Abstract:
Software reuse is a popular way of addressing software development issues—such as high
cost of development, low productivity and poor quality. However, organizations are yet
to realize maximum payoff from reuse due to reusability related issues. Reusability can
only be improved through measurement, because measurement is the only sure way of
monitoring and improving software quality. The goal of this study was to present a reliable
metrics-based framework for measuring the reusability of object-oriented (OO) software
components. A survey involving OO developers was conducted, where methods that they
currently use in assessing reusability were examined. The methods in question include
checking of source code, reading documentation, intuition, and checking of comments.
These methods were found to be largely subjective, hence not reliable. In addition, four
reusability assessment frameworks—found in literature were examined, and found to have
various challenges: all of the frameworks lack predictive power, hence they cannot be
applied at early stages of software development; three of the frameworks include
traditional metrics, which cannot be used to measure OO software components; and lastly,
one of the frameworks was platform dependent. The conclusions about the inefficiencies
of the reviewed reusability assessment methods and frameworks were arrived at,
following a comprehensive literature analysis, through which a criterion for evaluating the
efficiency of reusability assessment frameworks for OO software was defined. Literature
analysis revealed that an effective reusability assessment framework for OO software
should include the following key elements: major reusability attributes, factors that
influence the reusability attributes (reusability factors), measurable OO design constructs
that influence the reusability factors, and OO metrics for measuring the OO constructs.
This research culminated into the development and implementation of a framework that
conforms to this criterion. The developed framework was validated for superiority over
the existing reusability assessment methods and frameworks through a comparative
analysis.
