Abstract/Summary

We start our research by introducing the subject of Enterprise Architecture (EA), its content and purpose, as well as discussing what we mean by a ‘model’, and ‘quality’, building on concepts from semiotics and in particular on conceptual model quality. We set out to answer three questions. The first deals with how we measure the quality of a set of Enterprise Architecture models, and to answer this we produce a mathematical framework and then test it using a case study. This extends the conceptual model quality work done by Lindland and Krogstie into the realm of Enterprise Architecture, adding new aspects related to completeness of sets of models, modelling maturity as well as conditions for increasing quality. This incorporates mathematical concepts, including set theory and calculus, and proposes three specific metrics for the quality of sets of models (related to truthfulness, syntax and completeness). This uses a simple case study, based upon purely quantitative data, sampling the contents of an existing Enterprise Architecture repository. The second deals with how we measure the effectiveness of the language used in Enterprise Architecture models. We again use mathematical techniques to construct metrics, this time related to comprehension and utility: the former incorporating a triangulation technique based upon Kvanvig’s concept of moderate factivity of objectual understanding, and the latter being a more subjective measure (i.e. self-assessment). From these two metrics we provide a new conceptual visualisation of the effectiveness of language concepts. We then test this framework using a mixed-mode case study, carrying out 68 interviews, based mostly upon quantitative data again but with additional elements of qualitative data. Although the conceptual framework is independent of any particular language, in order to test it we actually need to select an Enterprise Architecture framework, or more specifically, the modelling language within such a framework; the framework we choose for this purpose is ArchiMate. Through the use of alternative modelling notations in the survey process, we gain insights not just into the understanding and utility of various ArchiMate concepts, as perceived by respondents, we also gain insights into the effect of understanding and utility of using the specific notation provided by ArchiMate through the use of differential analysis of the result sets thus obtained. The final question we address is more practically focused and deals with how we can specify and automate various kinds of changes to Enterprise Architecture models based upon the previous research. We construct a conceptual framework illustrating the kinds of transformations that may be required, given what we have learnt in the previous chapters, demonstrate that these can be deterministic and finally demonstrate, by use of a specific Enterprise Architecture modelling tool (BiZZdesign), that they can be implemented in software, and thus automated. vii In the course of our research, we deliver reusable methodologies and frameworks that will assist future researchers into Enterprise Architecture and related frameworks, as well as Enterprise Architecture practitioners.