An Invertebrate Semantic/Software Application Development Framework
for knowledge-based systems
, Girish R. Ranganathan
University of New Brunswick, 15 Dineen Dr., Fredericton, NB, Canada E3B 5A3
Received 12 March 2008
Accepted 12 March 2008
Available online 21 March 2008
This paper presents an Invertebrate Semantic/Software Application Development Framework (ISADF) for
building knowledge-based software systems. There are two main contributions of ISADF: it
proposes a framework for the development of software for interoperability, and it proposes a software
development approach, where all components, even application logic, are externalized from application.
Ó 2008 Elsevier B.V. All rights reserved.
The continuous change of requirements in the ﬁeld of software
engineering has led to the development of an enormous number of
business applications, which are speciﬁc to the needs of the consum-
ers. This has kept software developers busy with application devel-
opment. Recent achievements in knowledge engineering and rapid
development of the semantic web infrastructure [1,9] allow over-
coming this problem through the creation of frameworks, which
can assist developers in developing business applications. In this
work, an Invertebrate Semantic/Software Application Development
Framework (ISADF) is introduced as an ontology and rule-based
framework to be used by software developers and/or knowledge
engineers for the development of business applications belonging
to various domains. The framework is presented in Fig. 1 and ex-
plained in Section 2. Some of the many business applications which
can be developed using this framework are described in Section 3.
Traditionally software applications are developed on some
backbone, usually application logic. The proposed framework is
called ‘‘invertebrate”, because it assumes that there is no depen-
dent backbone for this framework and software is composed from
external parts. The proposed framework is particularly useful for
knowledge-based systems, because it allows disjoint development
and maintenance of ontologies by knowledge engineers, business
rules by experts (business analysts), and applications by software
developers. Recent achievements in the area of knowledge-based
software for interoperability between schematically and semanti-
cally heterogeneous information sources and consumers, espe-
cially in the business domain, have invoked externalization of
ontologies and business rules from applications. The reason for
such an externalization is that participants of the software devel-
opment process can use comfortable formats. For instance, busi-
ness analysts usually prefer user-friendly table formats (e.g.
Excel) to build business rules; knowledge engineers, however, pre-
fer ontology editors, e.g. Protégé , to build ontologies. Rules
then must be converted to a machine-readable format. The ontol-
ogies and rules are platform-independent; therefore, they can be
reused by any application tailored for particular groups of users.
The idea of externalization of the business logic from application
logic is one of the central points of the Model-Driven Architecture
. The proposed ISADF has some reﬂection on MDA at this point,
but ISADF develops towards the semantic enrichment of business
rules for their machine readability, semantic querying of business
rules, also proposes externalization of the application logic that
makes all participants of the software development process inde-
pendent. It must be noted that the present ISADF is now applied
to knowledge-based software for interoperability – other applica-
tions must be further researched. As a proof of concept, the paper
presents Methods (Section 2) and two Examples (Section 3)to
visualize the methods.
There are multiple challenges in the proposed invertebrate
framework: parsing the source document(s), creation of the do-
main ontology, population and maintenance of the ontology using
information from the source documents, creation and maintenance
of relevant conversion rules in a user-friendly format, transformation
0950-7051/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved.
* Corresponding author.
E-mail addresses: email@example.com (Y. Biletskiy), firstname.lastname@example.org (G.R.
Knowledge-Based Systems 21 (2008) 371–376
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/knosys