Document Type

Conference Proceeding

Faculty

Faculty of Computing, Health and Science

School

School of Computer and Information Science

RAS ID

2643

Comments

Johnstone, M. (2005). Using process modelling to capture requirements for a digital library. In Information, libraries and elearning : Proceedings of the Inaugural Transforming Information & Learning Conference.

Abstract

This paper describes the results of an action research study which explored how IT professionals used process modelling to capture requirements and design artefacts concerning a digital library system. The intention was to produce a web-based, non-relational (XML) database system to store publications of various types and to allow public domain access to the system. Traditionally, such systems are constructed with a data-oriented design, with little attention being paid to process concerns. In this study, process, described by business rules, was defined first which provides advantages as many business systems (including library systems) are often process-oriented. The team involved in this study modelled a reasonably complex system, articulating 168 business rules across 18 major functions covering many aspects of the operation of a digital library system. The team was able to produce a usable requirements specification that was used as the major input into the design phase of the system being developed. The design phase considered both process and network perspectives explicitly before modelling a data-oriented view of the proposed system. It is argued that this multi-perspective view led to a better-designed system than would have been obtained with a single perspective method. An issue for this study was that third year software engineering students were used in place of professional software developers. Using students instead of practitioners raises an interesting question concerning the authenticity of this study vis-a-vis action research. The fundamental question being tested in study was to ascertain whether a process modelling method could be used to model the domain of digital libraries i.e. does it contain a necessary set of constructs to satisfactorily model the problem domain? It was considered that the environment of student projects in a university with regular supervisory oversight offered advantages for the monitoring of this question more so than the alternative. As regards the authenticity, this study was a real project for a real client.

 
COinS