Information Technology Reference
In-Depth Information
Revisiting Naur's Programming as Theory
Building for Enterprise Architecture Modelling
Balbir S. Barn and Tony Clark
Middlesex University, Hendon, London, UK, NW4 4BT
b.barn@mdx.ac.uk
,
t.n.clark@mdx.ac.uk
Abstract.
The recent burgeoning interest in Enterprise Architecture
and its focus on artifact driven methods is taken as a motivation for the
re-appraisal of Peter Naur's notion of “programming as theory building”.
Naur strongly disputes the value of the role and orientation of the IS
discipline around artifacts and argues that algorithmic methods do not
lead to a theory of a domain. Such a viewpoint provides an alternative
lens with which to view current developments and may lead to addi-
tional insights by providing the reader with a source for questioning and
reflecting critically on the re-focusing of method design on conversation
rather than artifact production . It is suggested that such a conversa-
tional framework based on Toulmin and Pask may provide a means to
establish and test theory building views of enterprise architecture.
1
Introduction
This account sets out to re-appraise Peter Naur's influential paper on Program-
ming as Theory Building [11] in the context of model building and the recent
focus on Enterprise Architecture. It is the intention of this paper to evaluate how
theory building can play an important role in helping organizations make more
use of their enterprise architecture activity and in particular how theory building
may influence methods, techniques and tools to support enterprise architecture
by focusing on conceptual modelling as a conversation process.
The starting point for this work has been triggered by the extent of activity that
is currently surrounding Enterprise Architecture. As systems supporting business
become increasingly more significant and complex an important approach to man-
agement and planning of systems that has gained prominence is model-based
Enterprise Architecture (EA). EA has its origins in Zachman's original EA frame-
work [21] while other leading examples include the Open Group Architecture
Framework (TOGAF) [17] and the framework promulgated by the Department
of Defense (DoDAF) [19]. In addition to frameworks that describe the nature of
models required for EA, modeling languages specifically designed for EA have also
emerged. One leading architecture modelling language is Archimate [7].
Central to enterprise architecture is the notion of development and presen-
tation of models. Given the plethora of models available two concerns of note
arise: Firstly, given the range of models available, it is dicult to ascertain why
a particular model is relevant and preferable over others. This arises from a
