SOFTWARE OPTIMIZATION TECHNIQUES - Software Design for Six-Sigma: A Roadmap for Excellence - page 443

TABLE 17.1 Summary of Optimization Metrics

Type of Metric Comments Ranking

Cohesion High cohesion is an indication of a well-designed

system

1

Coupling Low coupling is an indication of a well-designed

system

1

Cyclomatic Complexity Only measures complexity as a function of control

flow

2

Halstead's Metric Difficult computations as well as questionable

methodology for obtaining some mathematical
relationships

2

Function Point Dominant metric for some types of economic and

quality studies. However, some software
applications so large that normal function point
analysis is too slow

2

Lines of Code Outdated, has not been used widely since the

mid-1980s

3

some software applications are now so large that normal function point analysis is too
slow and too expensive to be used. Second, as of 2008, there are at least 24 function
point variations in which the number of variations tends to make baseline studies
difficult.
The next optimization metric, cyclomatic complexity, also has drawbacks, as it
only measures complexity as a function of control flow. Instead, Halstead's metrics
are suitable for measuring how intensely the programming language is used. However,
the Halstead metric has been criticized for its difficult computations as well as its
questionable methodology for obtaining some mathematical relationships.
In contrast to LOC, function point, cyclomatic complexity, and Halstead's metric,
some simpler metrics to use are cohesion and coupling. Indeed, high cohesion com-
bined with low coupling is a sign of a well-structured computer system and a good
design. Such a system supports the goals of high readability and high maintainability.
Table 17.1 is a table summarizing each optimization metric, comments, and a ranking
in which 1 is the best, 2 is average, and 3 is worst. As seen in Table 17.1, both cohe-
sion and coupling rank the highest, followed by cyclomatic complexity, Halstead's
metric, function point analysis, and LOC.
Therefore, although there are many types of optimization techniques on the market
today, some of the best optimization techniques are probably cohesion and coupling.

17.3.1 Response Time Techniques

Response time is the presentation of an input to a system and the realization of
the required behavior including the availability of all associated outputs (LaPlante,
2005). Response time is important in real-time applications because it estimates the
maximum amount of time until an event, such as when a communication from another