BY BRUCE G. BUCHANAN
Department of Computer Science
University of Pittsburgh
" Expert systems require substantial knowledge bases, yet the process of acquiring the requisite knowledge can be cumbersome. Automating the learning of new knowledge has been the topic of considerable research. A new study brings together in one framework the various approaches to knowledge acquisition and uses it to evaluate their applicability to building expert systems.
V.L. Sham, E.J. Wisniewski, K.R. Levi, R.D. Scott, “A formal analysis of machine learning systems for knowledge acquisition,” International Journal on Man-Machine Studies, 29 (4), 429-48, October 1988.
" The traditional model of the stages of software production covers the “life cycle” of software from initial specification through coding and other activities to operation. Data from 13 large software projects (tens of thousands of lines of code) show that the traditional model does not reflect actual practice.