Phyllis A. Richmond

Classification from PRECIS: Some possibilities†

The PREserved Context Index System (PRECIS) developed for subject indexing for the British National Bibliography is discussed as a basis for various studies relating to classification which could be made from its initial phrases, strings, entries and back-up structure. Specific possibilities include use of methodologies such as Goffmans “indirect method,” Foxs “characterization” of rules, derivation of empirical hyperbolic distributions, and lognormal distributions. The relationship of existing methods of subject analysis to each other through PRECIS is explored. Brief samples are included.

An extended character set for humanities computer printout

With the new version of the United States of America Standard for Communication Information Interchange (USAScIl x.3.4-1967),1 it is possible to have extended character sets, triggered by escape codes in the main set of characters. One great advantage of having such extended sets is that, so long as there are sufficient escape codes, one can go on with many different groups of characters. For example, it should be possible to develop a set for music, reachable by an escape code from the standard USASCII. For such purposes, in the unlikely situation that music symbols were never to be used by the Library of Congress in catalog cards, it might be possible to instruct the computer to accept one of the USASCII standard graphics as an escape code, even though the Library of Congress was using the same graphic to escape, say, into Greek. If, however, one wanted to add an incipit, in musical notation, to the standard cataloging of musical scores now performed by the Library of Congress, it would be better to reserve a unique symbol as an escape code for musical notation. Lest one despair of limitation in the number of possible escape codes, it should be pointed out that the standard IBM print train for the 360 line of computers permits 240 characters and to date (August 1969) only 174 are used on the extended set now planned by the Library of Congress for catalog card printing in all Roman alphabets. There is no particular reason why this special print train should be the only one. Presumably there will be other extended trains for other alphabets and further graphics. In fact, there is no reason (other than cost) why any organization requiring special graphics cannot have its own extended print train to suit its own purposes. Undoubtedly the scientific community will do just this.