Automating testing by reverse engineering of software documentation

Reverse engineering of software documentation can play an important role in automating software testing. Since testers often compare the operation of a software system to the expected behavior as described in a document, information extraction from texts can provide valuable input to automated test systems. This paper describes the SIFT (Specification Information From Text) document parser and its use in generating tests for two different software systems: the OpenVMS operating system and the XCON expert configuration system.