A Software Engineering Framework for Context-Aware Service-Based Processes in Pervasive Environments - Advanced Design Approaches to Emerging Software Systems - page 126

FUTURE RESEARCH DIRECTIONS variabilities; and provides application developers

with context-specific programming constructs to
express their needs from context information. The
result is a more intuitive way to represent context
and improve overall systems performance.
On the other hand, we have described the Apto
model-driven approach for managing and gener-
ating process variants. One of the advantages of
using MDD is that the context management and
adaptation logic are included in models rather than
directly implemented in code. Based on logically-
viewed well-defined evolution fragments and evo-
lution primitive constructs; on the ability to group
evolution fragments in reusable components; and
on the ability to regroup these components in a con-
strained way, necessary adjustments of the basic
process can be correctly and easily realized when
creating or configuring a process variant. We have
adopted the viewpoint that this kind of adaptation
can often be considered as a crosscutting concern
with respect to the core application logic. Hence,
one of our main goals has been the decoupling of
the design and implementation of the adaptation
logic from the design and implementation of the
main process logic. Finally, Apto allows for the
dynamic configuration of process variants based
on the given process context.

Future work includes extending the proposed
context management framework to the distributed
context management architecture. Therefore, for
the purpose of interoperability, we need a formal
common semantics for context feature models
managed by different context servers.
In addition, in order to achieve the possibility
of making deep changes we intend in our future
work to extend the Apto idea to regroup different
process views' models. Indeed, as the number of
services or processes involved in a process grows,
the complexity of developing and maintaining
these processes also increases. One of the suc-
cessful approaches to managing this complexity is
to represent the process by different architectural
views. Examples of these views are collaboration
view, information view, orchestration view etc.
The idea is to give the developer the possibility
of applying the necessary evolution fragments in
each view and then the automated tool verifies the
integrity of the changes and generates the adapted
process variant artifacts accordingly. This involves
tackling the correct combination of evolution
fragments when creating a variant. Sophisticated
techniques are needed to prevent errors (e.g.,
deadlocks) or other consistency problems.

REFERENCES

CONCLUSION

Baldauf, M., Dustdar, S., & Rosenberg, F. (2007).
A survey on context-aware systems. International
Journal Ad Hoc and Ubiquitous Computing, 2 (4).

Change is the only constant in the software/service
development world due to the evolution in business
or user context and requirements. Therefore, there
is a need to customize processes by generating a
process variant that corresponds to the change
in the business and user requirements. We have
presented an approach for supporting context-
aware applications based on a flexible product line
based context model. The proposed approach to
model the context information allows the context
modeler to specify the context information in a
high-level and logical way that regroups context

Chen, H., Finin, T., & Joshi, A. (2004). An ontology
for context-aware pervasive computing environ-
ments. The Knowledge Engineering Review , 18 (3),
197-207. doi:10.1017/S0269888904000025

Czarnecki, K., Hwan, C., & Kalleberg, K. T.
(2006). Feature models are views on ontologies.
In Proceedings of the 10th International on Soft-
ware Product Line Conference (vol. 1). IEEE
Computer Society.