Information Technology Reference
In-Depth Information
In order to measure processes, we measure the outcomes of said processes to
gain insight into their effectiveness. Prior to the start of a new project, the team
will often measure the results of previous projects to establish baseline measure-
ments for estimation and goal establishment. These metrics will often include
measuring errors, defects, time expended, productivity, etc. These measurements
describe how effective the efforts and software development process was for past
projects. Once completed, the project's errors, defects, etc. will also be measured
and added to the records. These measures allow us to derive process metrics that
describe the efficiency of the processes used to complete these attributes.
In his topic, Pressman discusses the use of both private and public process
metrics. He explains that process data should begin at the individual level to be
used privately by the given individual. Private process data of this form is intended
to be used for individual improvement. Just above this level of privacy are metrics
which are private to a team and public to all members of the team to be used for
improving team performance. Metrics at this level may include errors and defects
for the entire software project or major parts of it while individual private process
data may include such measurements for their specific components and tasks.
Public process metrics are those visible to many, if not all, members within the
organization. These metrics often include the effort expended, defects found, team
size of, and other data for the project in its entirety (Pressman
2005
).
12.7 Post Release Metrics
Once completed, existing software can benefit from the use of metrics for the
purpose of maintenance. The software maturity index (SMI) proposed by IEEE
Std. 982.1-1988 (Pressman
2005
) is a metric to designate the stability of the
software being maintained. The calculation for SMI includes the following:
• M
T
= # of modules in current release
• F
c
= # modules in current release that have changed
• F
a
= # modules in current release that were added
• F
d
= # modules from preceding release that were deleted in current release
• SMI = [M
T
- (F
a
? F
c
? F
d
)]/M
T
An SMI approaching 1.0 is considered to be near stable.
12.8 Chapter Summary and Conclusions
Software metrics play an important role in every aspect of software engineering as
they provide a quantitative, objective view regarding the quality of the various
aspects of a system. Before the introduction of specific metrics, this chapter
explained the key differences between private and public metrics, between what
