Does Reuse Require Special Approaches to TT?

Is there any reason to think that the TT process applied to other technologies may not apply to Reuse? Participant opinions were divided.

Those who argued for the need of a special TT approach pointed out that reuse is not a ``technology''. It involves the application of well-established engineering principles--the application of best principles and practice, case-based reasoning--to software engineering activities. In the established engineering communities, reusability is part of the culture, like the principles of physics or mathematical notations. That culture is propagated through educational programs, training, engineering standards, codes of practice, and so on. Following this argument, rather than investing on technology transfer, we should invest in education and on consolidating good ``software engineering theories'' and ``domain-specific theories'' that could serve as a reliable foundation for the profession (by analogy with physics, materials science, structural analysis, and so on). Having redefined the TT problem as one of (i) building reusable theories and (ii) teaching people good engineering, the group could not agree on the question of how to reliable build reusable theories (i.e., What is the scientific method?). Software engineering research and the practice of domain analysis will continue to produce many of these. Practical validation and market needs will probably drive a Darwinian process of selection/evolution as it has happened to computer architectures, programming languages, design notations, and so on. An interesting question not addressed by the group was: How could we speed up the selection/evolution process?

Those who argued against the need of special TT approaches justified their answer on the observation that ``good'' TT approaches used by other disciplines do not seem to depend on what is being transferred but on:

• the structure of the organizations participating the transfer,
• funding schemes,
• the products or processes to which the technology applies,
• the economics of the markets of those products or processes, and
• the level of sophistication of the user community.

The discussion touched briefly on ``What is transferred in Reuse TT?'' There was agreement that the two major components of the transfer are: (i) systematic processes for building on what exists, and (ii) methods and tools for building reusable software and for assisting in reusing it.