Information Technology Reference
In-Depth Information
between requirements, we introduce an ontology-based approach for reporting analy-
sis results on a set of requirements, so-called ontology-based reporting.
The goal of the ontology-based reporting approach
OntRep
is making ReqM tasks
like requirements categorization, conflict analysis and requirements tracing more
efficient based on the automation of selected steps in these tasks. The following sub-
sections provide an overview on the approach and motivate research issues.
Fig. 1.
Components and numbered steps of OntRep
We developed a prototype tool for the
OntRep
approach a plug-in to
Trac
3
, an open
source collaboration platform consisting of a Wiki, ticket management system, and
subversion integration, which can be extended by Python plug-ins.
Fig. 1 illustrates the
OntRep
tool (together with
Protégé
) consisting of two main
components: 1)
Instance fetcher
takes input data, e.g., requirement tickets from
Trac
,
analyzes their contents and assigns them requirements categories (classes) defined in
the ontology; 2)
Reporting component
reasons on the input data and generates a re-
quirements conflict report based on the analyzed requirements.
3.1 Semantic Requirements Categorization
In a first phase natural language texts have to be linked to semantic categories as
preparation for further analysis and reporting. The following steps automate require-
ments categorization with OntRep (see numbered circles in Fig.
1
):
1)
Define the requirement categories
in Protégé, e.g., categories X, Y, Z. Each cat-
egory is defined as an ontology class in Protégé. It is important to define project-
relevant “semantic” categories and not formal ones in order to enable the automated
3
http://trac.edgewall.org/
