readable base holding the process’
documentation (RUP documentation on the
Web). Let us name this knowledge set as a
“Template”.
ii. The second knowledge base that holds an actual
software process that is used in an organization.
This knowledge base must be filled by an
ontology expert in processing and consulting
services using the same ontology dictionary as
the “Template”. The name of this set will be the
“Actual”.
Now, when we have two knowledge bases whose
contents are documented and real software processes
we can use the function from Definition 3.4 to
evaluate the similarity of them.
An individual content comparison can be
performed as an add-on to the assessment. We can
search for differences directly between individual
members of both sets. However, this means basically
a brute-force comparison of two sets, which is not
optimal. Results of such a comparison can tell us
where gaps are in the “Actual” knowledge base of a
process compared to “Template”. The way to search
for such differences is the main task of our future
research.
5 CONCLUSIONS AND FUTURE
WORK
In this paper, we presented our new approach for the
semi-automated assessment and evaluation of
software process similarity. The comparison of the
real and reference software processes is done by the
usage of machine readable knowledge bases. The
template (reference) knowledge base describes the
reference software process and the actual knowledge
base (assessed process) describes the software
process in the company. The template knowledge
base is created in advance, using specific ontology
for the particular software process and other
techniques that are necessary to build a machine
readable knowledge base. The actual knowledge
base is created during the assessment process. Both
knowledge bases are then compared and the result is
the number that represents percentage similarity.
This presented process is one part of our
comprehensive approach to the assessment,
evaluation and improvement of software processes.
The first step is the extension of the presented
process to the automated comparison of specific
parts of software processes. The next steps are then
e.g. automated comparisons with more than one
specific process at once, involvement of process
modeling to the approach that will be used for the
automated search, evaluation and improvement
suggestion using the template software process
models etc… A lot of future work is needed to solve
all the problems that arise during the development of
this new approach. Our work is also supported by
the experience that is gained through the practical
experiments of this approach in real software
companies.
Although, according to our preliminary use cases
studies, this approach seems to be very promising,
the further use case studies are needed to
continuously develop and enhance the approach and
support its inclusion into the software process
assessment models and improvement techniques.
ACKNOWLEDGEMENTS
This research has been supported by internal grant
agency of FEECS VSB-TU Ostrava - IGA 22/2009
Modeling, simulation and verification of software
processes. Author Michal Košinár is also grand
aided student of Ostrava City Authority, Czech
Republic.
REFERENCES
Alexandre, S. Makinen, T., and Varkoi, T. Implementation
of a Software Process Standard as an Electronic
Process Guide. In proceedings of SPICE 2008
Conference (Software Process Improvement and
Capability dEtermination), 26-28 May 2008,
Nuremberg, Germany.
Ronald Brachman and Hector Levesque. Knowledge
Representation and Reasoning (The Morgan
Kaufmann Series in Artificial Intelligence). Morgan
Kaufmann, May 2004.
Ciprich, N., Duží, M., Košinár, M.: TIL-Script: Functional
Programming Based on Transparent Intensional
Logic. In RASLAN 2007, Sojka, P., Horák, A., (eds.),
MasarykUniversity Brno, 2007, pp. 37-42.
Ciprich, N., Duží, M. and Košinár, M.: The TIL-Script
language. In the Proceedings of the 18
th
European
Japanese Conference on Information Modelling and
Knowledge Bases (EJC 2008), Tsukuba, Japan 2008.
Frydrych, T., Kohut, O., Košinár, M.: Transparent
Intensional Logic in Knowledge Based Multiagent
Systems. In RASLAN 2008, Sojka, P., Horák, A.,
(eds.), MasarykUniversity Brno, 2008.
Dragan Gasevic, Dragan Djuric, and Vladan Devedzic.
Model Driven Architecture and Ontology
Development. Springer, July 2006.
ENASE 2010 - International Conference on Evaluation of Novel Approaches to Software Engineering
110