Accting., Mgmf. & Info. Tech., Vol. 2, No. 2, pp. 57-76, 1992 Printed in the USA. All rights reserved. 0959~8022/92 $5.00 + .Oil Copyright D 1992 Pergamon Press Ltd. A FORMAL ALGORITHMIC MODEL COMPATIBLE WITH THE CONCEPTUAL MODELING OF ACCOUNTING INFORMATION SYSTEMS Andrew D. Bailey, Jr.* Kyeong S. Han? Ryan D. Stansifert Andrew B. Whinstons *University of Arizona t Yeungnam University $University of North Texas 0 University of Texas Abstract-Accounting and computer science have been uneasy partners in the development of account- ing information systems. The history of their association has generally led end users to do the best they could in matching the needs of accounting systems to the demands of current hardware and software developments. While the results have been impressive, recent developments in computer science make it possible to approach the modeling of accounting information systems in a manner more consistent with the accounting model than previously possible. This paper, motivated by program verification is- sues from an auditor’s viewpoint, introduces a functional approach to accounting systems development. The result is a formal algorithmic model of accounting that is consistent with current developments in computer science. A number of benefits are derived from the approach, including: (1) natural repre- sentation; (2) ease of coding, design and modification; and (3) verification capabilities. Keywords: Accounting systems, functional programming, program verification. INTRODUCTION Accounting and computer science have been uneasy partners in the development of ac- counting information systems. The history of their association has generally led end users to do the best they could in matching the needs of accounting systems to the demands of current hardware and software developments. While the resulting use of computers in sup- port of accounting needs has been impressive, from the user’s perspective it has remained a somewhat forced relationship. Recent developments in computer science allow us to cre- ate a model in which the computer and accounting concepts are naturally compatible. This may be best understood by comparing the complex process of converting an accounting sys- tem description to a COBOL program. This translation is, at best, awkward due to the lim- itations of extant software/hardware (Weber, 1982) in terms of both developing the system and understanding it after it has been developed. As a result, end users, such as auditors, devise various heuristic means to establish that what has been programmed does in fact meet the requirements of accounting theories and practices. Our research endeavors to fit 57