Arguments are presented in support of the evolutionary development of information systems. The traditional analysis and design tools are shown to be inappropriate for the evolutionary development, and the requirements for such a development environment are determined. Three major tools are borrowed from the fields of artificial intelligence, programming languages, and database management, and combined to create an environment that meets these requirements. The resulting novel system architecture is described by using a university registration system.
The interaction of formal information systems and organizational decision models will be examined in this article. It is conjectured that decision models and processes not only determine the information requirements, but that they are influenced by the organizational information systems that are designed to support them. This type of circular relationship between the decision models and the information systems undermines the success of requirements analysis which traditionally views information systems as supporting structures for the decision models, and ignores their counter effect on the decision models. Four different examples are presented to demonstrate the effect of information systems on organizational decision models. The examples range from algorithmic to highly unstructured and speculative, but they all suggest that information-intensive models are qualitatively different from their information-poor counterparts.