Abstract: The rapid development of mobile technologies has facilitated users
to generate and store files on mobile devices such as mobile phones and PDAs.
However, it has become a challenging issue for users to efficiently and
effectively search for files of interest in a mobile environment involving a
large number of mobile nodes. This paper presents SemFARM framework which
facilitates users to publish, annotate and retrieve files which are
geographically distributed in a mobile network enabled by Bluetooth. The
SemFARM framework is built on semantic web technologies in support of file
retrieval on low-end mobile devices. A generic ontology is developed which
defines a number of keywords, their possible domains and properties. Based on
semantic reasoning, similarity degrees are computed to match user queries with
published file descriptions. The SemFARM prototype is implemented using the
Java mobile platform (J2ME). The performance of SemFARM is evaluated from a
number of aspects in comparison with traditional mobile file systems and
enhanced alternatives. Experimental results are encouraging showing the
effectiveness of SemFARM in file retrieval. We can conclude that the use of
semantic web technologies have facilitated file retrieval in mobile computing
environments maximizing user satisfaction in searching for files of interest.
Keywords: Mobile computing, semantic, pervasive, ontology, file retrieval, SemFARM, ubiquitous
