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. - 89 - But this is a permanent feature of Flutter in the collection context and does not alter from one document description to another. But roles are expressly designed to clarify the local and varying relations between the terms of a particular description and perhaps this rules out role 5 from consideration. Flutter could also be regarded as the primary topic (role 8). However, firstly, role 8 is not strictly speaking a role at all; it does not show semantic relations between the different terms of an index description, only its subjective state as to a hypothetical reader; it is, in fact, a weighting device. Secondly, it could be argued that piston theory is a primary topic (role 8). Perhaps both terms could be labelled 8, but piston theory could also be regarded as an agent (role 10), although this would overlook the fact that it is not an a[OCRerr]_[OCRerr]p_t of Flutter but of its analysis. Supersonic flow could be considered as an environ- ment (role 5) or as a cause or influencing factor (role 6). If it is treated as (6) how- ever, Flutter, which it affects, ¢[OCRerr]ould be the factor influenced (role 7). When the questions to which document 1014 was one of the relevant documents are considered, the further difficulties in ensuring a match become apparent. Question 97 refers to the Prediction of flutter on lifting surfaces. Flutter could conceivably be given role 3 (undesirable component), or 8 (primary topic). It could also be considered as rote 9 in the sense that it is the object of analysis or predictive operations. Two other questions, numbers 98 and 276, to which document 1014 is also relevant, are on how flow characteristics or leading edge bluntness affect Flutter. So now Flutter is an influenced factor (role 7). Yet another question, number 3, is about aeroforces acting on high speed aircraft, and if Flutter is regarded as a kind of aeroforce, it may also be given role 6 (cause or influencing factor). The general import of these considerations seems to be that the term can be forced to play simultaneously a number of different roles in the same document according to what the particular user is seeking, and that attempting to label them too precise- ly is liable to result sometimes in unjustifiable rejection of indexing descriptions because they do not match exactly in the roles assigned them.