International Workshop on Software Reuse and Safety

One of the problems of reuse and safety is that there is no consensus on what constitutes "safe" software, much less a "safe" reusable component.

One approach toward evaluating safety involves a technique known as safety integrity levels. This is used in a number of standards, and is often misunderstood. Its centrality in some important standards makes it an important concept to understand, however, in the context of software safety.

Another approach toward evaluating software safety involves the idea of equating safety with reliability. "A reliable component is  a safe component." This has the advantage that reliability can be applied at the component level. But it leads to problems such as finding a reasonable interpretation of the idea of Mean Time to Failure of a software component.

Yet another approach to software safety involves treating safety as a characteristic of dependability. This leads to difficulties in the interpretation of what it means for a software component to be dependable (as opposed to an entire system).

Another approach that has been the subject of much interest in recent years - an approach advocated by some noted software engineers such as Bertrand Meyer - is design by contract.

The concept of proven in use is being explored by large organizations in particular, who want to find a way to reuse substantial software components in new systems.

Each of these approaches, together with associated resources, is treated in separate subsections.