9 Measuring external product attributes


A. Törn - Contents - - Previous chapter - Next chapter - - Previous page - Next page

The objective of this chapter is to focus on a small number of especially important external attributes. We brgin by considering several general software quality models. The models are then used to identify key external attributes of interest, including reliability, maintainability, and usability.

9.1 Modeling software quality

Because quality is really a composite of many characteristics, the notion of quality is usually captured in a model that depicts the composite characteristics and their relationships.

The oldest models are those of McCall (1977) and Boehm et al (1978), see Figures 9.1 and 9.2. Both models are hierarchical with successive refinements ending upp in primitives that are directly measurable, see Figure 9.3. Both are starting out from software use divided into three non-overlapping parts namely (naming from McCall):

Eriksson and Törn starts out not from software but from the parties involved with information systems (IS) and their main concern:

These models are example of static models. The use is described in Example 9.2.

Other models are defined separately for each project by identifying key attributes, see Figure 9.4.

The need to compare products led to work to define a standard model. One result of this based on the McCall model is ISO 9126: Software Product Evaluation: Quality Characteristics and Guidelines for their Use, describing quality as decomposed of: functionality, reliability, efficiency, usability, maintainability, and portability.