Programming Knowledge: On Indexing Software for Reuse and not Indexing Documentation at All

Abstract

This paper notes something of a paradox: Computer software reuse systems both intellectual and automated are making use of humansand documentation at a time when the role of both humans and documentation is being called into question by the computer industry itself. The challenge facing classification researchers and practitioners is to show how reuse systems ofall types depend on natural-language representations of the texts they contain, and to share the methods that classification research and practice have developed to create those representations.