A Critical Perspective on KRL

Bobrow and Winograd have presented to the A1 community two descriptions of KRL (Bobrow & Winograd, 1977, Bobrow, Winograd et al., 1977) which explicate both a high level A1 programming language and a theory of knowledge representation. In actual practice, the line between these roles is necessarily vague. As is the case with all programming languages, commitments made to specific data formats or control structures profoundly affect design decisions made by the user. In KRL, there are additional commitments to knowledge representation in the programming language as well. While these cornrnitments are neutrally presented as convenient features of a high level language, their impact on the user would be far less neutral. To a user who has not previously investigated problems of knowledge representation fust-hand, KRL either suggests a particular approach or imposes that same approach. In either case, the user is liable to be unconscious of the continual trade-off between low level design options and high level programming convenience. KRL is almost certainly meant by its authors to replace LISP as the basic A1 format of description, in much the way that FORTRAN once replaced machine language. A successful unification of the field at that level would be enormously positive; and would eliminate many of the purely notational disputes with which we are all familiar. Moreover, they have introduced as essential features of the language the notions of conceptual entities, inexact matching, and redundant representations, to name a few. Each of these has long been argued for (at least in the field of natural language processing), and their espousal represents a courageous break by the two authors from their past opinions. Indeed, they have gone further and, in a highly unusual move, published their own criticisms (Bobrow, Winograd et al., 1977) of the first paper.