2 
In reviewing several of his recent papers, however, one gains the impression 
that the universality of his concept is highly doubtful. He seems to be severely 
limited in his approach by his commitment to the 3X3 matrix. As a result, he 
ends up with some matrices that contain overloaded cells, and others that contain 
empty ones. He indicates that an empty cell represents a "not yet determined" 
subject or condition, but in some cases this is clearly a biological unlikelihood, 
and in those cases where the "not yet determined" holds throughout a row or 
column, the validity of the category itself is challengeable.--JAP 
TIME-SHARING IN MEDICINE 
COMPUTERS IN THE SERVICE OF MEDICINE-- Essays on current research and applications. 
Edited by G. McLachlen and R. A. Shegog. Vols. I and II. Oxford Univ. Press, 1968. 
Time-sharing appears only in the preface and in the predictive essays in 
these two volumes. Brian Flowers, writing the preface, says that "fast arithmetic 
was the first stage of the computer revolution; multi-access is the second." It 
is clear that the second revolution has not hit British hospitals, however, when 
one reads these essays. In every case where I could find time-sharing mentioned, 
it was as a future potential. Dale and Roberts (vol. 2, p. 83) are pessimistic 
about the use of teletypes by the nursing staff, and predict a high error level 
if the input must be typed by any and all users. Bennett and Holland (vol. 2, 
p. 127) touch very briefly on the potential storage of diagnostic information to 
be used for query by practicing physicians. James (vol. 2, pp. 157-167) discusses 
the development of multi-access systems in relation to the development of infor¬ 
mation systems, but regards it as in the future. Weir (vol. 2, 171-181) attempts 
to predict the next ten years in hospital computer use. He discusses a "national 
network of computers in medicine," (p. 178) but also (p. 179) says that "even by 
1977 the establishment of a large multi-access machine working solely within a 
hospital in the United Kingdom is by no means certain." It would seem that time¬ 
sharing is not'moving rapidly for this group of authors.--JAP 
AVAILABLE PROGRAMS 
9. NTERM--A numerical taxonomy program modified from the Smithsonian Institution 
program CLUSTR [see MUDPIE no. 2, program 8]; written in Extended 
BASIC for use on GE Mark-II T/S system. Performs: read-in of tri¬ 
angular or rectangular matrix; character standardization; computation 
of pair-wise estimates of similarity, with three options; and cluster 
analysis, using weighted pair group method with average linkage. 
Written by W. W. Moss. [W. W. Moss, Division of Entomology, University 
of California, Berkeley, California--note change of address from that 
given in MUDPIE 2. Tape also available from Peters, Smithsonian 
Institution.] 
