Extending worksheet modelling software with expert system technology

Abstract

As a result of the rapid changes in the “information processing technology”, existing programming languages like APL are facing a number of challenges to maintain or increase “market share”. Such technologies as networks, worksheets, expert or knowledge based systems, host / workstation shared processing, software quality, and end user computing are of concern to the consumer of programming languages. It is the contention of these authors and others that APL already contains many features to meet new needs (for example, tracing the execution of a function or networking users on a single mainframe), or can easily and elegantly be extended to meet new needs (for example, invoking assembler from APL or a host - workstation link). This paper describes a tool called Automated Ledger Book (ALB). ALB uses expert system technology to extend worksheet modeling software function and therefore enhance end user computing. The purpose of this description is to help demonstrate the applicability of APL to such new areas as expert systems, and help the APL community better understand the links between expert systems and decision support systems. ALB is a worksheet model generator with some enhancement based on an equation manipulating expert system and a natural language interface built by the authors in APL/APL2. ALB enables the user to build and run worksheet type models in high school algebraic notation and English without regard to equation order or specification of calculated and input variables. Solving sets of algebraic and conditional equations, ALB adjusts itself to input, responding to unlimited 'what ifs'. When challenged, ALB provides the rationale for its conclusions.