The DBin platform: A complete environment for Semantic Web Communities

Abstract

DBin is a Semantic Web application that enables groups of users with a common interest to cooperatively create semantically structured knowledge bases. Such Semantic Web Communities are made possible by creating customized user environments called Brainlets. Brainlets provide user interfaces and domain specific tools (e.g., querying, viewing and editing facilities) that enable community participants to interact with the data of interest. Brainlets are directly created by domain experts using an XML description language. DBin clients communicate and exchange annotations using a P2P infrastructure. Access control and digital signatures, put by DBin inside the authored RDF, enable trust and information filtering.

Introduction

The W3C Semantic Web initiative has been active for a consistent number of years, and Semantic Web programming tools and libraries have reached a certain maturity. It has been widely noticed, however, that very few, if any, applications are available today for the end user to clearly experience at least some of the promises of the Semantic Web vision. In this article we describe the DBin project,¹ which aims at creating a Semantic Personal Knowledge Manager (S-PKM) with the following main features:

• Being based on Semantic Web languages and usable in different domains by applying specific ontologies and settings, but yet enabling users to have a merged view of all the knowledge pertaining to different real world domains.

• Making use of ontology-based reasoning, whenever possible, for assisting the user in visualizing, editing and browsing semantic data.

• Working as a personal information manager and being integrated with the local desktop environment.

• Being interconnected with other S-PKM installations and with external data sources, thus enabling collaborative semantic knowledge authoring.

• Being powerfully adaptable to different domains and communities without the need for programming. Domain experts, rather than programmers, should be able to create domain specific applications on top of the platform and deliver them to end users in a simple, integrated, intuitive way. These domain specific applications should possibly co-exist in the same SPKM installation, interact among each other and share data.

In our opinion, seeking the realization of such an integrated tool is important. It serves both to validate the individual Semantic Web components as useful in large use cases, and to possibly discover the need for new components and infrastructures.

In developing DBin it became evident that, on top of the existing tools, there was the need both for a number of novel infrastructures and for pragmatic decisions, that in some cases would limit the excessive freedom, inherent in the Semantic Web vision, in favor of actual usability. Some of the topics which required novel solutions in terms of infrastructural components has been:

• Transport layer; this deals with the problem of publishing, importing and discovering semantic web knowledge.

• Authorship and trust; how the semantically structured information are digitally signed and how this enables personalized filtering.

• Semantically structured information visualization and domain specific, highly customizable user interfaces.

• Agreement on domain ontologies and entity names.

The first two issues are dealt in DBin with new methodologies and subsystems. In Section 3 we will overview the RDFGrowth P2P algorithm, that enables topic based sharing of data, and the way such semantic data can be published and retrieved from the Web. Furthermore, DBin offers a digital signature methodology enabling information authorship verification and local trust based filtering policies (Section 4).

DBin, on the other hand, pragmatically approaches the remaining issues with its Brainlet plug-in model. Brainlets are introduced in Section 2, where the overall scenario is described, and then discussed in more detailed in Section 5. Finally, in Section 6, we present the results of a user survey to verify reaction to the novel scenario and application model.

DBin has been completely implemented in Java, on top of the Eclipse RCP² framework, and therefore is platform independent and has a plug-in structure which supports agile extensibility. Current releases can be downloaded from the project’s web site.