The authors attempt to formalize some properties which any reasonable control-flow complexity measure must satisfy. Since large programs are often built by sequencing and nesting of simpler constructs, the authors explore how control-flow complexity measures behave under such compositions. They analyze five existing control flow complexity measures-cyclomatic number, total adjusted complexity, scope ratio, MEBOW, and NPATH. The analysis reveals the strengths and weaknesses of these control flow complexity measures.>

Abstract In 1982 it was decided to computerize the indexing activities of SMIC—the Sorghum and Millets Information Center—located in the Library of ICRISAT (International Crops Research Institute for the Semi‐Arid Tropics). The previously used manual indexing procedure was found to be unsuitable for computer manipulation, and it was decided to replace it by some other system which would be computer‐manipulative and could represent subject contents of the documents accurately and precisely. A survey of the existing systems was made and none of them was found entirely satisfactory for SMIC. A simple indexing system has been developed which is free from any classification system and which is not limited by cumbersome postulates. In this system, keywords chosen by the indexer (who is expected to have some subject knowledge) are arranged in a meaningful sequence (a logical string). The keywords are connected by punctuation marks depicting various types of associations. The keywords are rotated to provide access through each significant keyword. The system, which is computer‐manipulative, can also be used manually.