by Lucas Sarmento Santos

lucas_sarmento — 2015-01-26 15:31:03 UTC

Software metrics continue to be of interest for researchers

and practitioners. Metrics such as volume [3], McCabe Complexity

[26], the C&K metric suite [12] and a wide range

of architecture metrics (see Koziolek [23] for an overview)

are well-known and used in practice.

lucas_sarmento — 2015-01-26 15:31:03 UTC

Software metrics continue to be of interest for researchers and practitioners.

Metrics such as volume [3], McCabe Complexity [26], the C&K metric suite [12]

and a wide range of architecture metrics (see Koziolek [23] for an overview)are well-known and used in practice.

lucas_sarmento — 2015-01-26 15:34:38 UTC

the usefulness of two architecture level metrics, Component Balance [7]

and Dependency Profiles [8], which are designed to quantify the analysability

and encapsulation within a software system.

lucas_sarmento — 2015-01-26 15:37:01 UTC

The focus of these types of evaluation is to determine

whether the designed metric is related to the quality property it

as been designed to quantify, a property known as “construct

validity” [21].

2015-01-26 15:38:15 UTC

To present a convincing argument regarding the benefits of using metrics, one

needs to highlight the incentives and payoff. William T. Ward [11] describe

Hewlett-Packard’s (HP’s) “10 x software quality improvement” initiative. The

author uses data from a software metrics database and an industry profit-loss

model to develop a method to compute the actual cost of software defects.

The database is an important element of HP’s software quality activities and is

a valuable source for different tasks such as quality status reporting, resource

planning, scheduling, and calculation of cost defects. Sources of data include

product comparisons, analysis of source code size and complexity, defect

logging, project post-mortem studies, and project schedule and resource

plans.

lucas_sarmento — 2015-01-26 15:40:34 UTC

To summarize, this paper makes the following contributions:

• It introduces a methodology for evaluating the usefulness

of software metric in industry

• It describes the execution of this methodology in an

empirical study towards understanding the usefulness of

two specific software metrics

• It provides an overview of challenges involved in the

application of the two specific software metrics and lists

concrete areas for improvement

• It reflects upon the usefulness of the evaluation methodology,

concluding that the relative importance of challenges

involved in applying specific metrics cannot be

determined in a purely academic setting.

2015-01-26 15:56:59 UTC

Why software metrics is important

To achieve accurate measurements of
productivity and quality requires metrics collections and analyse
To indicate the quality of the
product
To assess the benefits derived from
new software methods.
To help justify request for new tool
or additional training

To Form a baseline for estimation