Information Technology Reference
In-Depth Information
6 Discussion
Much prior research shows that EA models are realized through a series of actions
by stakeholders, who not only design the models but also use them to transform the
enterprise. The operations described in the paper emphasize this dual nature by
capturing the evolutionary aspects of EA models. They build on the ontology of the
artificial as a foundation, and allow explicit acknowledgement of the progression of
EA models in practice (see, e.g. Grossman 2003; Ross et al 2006). Table 5 reconnects
the arguments so far to the unique aspects of EA models.
Table 5.
The Use of Operations for Conceptual Models of EA (see Table 1)
Dimension
Enterprise Architecture
Use of Operations
What
The underlying logic of a
business
Difficult to specify up-front, must evolve;
also stakeholders can have different
views:
Refinement and Specialization
When
Often
post-facto
, after a business
model is in place in the
organization
Refinement and Linking
allows multiple
stakeholders to take part in the definition
and evolution of EA models
How
A reflection of the architecture
as it currently exists, and with a
vision of a desired state
Decomposition and Specialization
allows fixing responsibility and allowing
stakeholders to participate in modeling
Who
Provides a decision vehicle and
what-if model for the
decision-makers
Decomposition and Derivation
can be
useful to communicate with different
stakeholders
Use
Used as a communication
mechanism to effect change
among stakeholders
Projection
allows stakeholders to use and
vary the different views of EA models
and use Linking to ensure consistency
Life cycle
Moving from extension to
intension, effecting change in
both via iterations
Multiple levels of
Instantiation
including
their reversal allows stakeholders to
refine EA models
Use of frameworks along with
Instantiation and Derivation
allows
consistency across EA models
We do not argue that the operations are complete. Arguing completeness will re-
quire greater scrutiny, and understanding of how the operations are carried out in
practice. Our intent is to distinguish the different modes of action by stakeholders
across the EA life cycle (see, e.g. Smolander et al. 2008).
This paper has argued for appropriateness of an ontology of the artificial for EA
modeling. To the best of our knowledge, no prior work has unpacked the space of
transformations that represent these changes in EA models. Our intent in describing
these operations is not to provide a formal basis for their automation. Instead, we seek
to identify these operations as a way to better understand the decisions and actions
of stakeholders involved in the design and use of EA models. We hope that the opera-
tions we have identified can provide a foundation for further studies such as the
locality of impact, traversing abstractions or construction of macros (e.g. Martin and
Robertson 2008) and identification of evolution patterns, backed by empirical studies.
Abstraction
Explicit use and instantiation
from frameworks with multiple
models; for ensuring compliance
