Browsing Linked Data with Fenfire Tuukka Hastrup Richard Cyganiak Uldis Bojārs University of Jyväskylä Digital Enterprise Research Digital Enterprise Research Jyväskylä, Finland Institute Institute Tuukka.Hastrup@iki.fi National University of Ireland, National University of Ireland, Galway Galway richard.cyganiak@deri.org uldis.bojars@deri.org ABSTRACT tion provides a useful Linked Data browsing experience, go A wealth of information has recently become available as through an example, describe the implementation and con- browsable RDF data on the Web, but the selection of client clude with some future directions. applications to interact with this Linked Data remains lim- ited. We show how to browse Linked Data with Fenfire, a Free and Open Source Software RDF browser and editor that employs a graph view and focuses on an engaging and 2. BROWSING EXPERIENCE interactive browsing experience. This sets Fenfire apart from The Fenfire application is a generic RDF browser and editor previous table- and outline-based Linked Data browsers. with features useful for Linked Data browsing. The user interface employs the conventional graph representation of the RDF data model. To make the visualisations scalable 1. INTRODUCTION in the number of nodes in the graph and to focus on one Linked Data [1] is Semantic Web data that emphasises the thing at a time, only one central node and its surroundings graph of relations between resources while recognising that are displayed concurrently. It is possible to switch between the data comes from Semantic Web documents that need to two views implemented based on the concept: a simple list be retrievable using standard conventions. Domain-specific view of objects associated with the focused subject (say, a applications can crawl Linked Data to pull together and dis- container), and the generic graph view from Fentwine [3]. play information from various sources. However, there is a need for generic Linked Data browsers as well, as they help A browsing session starts from some URI which is retrieved data producers to check what they publish and data con- for a document with RDF data. This URI will be the ini- sumers to check what is available. tial focus unless the document has a foaf:primaryTopic defined, in which case the primary topic URI will be chosen One generic Linked Data browser is Tabulator [2], but it as the initial focus. does not provide a graph view. Karger and schraefel pro- vide insights on why graph views often are not the best views For the surroundings, the generic graph view shows to the of Semantic Web data [4], but they acknowledge that graph left of the focus all triples that have the focus as an object, views have a high “cool factor” and a niche. We consider and to the right all triples that have the focus as a subject. a case that lies in this niche: we want to show the graph Each triple is shown as a predicate connecting the subject data as directly as possible. We apply a Free and Open to the object. This view is applied recursively to each dis- Source Software rich desktop application, Fenfire1 , and its played node until, with distance, the graph fades away to the graph view to Linked Data browsing because unlike other background. If the node has an rdfs:label, it is displayed browsers, this is a visually appealing, engaging and inter- instead of the URI. active demonstration of the Semantic Web’s capabilities. A graph view is a good way to explore a web of information, Graph navigation can be done entirely via keyboard by ro- and it is close to the nature of Linked Data as a heteroge- tating the surrounding nodes around the centre and moving neous, web-like environment with little high-level structure. focus to the node immediately to the left or right of the cen- tre. While there is limited space for the surrounding nodes, In the following, we demonstrate how the Fenfire applica- all nodes can be navigated to via the rotation. 1 http://fenfire.org/ To enable browsing of Linked Data, Fenfire dereferences the URI of the focused node and retrieves any rdfs:seeAlso related to this node whenever instructed to do so. For ex- ample, in Figure 1 a user has loaded the FOAF profile of one person, followed a foaf:knows link to another person and retrieved the FOAF profile of this person. As an important alleviation of incompletely linked data, Copyright is held by the author/owner(s). Fenfire adds triples asserting that the graph retrieved con- LDOW2008, April 22, 2008, Beijing, China. tains information about all of its disconnected components. Figure 1: Starting from Tim Berners-Lee, a user has followed FOAF data to a document about Henry Story. 3. USE CASES Some Linked Data browsers can edit the data as well. Fenfire We target two audiences primarily: can edit RDF graphs, but we need to implement remote • Semantic Web researchers, application developers and publishing interfaces in addition to the current feature of data producers need to explore available data on the saving to local files. level of individual triples. Fenfire provides a conve- nient alternative to manually downloading graph doc- Information display and navigation can be enhanced if there uments, reading the serialisation formats, and trying is knowledge about the user interaction requirements for a to match URIs to discover the links in the data. specific domain. Thus, Fenfire should have some automatic ontology-awareness and more should be configurable with • People who want to learn about or demonstrate the settings and plug-ins. Semantic Web and what data is available benefit from a visual presentation that truthfully shows the net- 6. ACKNOWLEDGEMENTS worked nature of the data. Here it is highly advan- The implementation of Fenfire in Haskell was initially devel- tageous that Linked Data documents that follow the oped by Benja Fallenstein and Tuukka Hastrup. guidelines include an rdfs:label for each node. 7. REFERENCES [1] T. Berners-Lee. Design Issues–Linked Data. Published 4. IMPLEMENTATION online, May 2007. http://www.w3.org/DesignIssues/ Fenfire has its origins in ZigZag [6], which is a system for LinkedData.html. managing interlinked, distributed data and is completely in- [2] T. Berners-Lee, Y. Chen, L. Chilton, D. Connolly, dependent of the World Wide Web standards. Fentwine [3] R. Dhanaraj, J. Hollenbach, A. Lerer, and D. Sheets. incorporated Semantic Web standards and became a graph- Tabulator: Exploring and Analyzing linked data on the view RDF browser, as did its sibling BuoyOING [5] (Buoy- Semantic Web. In Proceedings of the The 3rd Oriented Interface, Next Generation) user interface, which International Semantic Web User Interaction adds spatial navigation and concentrates on it. Workshop (SWUI06), Nov 2006. Fenfire is implemented in the Haskell programming language [3] B. Fallenstein. Fentwine: A navigational RDF browser to achieve high programmer productivity while meeting per- and editor. Proceedings of 1st Workshop on Friend of a formance requirements for real-time animation and for brows- Friend, Social Networking and the Semantic Web ing large graphs. It uses the Raptor library of the Redland (FOAF Galway), Sep 2004. project for RDF parsing, the GTK library for the standard [4] D. R. Karger and m.c. schraefel. The Pathetic Fallacy graphical user interface elements and the Cairo library for of RDF. In Proceedings of the The 3rd International the animated, vector-graphics based visualisations. Semantic Web User Interaction Workshop (SWUI06), Nov 2006. A major part of the architecture is the key-frame and identity- [5] J. Kujala and T. Lukka. Rendering recognizably unique based visualisation and animation system that provides the textures. In Proceedings of the 7h Internation user continuous feedback on how the navigation and the Conference on Information Visualization, 2003. IV switching between alternative views change what is visible. 2003., pages 396–405, Jul 2003. [6] T. H. Nelson. A Cosmology for a Different Computer Universe: Data Model, Mechanisms, Virtual Machine 5. CONCLUSIONS and Visualization Infrastructure”. Journal of Digital We applied Fenfire to the task of browsing Linked Data with Information, 5(1), Jul 2004. a graph view and highlighted the use cases of exploring, learning about and demonstrating Semantic Web data.