Integrating Code Knowledge with a Software Information System

Software Information Systems represent and provide access to knowledge about large software systems to support discovery, reuse, and maintenance. Such a system should represent various information about the actual code, including files, code objects like functions and global variables, and process structure and message traffic. This paper describes the CODE-BASE system, including the kinds of code knowledge it represents, how the knowledge is acquired and represented, and a scheme of `demand-loading` between the in-core system and a code database.