The DBin platform: A complete environment for Semantic Web Communities☆
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.
Section snippets
The Semantic Web Communities model: high level system architecture and user experience
In our system we distinguish between two different user’s behaviours: they might simply want to existing Semantic Web Communities, thus being able to cooperatively build the community semantic knowledge (which we call end users), or might be interested in starting up and/or maintaining communities (power users). The power user starts up a new community by first creating a customized user environment, called Brainlet, for the editing and exploitation of semantically structured annotations.
P2P and Semantic Web: related works and RDFGrowth
The P2P model as a transport medium for RDF has been investigated in several previous works. Edutella, described in [1], allows distributed queries within a federation of peers and has been later extended [3] to improve its scalability by introducing schema-based routing and clustering. RDFPeers, discussed in [2], is an other interesting approach to build scalable distributed RDF repositories. A P2P publish/subscribe system, as an alternative to explicit query based approaches, is described in
MSGs
Before describing DBin’s support for information filtering and revision, let us introduce the concept of Minimum Self-contained Graph (MSG). The RDFN of a resource (introduced in Section 3.2) can be decomposed in smaller pieces, called MSGs, which represents the minimum amount of RDF knowledge that can be exchanged in RDFGrowth. A formal definition of MSGs and their theoretical properties is given in [6]; intuitively, these are fragments of RDF graphs composed by a starting triple and, in the
Related works
RDF data visualization has a central role in Semantic Web research and many approaches have been proposed so far. Some of them use graphically represent RDF in the form of a graph, like RDF Gravity5 and IsaViz [9]. In general this approach suffers from difficulties in understanding and browsing data where the dataset is very large and connected even if sub-graphs browsing facilities partially
User survey and validation
To address the validation of the DBin application and scenario, we conducted a user survey. The user group was composed by 20 people. Most of the users reported at least some previous contact with Semantic Web ideas or tools. Some participants were experts with clear knowledge of the field, while others were professional figures, however selected for having familiarity with advanced software (e.g., Engineers, Web designers, etc.). The recruitment happened via email announcements on public
Related works to the DBin project
In this paper we have been so far presenting related works in a number of previous occasions. This has been functional to the explanations of the individual infrastructures, e.g., GUI, P2P. In this section we instead discuss the previous works which can be related to the DBin project in a more general sense.
During the last years many attempts have been done in creating applications that could show the users the Semantic Web in action. Piggy Bank [17], is an MIT project which consists in a
Conclusions
In this paper we presented a comprehensive overview and a first evaluation of the Semantic Web Community scenario enabled by the DBin platform.
DBin uses P2P to exchange the knowledge collectively authored by the communities of users via topic specific user interfaces named Brainlets. Brainlets and P2P channels can be configured, deployed by domain experts and then easily discovered, installed and used by people. The knowledge collaboratively created within communities can then be made available
References (21)
- et al.
EDUTELLA: a P2P networking infrastructure based on RDF
- et al.
RDFPeers: a scalable distributed RDF repository based on a structured peer-to-peer network
- et al.
SuperPeer based routing and clustering strategies for RDF based peer-to-peer networks
- et al.
Publish/subscribe for RDF-based P2P networks
- et al.
RDFGrowth, a P2P annotation exchange algorithm for scalable Semantic Web applications
- et al.
Signing individual fragments of an RDF graph
- et al.
SeRQL: a second generation RDF query language
Signing RDF graphs
Isaviz: a visual environment for browsing and authoring RDF models
- et al.
Fresnel: a browser-independent presentation vocabulary for RDF
Cited by (8)
Theoretical B2B knowledge management framework focused on value co-creation
2024, VINE Journal of Information and Knowledge Management SystemsModeling an web community discovery method with web page attraction
2021, Journal of Intelligent and Fuzzy SystemsCurating a document collection via crowdsourcing with Pundit 2.0
2015, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)Consistency criteria for a Read/Write Web of Linked Data
2014, CEUR Workshop ProceedingsAn efficient method for community detection based on formal concept analysis
2014, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
- ☆
This work has been partially supported by the European FP6 project inContext (IST-034718), by Science Foundation Ireland under the Lion project (SFI/02/CE1/I131), and by the European project DISCOVERY (ECP-2005-CULT-038206).