Information Technology Reference
In-Depth Information
Transforming Enterprise Architecture Models:
An Artificial Ontology View
Sandeep Purao
1
, Richard Martin
2
, and Edward Robertson
3
1
College of Information Sciences & Technology, Pennsylvania State University, PA 16802
spurao@ist.psu.edu
2
Tinwisle Corp., Bloomington Indiana
richardm@tinwisle.com
3
Indiana University and Persistent Systems Inc.
edrbtsn@indiana.edu
Abstract.
Enterprise Architecture (EA) is, by definition, an artificial construct.
It includes conceptual objects and attributes created for human purposes.
EA models, therefore, require an ontological foundation that goes beyond the
'furniture of the world' metaphor. We develop an argument that supports this
premise, and demonstrate how the perspective can help us understand opera-
tions on EA models. The paper demonstrates these operations with an example
and briefly points to formalization efforts detailed elsewhere. The paper
concludes with implications for research and practice.
Keywords:
Enterprise Architecture, Artificial Ontology, Operations.
1 Introduction
Enterprise Architecture (EA) is the logic underlying a business (Ross et al. 2006).
An EA outlines how the different technological, human and organizational elements
within the business are structured, describes how they can be coordinated, and makes
plain possibilities for manipulation for business improvements. Accurate yet mutable
representations are, therefore, important for the practice of EA. These representations
must also include not only the furniture of the world (Bunge 1977) but also concep-
tual constructs created for satisfying human goals (March and Allen 2007). Decision-
makers can use these representations to chart the course for an enterprise; managers
can use these to communicate with stakeholders (Schekkerman 2008).
A number of meta-models, frameworks and standards have been suggested for
EA representations (e.g. Zachman 1987, DoD 2010, ISO Standards 42010, 15704).
The manner in which practitioners use these meta-models and frameworks, however,
remains unclear. Their use of these meta-models and frameworks cannot be the
same as that for conventional software engineering where the designers' aim is to
move from models to executable software. In contrast, EA efforts often include an
archaeological expedition with the intention of charting a new course of action for the
enterprise and facilitating stakeholder buy-in.
The research reported in this paper is presented against the above backdrop. We
explore an alternative ontological basis for EA models and identify transformations
