Chapter 4. Computer-based Accounting Systems: Practice
The previous two chapters attempted to identify the characteristics of "good" accounting systems by reviewing the normative and empirical literature. This chapter begins the right-hand branch of the four-chapter Requirements Determination process illustrated in the lower half of Figure 1.1: it surveys the present state of computer-based accounting practice; Chapter 5 reviews computer-based accounting systems theory. Opportunities and problems identified in this chapter should indicate areas for improvements in future computer-based accounting systems.
Despite the widespread use of computer-based accounting systems (see survey evidence in Section 4.1 below) there has been very little research into the features that make one system or suite of systems better than another. It was therefore necessary to conduct some first-hand empirical studies specially for this thesis. It is unusual to report original empirical research in the "literature survey" stage of a thesis, but in this case there was no choice. Most of Section 4.1 reports on results of two mail surveys of general ledger system users. Mail surveys can provide useful, unambiguous answers to simple questions, like "Do you have a computer-based general ledger?", but to gain an understanding of how today's systems work, and what they are intended to do, it was necessary to examine some specific systems in depth. Sources of information were first-hand experience with PC software, visits to user sites, and vendor's documentation. It was mainly through first-hand experience and reading documentation that some useful insights into the nature and purposes of computer-based accounting software began to emerge. These insights are discussed in Section 4.2.
4.1An Overview of Computer-based Accounting Systems in Practice
Following Anthony  and Keen and Scott-Morton , computer-based accounting systems are usually classified as either transaction processing systems or decision support systems. Transaction processing accounting systems (TPAS) are used for capturing, processing, storing, and reporting, the sometimes millions of day-to-day transactions that occur in a firm each month. Decision support systems (DSS) use spreadsheets and financial modelling packages for budgeting, forecasting, and analysis purposes. DSS packages are not normally used for capturing and processing the large volumes of day-to-day transactions required for accountability reporting. Some firms still develop their own TPAS software but, as documented below, there is an increasing trend towards use of packages for both TPAS and DSS.
Sales revenues of software suppliers provide an indication of the economic importance of accounting TPAS and DSS software. Industry sources estimate combined 1988 sales of the mainframe TPAS suppliers MSA, McCormack and Dodge, and Software International were around $US600 million. PC-based TPAS systems and in-house developed systems add to this figure. In the area of DSS software, Lotus Development Corporation reported 1988 sales of $US469 million. It would therefore seem reasonable to suggest that 1988 world revenue for packaged accounting software was around $US 2,000 million. This is only about 1% of world-wide hardware sales (in 1988 IBM and Digital's sales revenues were $US60,000 million and $US11,500 million respectively [Fortune, April 24, 1989, p.168]). Nonetheless, at $US2,000 million p.a., design, development, and marketing of accounting software is a significant economic activity.
In Chapter 2 (see Figure 2.3, p.2-9) it was decided that this thesis would concentrate only on routine accounting information systems, i.e., only on TPAS. The annual $US600 million market for mainframe TPAS is offered as evidence that even this more narrow focus is important enough to warrant academic study.
4.1.1 A description of TPAS software
This section provides a...