Workshop on Software Reliability held in Paris, France

Pierre Dersin, Marsha Abramo

On September 5th the French Chapter of the IEEE-Reliability Society organized an international workshop on the current status of Software Reliability Engineering theory and practice. The one day workshop was sponsored by the IEEE RS. Organizers were Dr. Pierre Dersin, AdCom member, and Dr. Frédérique Vallée, French Chapter Chair.

The Speakers included Dr. Samuel Keene, IEEE, Prof. M. Vouk, South Carolina State University, Prof. T. Dohi, Hiroshima University (Japan), Dr. Vallée (All4 Com,France), Dr. M Antoni (French Railways SNCF), and Dr. K Kanoun (LAAS, French Research Institute). Ms. Marsha Abramo, IEEE-RS VP Membership, concluded the day with a presentation on the IEEE-RS, its organization and benefits.

By all accounts it was a great success, as an active and interested audience of some forty gathered in the ALSTOM Headquarters at Levallois, a suburb of Paris. Pierre Dersin an employee of Alstom secured this venue for the event. ALSTOM, is a worldwide infrastucture specialist company involved in electrical power generation, transmission and distribution, as well as rail transport including high-speed trains and driverless subway systems. We are very grateful for their hospitality and flawless logistics.

The topic of software reliability was selected because it is a subject that continues to be controversial. A lively debate during the roundtable discussion followed the technical presentations. Even though software reliability assessment and prediction techniques have been around for more than 40 years and have given rise to numerous academic publications, and are even described in standards such as IEEE 1633, their penetration in and acceptance by industry is still not uniform and limited.

The reasons that were identified for this state of affairs are essentially cultural. Software engineers are often not familiar with the language and concepts of probability and statistics and may find it difficult to grasp software reliability engineering techniques. Also managers may be unfamiliar with software reliability which is not part of the curriculum in business schools.

Yet, it was illustrated by examples in some of the presentations that overcoming those barriers pays off. The companies which have invested in those concepts and techniques have reaped the benefits.

The main advantages are: a decision support tool to decide when to stop testing and when to introduce a new software on the market; fault density reduction; rework cost decrease and last but not least, significant improvement in quality and customer trust.

A caveat was made by some speakers: the scope of application of software reliability engineering techniques must be carefully delineated. They primarily address non-safety-related faults. Safety-impacting faults must be exhaustively eliminated. In Dr. Antoni's words, ‘the black swan must be eradicated’.