Using a document parser to automate software testing
Patricia Lutsky · 1994
The SIFT document parser is a software engineering tool that extracts Specification Information From Text.It can be used for reverse engineering, since it can recreate a specification-level description of a software system from documents such as computer manuals.The information extracted from these texts can then be used to test that the system behavior matches the documentation.This paper describes the SIFT architecture and reports on experiences in using it to generate indices into a database of reusable software testcases and to augment expert system database tests.SEFT's design separates domainindependent and domain-dependent sections to make it easy to apply to new document types and new tasks.