CIS Colloquium, Nov 14, 2007, 03:30PM - 04:30PM, TECH Center 111
Come Out, Come Out, Wherever You Are (Predicting Where Faults are Hiding)
Dr. Elaine Weyuker, AT&T Research
Paper Title: “Some Progressive Ideas on Regression Testing”
Authors: Andy Podgurski and Elaine J. Weyuker
All software systems evolve, and there are many causes for this evolution. Some changes are attributable to fault identification and location, some are due to specification or requirement changes, some are based on a changing customer base or usage patterns. Regardless of the source of the changes, and regardless of how stable the software has been, it is always necessary to do some amount of additional testing as a result of any type of system modification.
Historically, the term regression testing was introduced to indicate that retesting of software was being performed in order to verify that the software had not ``regressed'' in the sense that changes in the software or specification had caused portions of the software that functioned properly at one time on a given set of test cases, to function improperly. In this paper, we broaden the use of the term ``regression testing'' to include any retesting of software during maintenance whether or not the software has been modified, and regardless of why it has changed. As long as some facet of the observable behavior of a software system has changed, retesting must be done, and this will be refered to in this paper as regression testing.
In this paper we concentrate on a foundational investigation of regression testing, and discuss various ways of categorizing regression testing criteria.
Dr. Elaine Weyuker is an internationally-renowned expert in software engineering and testing. Dr. Weyuker has helped to develop the formal foundations of testing, and has been a leading proponent in the effort to make testing a recognized, professional specialty. In particular, she has provided the theoretical foundation and several practical techniques for a shift from "static" measures of testing (such as the number of source code lines covered) to "dynamic" measures (such as coverage of user profiles and impact of potential features). Her work has helped make it possible to scale testing up to source code that is several orders of magnitude larger than what was previously possible to test. She has numerous awards and distinctions at the highest professional levels, including being a Fellow of IEEE, a Fellow of ACM, a member of the National Academy of Engineering, and an AT&T Fellow. For the past decade, she has been a senior researcher in the Information and Software Systems Research Lab at AT&T, and she is also on the board of directors at the Computing Research Association.