Using OPEN's Deontic Matrices for E-Business

1 INTRODUCTION Method engineering (e.g. Brinkkemper et al., 1998; Ralyté and Rolland, 2001) focuses on how to create a method (here a software development method) that is highly suitable and " personalized " for a particular project or organization. Successful method engineering needs, as input, a repository of process elements (Chroust, 2000) that describes individual elements of the methodological approach; these include descriptions of activities, phases, tasks, techniques, roles, deliverables and so on. Given such a repository, sometimes called a process reference model (Henderson-Sellers et al., 2002), a number of various kinds of method or methodology can be constructed (Rupprecht et al., 2000; Henderson-Sellers, 2002a,b). In this paper, we examine how the use of a deontic 1 matrix, originally introduced by Henderson-Sellers and Edwards (1994) and Graham (1995a) provides possibility values for identifying useful process components and how they should be interconnected. We use industry data for one particular business domain, that of e-commerce, to realize the deontic matrices as used in the OPEN Process Framework (Firesmith and Henderson-Sellers, 2002). The software development paradigm used will be that of OO (object-oriented) and CBD (component-based development). Values are given for both large and small B2C examples. Before progressing further, two points are important to note. The first is that the nature of Web development processes is currently rather poorly addressed within the research literature (see, for example, Lowe and Henderson-Sellers (2001a; 2001b). The purpose of the research described in this paper is to begin to understand and structure our understanding of how to effectively construct Web development processes. In this context, we have adopted a grounded theory research approach –progressively collecting and analysing data on specific 'experimental situations' (i.e. actual commercial projects) within the context of an emerging theoretical framework. Secondly, although the discussion in this paper focusses on applying concepts within the framework provided by OPEN, the general lessons with regard to selection of process components is much more broadly applicable.