Abstract :
A computer-based economy implies a computer-based education. Learning and teaching are optimized by the student, as well as for the professor, if instructional redundancy is captured for reuse. One of the fundamental problems, to be confronted in the construction of complex learning programs, is that of providing ever higher-level authoring languages. This paper shows how instructional reuse, as mediated by a random seeded crystal learning algorithm, can facilitate the evolution of complex learning frames. A natural language interface enables the sharing of large software projects across a team. Here, a knoqledge-based system acquires a capability to map natural language phrases onto object invocations. The natural language phrases are iteratively normalized through a transformative process, which utilizes band theory. A human-machine system is described, which entails the construction of a self-referential program editor. The presented concepts and arguments are codified with examples taken from information systems and flexible manufacturing. Learning is shown to be a consequence of randomization and reuse.
