Rutas Cortas Colonia De Hormigas

ISO/IEC 9126 in practice: what do we need to know?
P. Botella, X. Burgués, J.P. Carvallo, X. Franch, G. Grau, J. Marco, C. Quer

Abstract
ISO/IEC 9126 is currently one of the most widespread quality standards. In its actual form
it embraces both quality models and metrics. Due to its generic nature, some of the concepts
presented therein need to be refined before using the standard in areal project. In our paper,
we aim at exploring which are the concepts that require being more elaborated before
putting the standard to work. Specifically, with respect to the part 1 of the standard we focus
on the hierarchical form of the quality entities (i.e., characteristics, subcharacteristics and
attributes) that appear in quality models; we propose some criteria to distinguish amongsubcharacteristics and attributes; we distinguish among derived attributes and basic
attributes; and we propose an extension for covering not only technical factors but also
managerial and political ones. Concerning the other 3 parts of the standard (one of them
currently pending of approval), we distinguish different categorization criteria of metrics
(scale, type, objectivity,qualitative/quantitative) and we make explicit the rationale behind
their definition with the Goal-Question-Metric approach. As a final contribution, we show
the use of UML class diagrams for representing all the concepts of the standard and their
relationships, and we highlight the need for having tool support for quality model
development and metrics definition.

1. Introduction
Software quality models [1]are artefacts used for describing the quality factors of a single
software product of any nature (e.g., a bespoke application, a commercial off-the-shelf
component or a web service) or a software domain (e.g., ERP systems or document
management tools). A software quality model provides a taxonomy of software quality
factors. Software quality models and their metrics may be used in manycontexts, for instance
during the development of a new application [2, 3] or when selecting commercial components
[4].
There are many approaches to deal with the identification and representation of software
quality factors. All of them rely on the existence of a catalogue of factors which in general is
defined as extensible. There is also a great variety of those catalogues, many of them in theform of standards. In this paper, we focus on one of such standards, the ISO/IEC 9126 quality
standard [5], which we have used in our experiences.
The ISO/IEC 9126 is currently one of the most widespread quality standards. In its actual
form it embraces both quality models and metrics. Due to its generic nature, some of the
concepts presented therein need to be refined before using the standardin a real software
procurement project. In our paper, we identify those issues that require clarification before
putting the standard to work, and we highlight also the role that conceptual models and tool
support may play. Specifically, our paper addresses the following points:
• Some concepts appearing in the standard are not clearly defined. When we started to use
the standard, this factmade harder return on investment and quality model reuse.
Therefore we felt compelled at some time to define clearly each concept of the standard
and the relationships among them.

In addition to quality factors, there are other type of factors that are utterly important
when procuring software systems. We have found that these factors may be managed in a
similar manner as quality ones,providing therefore a unified framework that has revealed
to be comfortable during software procurement.
• The elements of a software metric were not made clear in the standard definition. For
instance, it was not clear what type of metric scales were allowed, nor which classification
the standard considers. We have decided then to make this information also explicit.
• The definition of the...