OntoWiki a Semantic Data Wiki Enabling the Collaborative Creation and (Linked Data) Publication of RDF Knowledge Bases Sebastian Tramp Philipp Frischmuth Norman Heino {tramp|frischmuth|heino}@informatik.uni-leipzig.de Agile Knowledge Engineering and Semantic Web (AKSW) University of Leipzig / Institute of Computer Science Johannisgasse 26, 04103 Leipzig, Germany ABSTRACT text search. These different navigation facilities can also In our demonstration (with supportive poster) we want to be used as a combined navigation, for example a user present the Semantic Wiki application OntoWiki, which is can start with a full-text search and refine the results us- an extensible tool for managing structured information in a ing faceted-search or restrict them to a certain part of the collaborative, web-based environment. OntoWiki provides selected hierarchy. OntoWiki translates these different nav- sophisticated means for navigating, visualising and author- igation features into a single SPARQL query. In addition to ing of RDF-based Knowledge Bases. It serves and consumes rendering the results of such a query as a resource list, On- Linked Data and comprises a comprehensive middleware toWiki provides an extension mechanism for implement- API for building custom Semantic Web applications. ing dedicated views like maps and calendars. Once a spe- We refer to it as a Wiki, since our focus is on simplicity, cific entity has been selected, the resource view displays adaptability and collaboration. However, other than anno- all available information for that entity. Besides render- tating text-based Wiki pages with a special syntax (as sug- ing this information generically in a tabular way, OntoWiki gested by text-based Semantic Wiki approaches), OntoWiki can be extended with domain specific resource views uses RDF in the first place to represent information. For (i.e. vocabulary). Specific resource views for the SKOS and human users, OntoWiki allows to create different views on FOAF vocabularies are already included. Figure 1 shows the data, such as tabular representations or maps. For machine domain-independent list and resource view of OntoWiki. consumption it supports various RDF serialisations as well as RDFa, Linked Data and SPARQL interfaces. Since its Authoring introduction [1], the application has evolved into a frame- All views in OntoWiki can be equipped with corresponding work for building Semantic Web applications [2] and was RDFa annotations. OntoWiki employs the RDFauthor recently updated to support the collaboration across multi- mechanism [4] to automatically transform these views into ple domains and application via Semantic Pingback [3] and editable forms. As a consequence, all information items RDFauthor [5]. displayed in OntoWiki can be directly edited in place. This allows OntoWiki users to edit the knowledge base even with- out being acquainted with the RDF, RDF-Schema or OWL 1. DEMONSTRATION CONTENT data models. Following the Wiki approach all changes are OntoWiki can be used to author, manage and publish put under version control and can be easily rolled back. RDF-based knowledge bases. In particular, our demonstra- In addition to this, the integration of RDFauthor as our tion will focus on the following topics1 . We have highlighted main authoring mechanism led to two additional usage op- several phrases and keywords to give the reader a better tions: overview on the main demonstration points. • By using OntoWiki’s RDFauthor bookmarklet, users Navigation and Visualisation can collect data from different web pages and OntoWiki provides a number of ways for navigating through import it directly into their personal knowledge bases RDF knowledge bases. These include taxonomy and hi- (e.g. contacts or events). erarchy browsing (e.g. via SKOS taxonomies or the class hierarchy), facet-based browsing (with complex filter • OntoWiki can be used as a service for hosting ed- conditions and attribute based tag clouds) and full- itable mash-ups. These mash-ups use data from On- 1 toWiki’s SPARQL endpoint and other data sources The feature list is far away from being complete, so please (which are interpreted as named graphs) and provide refer http://ontowiki.net (for end-user centred function- a merged view on this data. The bookmarklet is able ality and screen-casts) and http://code.google.com/p/ ontowiki/ (for administration features and extensibility as to distinguish the statements from these named graphs well as development information) for a more complete view and allow the user to edit these data directly inside the on our application. mash-up as well as to propagate the changes back Figure 1: Generic OntoWiki views – resource list on the left (with two selected properties, an applied filter, map display and an attribute cloud) vs. resource view on the right (with similar and linking instances as well as tagging module). to the different sources (please refer to [4] for more 2. REFERENCES information). [1] Sören Auer, Sebastian Dietzold, and Thomas Riechert. Ontowiki - a tool for social, semantic collaboration. In Knowledge Base Evolution The Semantic Web - ISWC 2006, 5th International OntoWiki tries to support the creation of RDF Knowledge Semantic Web Conference, ISWC 2006, Athens, GA, Bases from the scratch (rather than using predefined on- USA, November 5-9, 2006, Proceedings, volume 4273 of tologies only) in a wiki way. This means that our users often LNCS, pages 736–749. Springer, 2006. follow the learning by doing path, where requirements and [2] Norman Heino, Sebastian Dietzold, Michael Martin, priorities change over time. Our evolution framework al- and Sören Auer. Developing semantic web applications low OntoWiki users to apply predefined evolution pattern with the ontowiki framework. In Networked Knowledge from a pattern repository to their data, as well as to cre- - Networked Media, volume 221 of Studies in ate their own pattern in order maintain the data quality Computational Intelligence, pages 61–77. Springer, of their Knowledge Base with little effort. Examples for such 2009. evolution patterns include the splitting of classes by us- [3] Sebastian Tramp, Philipp Frischmuth, Timofey ing their property values or the transition from a data Ermilov, and Sören Auer. Weaving a social data web property to an object property (e.g. the transition from with semantic pingback. In P. Cimiano and H.S. Pinto, literal based tags to resource based tags). editors, Proceedings of the EKAW 2010 - Knowledge Engineering and Knowledge Management by the Linked Data & Semantic Pingback Masses; 11th October-15th October 2010 - Lisbon, Portugal, volume 6317 of Lecture Notes in Artificial OntoWiki acts as a Linked Data server as well as an Intelligence (LNAI), pages 135–149, Berlin / Linked Data client. Entities that use identifiers being Heidelberg, October 2010. Springer. governed by the OntoWiki installation are automatically served as Linked Data. OntoWiki supports access con- [4] Sebastian Tramp, Norman Heino, Sören Auer, and trol on Linked Data employing the FOAF+SSL protocol Philipp Frischmuth. Making the semantic data web as well as traditional local account based authentication. easily writeable with rdfauthor. In Proceedings of 7th References to entities using foreign identifiers can be de- Extended Semantic Web Conference (ESWC 2010), 30 referenced (i.e. consumed by OntoWiki) in order to obtain May – 3 June 2010, Heraklion, Greece, volume 6089 of additional information from the original source. Even LNCS, pages 436––440. Springer, 2010. non-linked data resources can be consumed by utilising spe- [5] Sebastian Tramp, Norman Heino, Sören Auer, and cific data-wrapper, that allow the generation of RDF Philipp Frischmuth. Rdfauthor: Employing rdfa for data (e.g. from images, videos or web services). collaborative knowledge engineering. In P. Cimiano and Often it is beneficial for the owner of a Linked Data re- H.S. Pinto, editors, Proceedings of the EKAW 2010 - source to be informed about the usage of their resources. Knowledge Engineering and Knowledge Management by When publishing a FOAF profile, for example, it is use- the Masses; 11th October-15th October 2010 - Lisbon, ful to get notified when others establish a foaf:knows re- Portugal, volume 6317 of Lecture Notes in Artificial lation. We therefore developed the Semantic Pingback Intelligence (LNAI), pages 90–104, Berlin / Heidelberg, protocol. By employing Semantic Pingback, OntoWiki users October 2010. Springer. can communicate RDF statements across domains and ap- plications.