Colloquium Computer Science - Prof. dr. E.O. de Brock RUG/FEB

Title: Developing Information Systems revisited

Abstract:

Although Developing Information Systems is an ‘old’ research topic, there are still several problems (both old and new ones). Some ‘old’ but lasting problems in developing information systems are that user wishes are often unclear (at least initially) and that the language/thinking of users is (completely) different from the language/thinking of developers. Later problems arose because we went from automating known processes to enabling entirely new business models (nowadays even before the business will start). Newer problems, which are becoming stronger and stronger, are a.o. that circumstances are changing (very) quickly and continuously - and hence the requirements as well - and that ‘times to market’ should be shorter and shorter.

Consequently, IS development projects are still often failing on the basic project requirements: They are not within budget, are not within time, and/or have inadequate functionality (e.g., see NVWA for a recent dramatic example).

Many problems are requirements-related and/or related to the use of natural language. The problems also ask for increased traceability, transparency, and development speed during the (continuous) development and evolution of information systems.

We will concentrate on development paths for functional requirements. We introduce a development path that enables ‘stepwise clarification’ and ‘stepwise specification’. It gradually goes from the informal natural language and way of thinking of users to the more formal language and way of thinking of developers (schema’s, models, inputs, outputs, parameters, procedures, etc.). Some general structures and forms for such development paths are presented, w ith an emphasis on ‘patterns’. Moreover, we present an evolutionary development path for the system-level . We are also incorporating this in our CS-course Problem Analysis and Software Design.