GLIMPSE - A Statistical Adventure

This paper describes advances in the development of a know ledge-based front-end for the s ta t is t ics package GLIM 3.77, employing logic programming methods and tools, including sigmaPR0L0G and APES. The domain of application computer-based s ta t i s t i ca l analysis requires that a changing knowledge base be handled by the system, GLIMPSE*. The developing system is suitable for users with different levels of expertise, including those who wish to act independently of the system' s advice. Two features are incorporated into GLIMPSE that are suitable for other front-ends and expert systems a f ac i l i t y for suggesting answers to a user who requires guidance; an evaluator that suspends when information is missing and suggests an action to f ind the missing information. Previous and current methods of providing advice are discussed.