DEPARTMENT OF GENETICS                                                                                 (Dictated before my call)
School of Medicine

March 16, 1962

Dr. Eugene Garfield
Institute for Scientific Information
33 South l7th Street
Philadelphia 3

Dear Gene:

I really am very sorry that we didn't have more time together in New York to go through all of the points in your recent correspondence. I was somewhat at a loss to know how to deal with two people at one time and I appreciate the graceful way in which you took me out of a predicament.

1.    Dr. Bertram Brown of the National Institute of Mental Health called me up yesterday to ask about a rumor that Stanford was contemplating setting up a clearing house on information on mental retardation. There was, of course, some ultimate basis for the rumor but no real substance to it and it probably stems mainly from my own general interest in the problem. I urged Brown to put whatever political pressure was available to him behind a larger scale effort for medical research information as a whole but he seemed to feel that it was important to go ahead with some interim measures that could show some progress within the next year or two. I am rather frightened about the implications of this since we may end up having about eighteen different information agencies within the NIH itself, but there may not be a great deal we can do about it. I have no reason to believe that the NIMH has had access to any really thoughtful study of what they could do, and I think you might be able to do a national service by making an input of your own. In particular I think it would be constructive if you could get in touch with Dr. Brown or with Dr. Joe Douglas at the NIMH to get more information and to make some fresh proposals. For example, along some lines that we have already discussed, couldn't you rather quickly organize to get out a selective permuted title index to Current Contents that would cover the indicated areas of interest of the NIMH plus mental retardation. One approach to this would be 

a) to specify a number of journals that would be indexed fully and 

b)to establish a thesaurus of key words, either by personal intuition or by an analysis of the terms which appear in the indicated journals which would be used to select additional individual titles from the total literature covered by Current Contents.

An index of this sort computed, say monthly and accumulated semi-annually could, it seems to me, be gotten out very expeditiously and be a useful interim measure until the larger systems are put into effect. It would also have the advantage that the basic input data would not represent any very large multiplication of effort, but would be another specialized used of the general inputs that you are now developing anyhow. An important aspect of this is how long it would take to generate the program necessary to do the computer selections.

I really don't disagree with you about the relative significance of permuted title indexing, but I think it can play a definite role as part of a unified index in specialized areas especially. The detailed format of presentation of the index is, it seems to me, very important.

It would also be profitable to think how citation indexing would fit into this scheme, although I am not sure there is very much point in being highly selective about citations. That is the main value of SCI in its interdisciplinary and breadth of coverage, not the segregation of a specific subject area. However, it is not at all impossible that NIMH would be responsive to some explicit suggestions in this field. One might, for example, as a somewhat perverted use of SCI, think of adding to the sample for the permuted title index in mental health, in addition to the above stated criteria, titles of all additional articles which had a citation to any one of some dozen key journals in the field of mental health. This would hardly be a perfect mechanism but for a selective index might help to give some additional penetration into the unexpected manifestation of relevant articles in totally unpredictable places. There might also be some room here for a certain amount of personal editing and selection from the samples generated in this way.

2.     Can you program underlining of key words in the more elegant presentations of permuted title indexes?

3.     That discussion about the cost of replacing the / still fresh in my mind. I was very much Interested when Gil King mentioned some recent experimental work that he is doing on the adaptation of a Mergenthaler Linotape machine to rapid tape controlled printing. He is having his own problems at IBM in convincing the front office that something more esthetic than the high speed IBM printer would have a useful role, and I think he might be very amenable to some constructive help in illustrating the possibilities of a more flexible font with some relevance to your own printout problems.

4.     The more I think about the information problems of the science industrial complex out here in the Bay area, the more apt it seems to me it would be for you to investigate the possibility of establishing a centralized information service. There are a very considerable number of quite small companies which must be quite desperate for good technical information services and who could probably buy much more for the annual cost of a librarian by contracting with your agency than they could on their own hook. I may have some opportunity to explore some of the implications of this idea and if anything more relevant comes up I will, of course, let you know about it. Meantime, it might pay for you to give some thought to the kind of services that an ISI agency at Stanford Industrial Park might be able to organize. Itek has a couple of divisions out here which may not be completely irrelevant to the present point.

5.     You can tell me this figure faster than I can find it in my files - how many articles do you now index per week in Current Contents?

More later,