TMorgan_posn

Producing business solutions:

Making better use of our intellectual capital

Tony Morgan, Unisys Limited

Note: The views expressed in this presentation are solely those of the author and should not be taken to represent the views of any other person, group or organization.

Although a great deal of knowledge and experience exists relating to the development of complex software systems, it's clear that we don't use this in the most effective way. The two main sources of know-how - industry and academia - usually have quite different perspectives. In general, academia tends to focus mainly on issues relating to software engineering, as opposed to industry's 'solutions' (of which software is a part). Well-respected text-books frequently show only token coverage of key industry concerns. For example, although testing can typically consume 30% of project budget, it receives scant treatment in many software texts.

Industrial software production has more in common with Flexible Manufacturing Systems than with the traditional view of Software Engineering, in which coding takes a dominant role. Coding has traditionally been treated as a craft skill, carried out by ‘clever’ individuals with specialist training. In reality, coding in an industrial setting is a low-level task that can be easily contracted out to a low cost area.

Much more important (and valuable) is the knowledge used to define and specify the code. Unfortunately this is rarely captured in any systematic way. This makes software expensive, and leads to calls for more software re-use to reduce per-unit costs. An alternative (and perhaps better) approach would be to incorporate more structured knowledge management into the production process.

To some extent, industry is moving in this direction, with more emphasis on software components, design repositories, etc. However, much of this effort is based on ad-hoc approaches and driven by short-term objectives. Industry would have much to gain from collaboration with academia (and vice-versa), if only this can be organized in the right way. This should not be misunderstood to be a call for academia to carry out development work for industry. Rather it is an expression of the need for a sound framework within which industry and academia can better understand each other's objectives, and focus on the areas giving the best joint pay-off.