Can the case for CASE technology be advanced by Process Improvement?

Abstract

Findings of a case study that focused on understanding how software development proceeded in a small Information Systems Department (ISD) located within a major public sector service are presented. The observations collected relate to a period prior to, during and after the introduction of a CASE (Computer Aided Software Engineering) tool. They are based on a combination of minutes from quality circle meetings, interviews and regular on-site observations by the first author. The study used a framework loosely based on SEI's CMM model to characterize the state of practice before and after the introduction of the tool and to assess process improvement. Prior to the introduction of the first CASE tool a typical ISD software development task was a standalone single reporting application. The software development process model consisted simply of requirements acquisition and program development. Both these activities for a particular application were carried out by an individual developer within a few months. The former activity involved the client/user, with the level of participation dependent on the developer. Similarly, program development varied from one developer to another, with most effort being expended on coding and testing. Not surprisingly, maintenance was locked to the original developer. A decision that ISD should develop larger and more complex applications triggered the purchase of a CASE tool. Typically, the larger applications would be developed by a team of about five people over a period of twelve or more months. The introduction of CASE tools, first Case Designer 5.1 then Designer 2000, had a marked effect on the working practices of ISD. Specifically two more stages were introduced into the development activity: designing via an entity-attribute relationship model and validation via rapid prototyping; and greater attention was paid to the management of testing and fault reporting. The paper explores whether the benefits accrued by ISD can be attributed to the CASE tool, the changes in work practices, or a combination of both.