B.C. in Egypt, China and North America with the use of stone, clay, bark and skins as convenient media for carving and imprinting. The representation of ideas through the use of symbols and letters took place about 2,000 B.C., particularly in Mediterranean civilizations, where phonetic alphabets were developed to spell words for objects and actions. The cuneiform symbols of the Sumerians represent the earliest written language. These steps in communication history led to calligraphy and in turn to printing which has been the principal means of recording and disseminating knowledge for the past 500 years. The first dated book was produced in 1457 by Johann Fust and Peter Schoeffer, associates of Gutenberg, who printed religious materials and Bibles at Mainz in the mid-fifteenth century. Medicine adopted this modality of recording information soon thereafter by producing Abulcasis's "Liber Servitoris" in 1471, although it had been using laboriously produced manuscripts for many centuries with success. Ancient medical writings are attributed to Emperor Huang-Ti of China nearly 3,000 years B.C. The invention of the hand press was followed by the power press in the 18th century, and this in turn was followed by the typesetting machine invented by Mergenthaler in 1886. American medicine was blessed in 1865 with the appointment of Dr. John Shaw Billings as the first Director of the Surgeon General's Library. Trained as a surgeon in Cincinnati, he became this nation's foremost medical bibliographer largely because of the difficulties he had encountered earlier as a medical student writing his dissertation. After searching the libraries of Cincinnati, Philadelphia, and New York, he came to the conclusion that nowhere in the United States was there a complete medical library, and that the indexes of medical books and journals could not be relied upon as a guide to their contents. According to Rogers,' he wrote, "It was this experience which led me when a favorable opportunity offered at the close of the war, to try to establish, for the use of American physicians, a fairly complete medical library, and in connection with this to prepare a comprehensive catalog and index which should spare medical teachers and writers the drudgery of consulting ten thousand or more different indexes." He began publication of the monthly INDEX MEDICUS in 1879 and published the first volume of the Catalogue of the Library of the Surgeon General's Office in 1880. As early as 1876, he had submitted a specimen of a catalog to the Surgeon General and wrote in his letter of transmittal: "The use which is made of the collection by physicians from every section of this country, and the number of persons who consult it, either in person or by letter, is steadily, and of late, rapidly increasing; and I cannot doubt that if a sufficient number of the catalog, of which this is a specimen, be printed, and distributed to our medical writers and teachers, so that they may at their leisure learn what aid they can obtain in their researches, no collec

Page  3 tion of the government will be more used or be of more practical utility; that it will soon tend to elevate the standard of medical education, literature, and scholarship of the nation, and will thus indirectly be for the benefit of the whole country, since the general knowledge and skill of the medical profession become a matter of personal interest to almost every individual at some time during the course of his life.'[2] Although he had no authorization to do so, Billings called the Surgeon General's library the" National Medical Library" and saw to it that it served a national function. This interest in the development of a "National Medical Library" for the support of medical education and medical practice has been sustained in the century which followed Billing's original inspiration. In like manner, the relationship of the National Library of Medicine to the development of medical research in this country also stems from the farsighted actions taken by Dr. Billings when he argued that monies appropriated to the National Board of Health be spent for special scientific investigations. In a report to the Honorable John Sherman, Secretary of the Treasury, dated January 1, 1881, there is an account of what we believe to be the first grants for medical research awarded in the United States.3 The awards were made to universities and medical schools, including Johns Hopkins, Harvard, the University of Virginia, and Cornell. The subjects chosen for study included yellow fever, diphtheria, malaria, air pollution, sanitary control of water, and studies of the adulterations of food and drugs. It seems remarkable that it took another sixty years for the research grant to be rediscovered and that many of the problems selected for study in 1880 are still major areas of concern to this nation. The invention of the telegraph, telephone and radio contributed significantly to advancement of communications generally, but in the field of biology and medicine, the development of the electric calculator and its modern prodigy, the computer, are important instruments for advancement of medical research. In this stream of events, Dr. Billings appears again as a major contributor to the scientific revolution taking place in the late 19th century. In 1880, Dr. Billings, still Director of the National Medical Library, was appointed Consultant to the United States Census Bureau. In this capacity he played an active role in the development of vital and medical statistics. He is credited by his biographer, Dr. Fielding H. Garrison, and also by the Dictionary of American Biography with being the actual inventor of the modern tabulating machine. Dr. Billings suggested that the various statistical data of the living and the decedent might be recorded on a single slip or card by punching small holes in it, and that these cards might be sorted and counted by mechanical means according to any selected group of these perforations. According to Garrison, this suggestion was adopted by Mr. Herman Hollerith in the electrical accounting and integrating machines used by the United States Census. The Dictionary of American Biography added in its article on Hollerith, that he gave Billings the credit for originating the idea and offered Billings shares in the tabulating machine company he started, and which eventually grew into IBM. It is said that the present size of the IBM punched card results from Billings' suggestion that the U. S. dollar be used as the standard for its dimensions. Now, 85 years later, we are using this unique idea of Billings in a

Page  4 more sophisticated way to produce the INDEX MEDICUS and other bibliographic materials with our computer-based information storage and retrieval system called MEDLARS. More about that later. Libraries remain society's means of storing and retrieving the world's fund of knowledge. New libraries such as the one being dedicated here today are essential to fulfill the needs of students, teachers and practitioners of medicine. They obviously serve the information requirements of the researchers also. For all these the library serves a vital purpose and its vitality is something to be nurtured. Above all the library is a place to study, to learn and to produce creatively. It should not be created to function solely as a depository. Blake4 has commented on the use of libraries and books as follows: "The library is the historian's laboratory. A medical historical collection should be not a museum where one comes to admire the rare and curious, but a working center providing first and foremost the essentials for teaching and research: significant texts and translations in modern editions and as strong a reference collection as possible."4 "Historical collections are not just for classical scholars. Physicians who want to read Hippocrates, or find out what Sydenham really said about gout, will usually - and quite rightly and reasonably - look for a good modern translation. Medical students can rarely be expected to read the original text of Harvey's De Motu Cordis, but all may read with profit one of several translations available at modest cost. Even the specialist will prefer for most purposes an edition like Kenneth J. Franklin's, which gives both the original text and a translation, to the fabulously expensive first edition or very expensive second edition. It is fortunate indeed for libraries building up a research and teaching collection in history of medicine that a rapidly growing number of quality reprints, in hard cover and in paperback, is now being published."4 Libraries should not be considered the exclusive domain of scholars. It should be a major learning resource for students since the use of books by students remain a most important means of learning medicine. Each physician should develop his discrimination of books while exposed to them as a student. Wilens5 has discussed the relationship of students to books in an article entitled "Requiem for a Textbook." "It is the student who develops the most intimate relationships with textbooks, which represent for him the chief and dependable storehouses of medical knowledge. They protect him against ambiguous, iconoclastic, and speculative lecturers. They can serve in a gyroscopic fashion to help him maintain his balance when he is presented with newly hatched and unproved hypotheses. "During his student days, he may not develop feelings of affection or attachment to any of his textbooks. They constitute formidable obstacles to be overcome in the process of acquiring a medical degree. In later years, however, when he comes across one of his old books, the past may be suddenly recalled and he may regard it with sentimental feelings. This is particularly true of old texts in his own specialty. Knowing their past history, he can better evaluate their qualities and their contributions to medical education."5

Page  5 A student's view of medical books may be somewhat different. A medical student at Yale wrote6 "The most concentrated market for medical books is the medical student, and the factors which influence his choice are many. Perhaps the ideal way to select a book is to go to a book store or library, examine the appropriate volumes, compare their features, read some of each, and then decide. But obviously, the student, by the very fact that he is a student, has little ability to judge whether a book on a new subject is accurate, complete, and generally desirable. He needs help. "Fortunately, help is never very far away. Sometimes the student is spared the task of decision. On the first day of class a professor may strongly recommend a book. The student, after purchasing this 'best work available', may then have his teacher autograph it. But more commonly the professor, in the first session, distributes a list of 12 to 50 books with titles that sound very similar. Faculty members who prepare such lists believe each book treats one part of the subject well, and therefore all are valuable. The chances are good that the list will be read, somewhat less good that the books will be. Students believe that the law of diminishing returns applies after the third book (on the 'how many can-openers do you need' theory)."[6] Electronic computers emerged after World War II as devices to provide rapid and precise mathematical processing of numerical information. They have been used successfully for more than a decade for business, engineering and other purposes. Thus it seemed logical that their speed and power might be applied for the handling of other types of information. After much study and careful rapid and precise mathemat rapid and precise mathematical processing of numerical information. They have been used successfully for more than a decade for business, engineering and other purposes. Thus it seemed logical that their speed and power might be applied for the handling of other types of information. After much study and careful rapid and precise mathematical processing of numerical information. They have been used successfully for more than a decade for business, engineering and other purposes. Thus it seemed logical that their speed and power might be applied for the handling of other types of information. After much study and careful systems design computers have been adapted to perform certain library functions at the National Library of Medicine. The story of MEDLARS (an acronym for Medical Literature, Analysis and Retrieval System) starts with the fact that, as the world's largest biomedical library, the National Library of Medicine has the responsibility for the comprehensive indexing of the world's literature in biology and medicine. It is the only institution in the United States performing this awesome task. To insure that the world's biomedical community has bibliographic access to this literature, NLM since the days of Billings published subject indexes, of which the INDEX MEDICUS is best known. But because of the tremendous growth of the biomedical literature, the task several years ago became too large for the manpower and the limited degree of mechanization then being employed. MEDLARS was conceived to join the intellectual talents of trained literature analysts to the tremendous processing capabilities of an electronic computer. Trained literature analysts review each article to determine how it should be indexed. The data they develop are punched on tapes, input into the computer, and transferred to magnetic tapes for storage and retrieval. Each reel of tape, about 12 inches in diameter, holds approximately 25,000 bibliographic citations. MEDLARS does three principal jobs: First, it produces INDEX MEDICUS. The computer is used to process, i.e., compile and print a comprehensive author-subject monthly index to articles from more than 2,400 of the world's biomedical journals. Approximately 150,000 articles are reviewed and indexed annually. This

Page  6 process formerly required 22 days per month; with MEDLARS it is now being done in 5 days. Approximately 60% of the material in INDEX MEDICUS comes from journals published in foreign languages requiring a massive translation effort. Secondly, it prepares special recurring bibliographies. These are periodically issued lists of citations in specialized medical subject areas such as cancer, heart disease, mental health, and recent reports on drugs. These specialty listings are compiled at regular intervals from data in the computer store, and printed for distribution by government agencies and other national professional organizations working in the specialty fields. Finally, it produces citations on request which we call demand bibliographies. A major reason for computerizing the Library's indexing is to perform rapid machine searching of the computer's data files and provide answers to complex reference questions which cannot be effectively handled by existing printed indexes or catalogs. More than 200,000 articles have been analyzed and stored on magnetic tape since 1963. By 1969 the MEDLARS computer file will contain over 1,000,000 citations with an average of 10 subject headings for each article. A search of all these citations for answers to 75 different complex reference questions will take less than three hours. Reports summarizing information from the computer are printed out by two processes -the high speed printer, which operates at the speed of 21,600 words per minute, and the Graphic Arts Composing Equipment (GRACE), at 3,600 words per minute. GRACE is a revolutionary computer-driven photocomposer developed by the Library for its own needs but has been characterized by some as the most important development in printing since Gutenberg's invention of movable type. This system will now be decentralized by providing the magnetic tapes which contain this vast source of references to medical libraries throughout the country so they may share this power of storage and retrieval of the world's biomedical literature. It should be remembered however that MEDLARS is a new instrument of the research library, not a replacement for libraries. The development of new communications devices has not replaced older ones. The telephone did not replace the telegraph, radio did not replace the phonograph, television has not replaced the radio nor shall the computer replace books. Man's thirst for information and knowledge inspires him to add new means of communications to older ones. We retain the wheel while we develop jet engines and rockets; in like manner the printed word will be retained in the face of other technologic developments. Even computers print their output so man may read their calculations or findings. It is now possible to have computers communicate with each other. This development however results from man's urge to receive, store and transmit knowledge. Medical libraries will continue to participate in this process. That they will use computers to serve this function is no longer a matter for speculation; it is merely a matter of time. Like Billings, we must have the vision and energy to apply new ideas and techniques to assist those who are dedicated to the advancement of medical education, research and practice.

Page  7 REFERENCES 1. Rogers, F. B.: John Shaw Billings, 1838-1913. Libr. J. 88:2622, 1963. 2. U. S. Surgeon General's Office: Library. Specimen Fasciculus of a Catalogue of the National Medical Library. Washington, D. C.; Government Printing Office, 1876. 3. U. S. National Board of Health: Annual Report, 1880. Washington, D. C.; Government Printing Office, 188L 4. Blake, John B.: Books, Libraries, and Medical History. J.A.M.A. 188:263, April 20, 1964. 5. Wilens, S. L., M.D.: Requiem for a Textbook. J.A.M.A., Vol. 188:249, April 20, 1964. 6. Cassidy, Eugene P.: A Medical Student Looks at Medical Books. J.A.M.A. 188:281. April20, 1964.