Semantic Requirements Engineering - Intentional Perspectives on Information Systems Engineering - page 68

should know the meaning of elements such as Message and Object included in
sequence diagrams, in order to write the diagrams easy to understand. Thus, we
can consider the semantics of elements of sequence diagrams, i.e. the semantics of
a meta model of sequence diagrams, and this is the case where our subject domain
is meta models, not a problem domain of lift control.
In this chapter, we propose one of the techniques to provide semantics to require-
ments and RE techniques. We consider an ontology as a semantic domain so as
to provide the meaning for requirements and discuss the potentials of the RE
techniques using an ontology as a semantic basis. More concretely, we adopt the
technique of denotational semantics and consider mappings from artifacts (incl.
requirements) to the ontology. An ontology consists of a thesaurus and inference
rules on it, and the thesaurus includes the words and their relationships. Each word
in the thesaurus is frequently used in a certain subject domain and it denotes an
atomic semantic element that has a unique meaning in the domain. We can map
artifacts to the words in the thesaurus and the meaning is provided for the arti-
facts by the mapped words. As a result, we can reason about semantic properties of
the artifacts by using the inference rules on the words. This is a basic idea in this
chapter. As a result, we can have a light-weight semantic processing of artifacts for
assisting RE activities, e.g. semantic consistency checking of requirements, retriev-
ing the requirements that are semantically similar as reusable components, etc. Our
technique is inspired by the technique of Semantic Web [ 20] , where HTML texts
are annotated with semantic tags derived from ontological components. In Semantic
Web, the meaning of information on the web is defined using the annotated semantic
tags so as to make it possible to analyze and process the web contents by computer.
The semantic tags represent the meaning of information where they are annotated.
In our technique, any kind of artifacts related to RE process, including specified
requirements with natural language sentences or UML diagrams, meta models, etc.,
are annotated with (mapped to) semantic tags (ontological components) so that the
existing RE techniques can semantically process them by computer. Thus we can
use the word Semantic Requirements Engineering by the similarity to the idea of
Semantic Web.
The rest of the chapter is organized as follows. The next section presents the
essential idea of our approach more concretely. As a concrete application, in Sect. 3
we show an extended version of goal-oriented requirements analysis (GORA) tech-
nique combined with ontologies. GORA techniques are for eliciting and specifying
requirements and one of the major RE techniques supporting the first and the second
in the four activities mentioned above. In Sects. 4 and 5, we illustrate the other appli-
cations of our technique in a different activity, requirements management. Section
4 presents metrics to measure the quality of requirements specifications consider-
ing their meaning, while we discuss the version control based on the meaning of
requirements, so called semantic version control. We would like to show how our
technique using ontologies can reinforce the existing RE techniques in each of the
four activities, throughout the whole of RE processes.