Information Technology Reference
In-Depth Information
counted and analyzed, however, things such as the use of Object-Oriented
concepts verse not using such concepts is much more capable to subjectivity on
the part of the data collection team.
Most developers will experience a slew of metrics and measurements in the
course of their career. By paying careful attention to the five core values of a
quality metric, it will be possible to determine which measures should be carefully
reviewed and which should be reworked or eliminated entirely (Pressman
2005
).
12.1.7 Establishing Uniform Measures
Measurements not only rely on consistent data collection and rigorous analysis,
but they must also rely on comparative scales of the proper nature (Card and Glass
1990
):
• Nominal scales provide discrimination between two objects based on simple and
basic characteristics such as color or shape (rounded or square).
• Ordinal scales rely on some sort of ranking, shortest to tallest or heaviest to
lightest.
• Interval scales have the characteristic that the distance between two points
signifies a degree of distance proportional to the distance.
• Ratio scales are similar to interval scales, but have a fixed absolute zero point.
Human error and bias is one of the main contributors of non-uniform measures.
With proper training and experience software engineers can curb their biases and
create consistent and uniform measurements. The use of software measurement
frameworks, however, facilitate the most accurate and consistent measurements
possible.
Software measurement frameworks allow for a consistent and complete soft-
ware development lifecycle, data collection cycle to occur. By using a software
measurement framework, you are assured that data is collected without bias and
uniformly amongst all projects implementing it. A software framework collects
metrics from the code base itself, as well as from the developers during the
development process, in an automated fashion. A good framework is able to
collect and format data in a fashion that is consistent with the requirements of both
the stakeholders and developers (Zuse
1997
).
12.2 Quality Metrics
Software quality is one of the most subjective terms available when determining the
goals of an organization. Pressman defines software quality as the following ''an
effective software process applied in a manner that creates a useful product that
