M. Odeh, I. Beeson, S. Green and J. Sa Systems Modelling Research Group Faculty of Computing, Engineering & Mathematical Sciences University of the West of England Coldharbour Lane BRISTOL BS16 1QY tel: +44 (0)117 965 6261 fax: +44 (0)117 344 3155 email: firstname.lastname@example.org (or ian.beeson or stewart.green or email@example.com) Abstract In Software Engineering and Information Systems, increasing attention has been focused lately upon modelling organizational processes - as a starting point for developing computer-based systems to support (or control) such processes. A number of process modelling methods are available, but it is not yet clear what the relative merits of these are, nor whether they might bemore or less useful in particular contexts. We have applied two well-known process modelling techniques, Role Activity Diagramming and UML Activity Diagramming, to a particular process in our own organization, that of managing the registration of research students. We developed an RAD first and then translated it into a UML AD, to compare the two techniques and check the feasibility of suchtranslation. We conclude that translation from RAD to UML AD is likely to be feasible in particular cases, but will rely on the ability of the translators to establish and maintain the equivalence between the two (i.e. the equivalence will be partly a matter of local interpretation).
As IT systems have penetrated further and further into everyday organizational processes, interest hasbeen growing in how we might understand and represent those processes in a sufficiently comprehensive but also lucid manner, to serve as a basis for introducing new technological arrangements which will effectively support, and perhaps in some cases direct, the organizational processes. From a software engineering perspective, the task here is not so much to capture a process in order to automateit, as to comprehend a process in detail in order that the human activities and interactions which constitute the process can be given appropriate technological support or can be in some ways improved or better integrated through the inclusion of computer-based components. Ould has observed that this has involved a switch of emphasis away from information structures and flows as such towards theactivities that individuals and groups carry out in their work in organizations: "...in our approach to process modelling we concentrate unashamedly on what people do, rather than on what people do it with. Once we have chosen an organizational structure and the processes it will operate, then we can decide on the information needed by individuals and groups to perform those processes in thatorganizational structure." (Ould, 1995, p. 16) In an earlier study (Beeson et al, 2002), we used Role Activity Diagramming to investigate the processes linking strategic decision making with information systems provision in an insurance company. We argued in that paper that the interconnection between the general business and IT/IS units in the company, and between strategic and operational levels, wasbeing achieved more by dynamic and continuous processes of negotiation and communication than by the implementation of a
fixed strategy or architecture. We found that RADs were an effective tool for charting these processes and had some evidence that they also helped organizational members grasp and verify the processes. The present paper emerged out of our interest in the similarities ordifferences between role activity diagramming as we had used it in the previous study and activity diagramming in UML. We were aware that the Unified Modelling Language (UML) was becoming increasingly important in software engineering, and that it included Activity Diagrams as one of its techniques. We understood that UML aimed to provide system developers with a coherent set of notations which...