F.-P. Yang, C.H. Kao, and H.C. Jiau (Taiwan)
Design principle, perfective maintenance, experimental design, concept analysis, software metrics, statistics.
Maintenance effort becomes a major burden of many
companies. Software developers hope to consider the
maintenance related issues early in development phase.
They can reach this goal by applying design principles.
Applied design principles usually lead to higher quality of
software, and such higher quality often reduce the main
tenance effort dramatically. However, many other factors
are not able to control through applying design principles,
and their confounding effects might also be significant on
software maintenance. This experiment conducted for an
alyzing the impact of design principles on the maintenance
effort under a varied confounding factor--the complexity
of collaboration complexity. In addition to analyze the ef
fect of these two factors separately, a global analysis is car
ried for observing the compound effect. Finally, this exper
iment used a finer granularity analysis that reflects the two
phases, understanding and modifying, of maintenance pro
cess to get a more operational insight of the effect. Man
agers may decide to introduce some phase-specific tools
or methodologies based on the measured results. All the
analyses in this experiment are conducted by a technology,
which is based on concept analysis, impact analysis and
t-test, called CIT.