The evolvability of software systems is one of the key issues when
considering their long term quality. Continuous changes and extensions of
these systems are neccessary to adjust them to new or changing
requirements. But the changes often cause quality deficiencies, which lead
to an increase in complexity or an architectural decay. Especially quality
deficiencies within the specification or the architecture of a software
system can heavily impair a software system.To counteract this, a method is
developed in this work to support the analysis of a quality goal in order
to identify the quality deficiencies which hinder the achievement of the
quality goal. Both the detection and the removal of quality deficiencies
are accomplished in a systematic way. The method integrates detection of
these quality deficiencies and their removal by reengineering activities
based on rules. The detection of quality deficiencies is performed by means
of measurable quality attributes which are derived from a quality goal,
such as evolvability. In order to demonstrate the practicability of the
method, the quality goal evolvability is taken as an example. This work
shows how a software system can be evaluated with regard to evolvability
based on structural dependencies and which reengineering activities will
improve the system in the direction of this quality goal.To evaluate the
method, it was applied within an industrial case study. By analyzing the
given software system a large number of different quality deficiencies were
detected. Afterwards the system's evolvability was improved substantially
by reengineering activities proposed by the method.