CRANV1P1 ASLIB Cranfield Research Project: Factors Determining the Performance of Indexing Systems: VOLUME 1. Design, Part 1. Text Formation of Index Languages chapter Cyril Cleverdon Jack Mills Michael Keen Cranfield An investigation supported by a grant to Aslib by the National Science Foundation. Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. -75 - but it was subordinated to Ablation cooling since its functions in the literature indexed was that of an agent of the cooling process. The fact that Heatshield was subordinated to Ablation cooling devices did not mean, of course, that it was unavailable as a member of the class Structure if this latter subordination had also been required. It was placed under Ablation solely on the score that in the test collection, or the subset, this was its most probably use- ful hierarchy. The concept schedules were essentially 'one-place' schedules in linear sequence, in the sense that no attempt was made to repeat one concept in several different positions should it happen to belong usefully to several different hierarchies. This last event was provided for by the rotated A/Z index, and by references within the schedules described later. It must be emphasised that the function of these sched- ules was simply to show as clearly and as comprehensively as possible the hierarchical relations (generic and non-generic) between the terms (concepts) so that searches could be programmed from them. The major relations were most economically displayed by physical juxtaposition. Other hierarchical relations were established via the A]Z index and by internal references (linking, for example, Heat transfer, subordinated to Thermodynamic processes, to Transport properties in general). The index in the physical sense (the matrix of index descriptions and document numbers) consisted of the separately entered concepts designed to be searched post- coordinately (the strip, or the scan-column, method used is described in Chapter 6 ). So a compound like Fully developed laminar channel flow could be sought equally in any of the various arrays concerned, or combinations from them - Fully developed flow, or Laminar flow, or Laminar channel flow, etc. It follows from the above that the problem of citation order was very much reduced, compared with a real life schedule, since it was confined entirely to the choice between two (and sometimes, but rarely, three) elements; e.g. Jet noise, Interference rocket, Laminar boundary layer heating, Surface stress, Slot blowing. But whilst these particular examples offered a serious choice between two or three equally important elements, the great majority did not even demand this; they con- sisted of combinations such as High pressure ratio compressor, or Hinged flap, where the major element was obvious and the other elements trivial; no hierarchy of Ratios, or things High, or things Hinged was necessary. Since the concepts repre- sented the limits of precoordination, the problem of providing for the accurate pre- diction of the exact location of all potential synthesized combinations (a major function of citation order) did not arise. The problem of 'distributed relatives' was solved by postcoordination of the concepts. It also follows from the above that problems of notation were virtually non- existent. A purely ordinal notation to identify quickly the 1oc[OCRerr]tion of particular simple concepts was the only requirement. Problems of hospitality and expressive- ness did not arise; no additions to the schedules were contemplated and no aids to display were necessary in schedules which were relatively homogeneous and designed for internal test-programming entirely. Within each major area the various facets and arrays (subfacets) were now distinguished. At this point the problem of displaying generic and non-generic hier- archies arose; it was met as it usually is in conventional library classification, by subordinating to a thing its various categories - its kinds, parts, properties, pro- cesses, etc. Below is a brief extract from the schedules which are given in full in Appendix 5.4, followed by an explanation of some of its features: