Standardizing at the leading edge

Ⅵ International standardization activities are often thought to be undertaken only when the subject is mature. It is further presumed that as concensus takes so long to achieve, the technology concerned will often have moved far ahead by the time the standard is published. In one area of information systems activities, however, the reverse is the case: the standards-makers are working at the leading edge of the subject. The work in question is that of ISO/IEC JTC1 Subcommittee 7, Working Group 12, which is concerned with standardizing methods for sizing software. This methodology is important in software development for a range of practical purposes such as productivity measurement, effort estimation, controlling the scope of a project as it progresses, etc. Various sizing methods exist, and it is clearly desirable to achieve some common international agreement on the best approach. The first step taken towards standardizing sizing methods has been to seek the underlying principles of software sizing, and to define these in a 'meta' standard, rather than to attempt to define a single international software sizing standard. This approach has brought new challenges, not the least of which has been agreeing on abstract concepts across diverse languages and cultures. But the result has been new insights into how to size software , which it is difficult to believe could have been achieved other than through such an international forum. There are clear lessons to be learned about the value of this way of working; it is equally clear that attention must be paid to the dissemination and promotion of new ideas so that they can be quickly taken up in the market place, and their benefits made available to the user community. hould the development of international standards in a fast-moving area such as information technology be pursued at the leading edge of its expanding technology and ideas, or should standards be developed only when the subject is well understood, stabilizing and matur-ing? And does it matter in practice which path is pursued? The common perception is that most standardization takes place well back from the leading edge. The development of standards for technologies as diverse as programming languages such as COBOL, operating systems such as Unix, and hardware and telecommunications generally, has proceeded at a pace which left those standards, by the time they were agreed on and published well out of date. To those in industry eagerly following …