GLISP: A High-Level Language for A.I. Programming

Gordon S. Novak Jr.

Department of Computer Sciences
University of Texas, Austin, TX 78712

This paper was written while the author was at the Heuristic Programming Project, Computer Science Department, Stanford University, Stanford, CA 94305.

Copyright © 1982 by AAAI.

This article appears in Proc. National Conference on Artificial Intelligence (AAAI-82), Pittsburgh, PA, Aug. 1982, pp. 238-241.

This research was supported by NSF Grant SED-7912803 in the Joint National Science Foundation - National Institute of Education Program of Research on Cognitive Processes and the Structure of Knowledge in Science and Mathematics.


GLISP is a high-level LISP-based language which is compiled into LISP using a knowledge base of object descriptions. Lisp objects and objects in A.I. representation languages are treated uniformly; this makes program code independent of the data representation used, and permits changes of representation without changing code. GLISP's object description language provides a powerful abstract datatype facility which allows the structures and properties of objects to be described. Reference to objects is permitted in an English-like syntax, including definite reference relative to the current context of the computation. Object-centered programming is supported. When interfaced to a hierarchical representation language, GLISP can perform inheritance at compile time, resulting in substantial performance improvements. In addition, a LISP structure can be specified as the way of implementing a class of objects in the representation language, making simple objects efficient in both time and storage.