The Speci cation of \ Speci cation

The notion of \speciication" plays a key role in the developing science of computing. It is typically considered to be the keystone in the software development process. However, there is no single, generally agreed meaning of \speciication" that bears close scrutiny. Instead there is a variety of diierent, although partially interlocking and overlapping interpretations of the term. We catalogue this varietal profusion and attempt to lay bare both the sources and consequences of each major alternative. We attempt to present the full range of possibilities, and the biases inherent in each style of interpretation. By initiating debate on the meaning of \speciication" (and several other important terms) we would hope that a well-deened consensus will eventually emerge to replace the current very unsatisfactory state of aaairs in which a science and a technology are being built on shaky foundations. 1 The most deadly thing in software is the concept : : : that you are going to specify what you are going to do, and then do it. And that this is where most of our troubles come from. Ross, Garmisch. What happens is that speciications of software are regarded as functional speciications : : : It is my belief that anybody who is responsible for the implementation of a piece of software : : : must specify the design, the form. Sharp, Rome. No matter how precisely you try to specify a system, once you have built it you nd it isn't exactly what is wanted. Oestreicher, Rome.