ESWC 2007 - European Semantic Web Conference

This are just the most important facts (for us) from the Conference in Innsbruck.

People encountered

  • Sebastian Kruk, DERI Galway, JeromeDL project
  • Bernd Haslhofer, Vienna University, BRICKS project
  • Eyal Oren, DERI Galway, ActiveRDF author

Digital library tutorial

This was about "Semantic digital libraries", and presented three existing systems. All of those

  • Used an RDF triple store for semantic information
  • Tried to import existing data (e.g. Marc21, see MarcOnt?)
  • Had a controlled vocabulary for metadata for librarians
  • Allowed for personalization and/or user content with an "uncontrolled" vocabulary
  • Were deployed as Java applications on Tomcat
  • There are comparisons tables in the slides

JeromeDL

  • Created at DERI Galway (and others)
  • 10-12 instances installed, not all institutional
  • Aims at: Librarians, Scientists and normal users
  • Metadata: DublinCore?, BibTex?, Marc21, represented in RDF
  • Uses predefined ontologies (e.g. MarcOnt?)
  • WWW user interface (aims to be simple)
  • P2P for searching and authentication (FOAF)
  • FOAF based community framework, for authentication, advanced bookmarking etc.
  • Hierachical mode for browsing
  • ELP (their own) protocol for communication (Hypercube)
  • Identifier: URLs

BRICKS

  • Distributed software framework
  • Prototype, going to stable version
  • About 8 installations, small research institutions and museums
  • P2P nodes form network (BNodes)
  • Interfacing through web services
  • Digital object store
  • Web interface on nodes (not part of the "standard" package)
  • Search and browse
  • Uses Physical/Logical "Collections"
  • Data can be stored within Bricks or be referred to via URL
  • Uses JCR APi
  • Plug-In ontologies (OWL)
  • Uses sparql
  • Architecture and security: See slides
  • Demo Interfaces: Archeological fact finder, Annotation Tool

FEDORA

  • Complex, mature system
  • Advanced functionality, workflow, versioning, replication, Integrety checking, ...
  • User interface is not part of the package
  • Uses persistent IDs
  • About 50 "monitored" installations
  • Each object consts of several datastreams
  • One contains RDF data
  • RDF data is combined in "external" store
  • Has own "base ontology" and "community ontology"
  • They ested various triple stores
    • Jena, Sasame - soso
    • Kowari - Fast but unstable?
    • MPTStore - Fast
    • Mulgara -> Kowari++ ?
  • Several high-profile projects
  • Eg. 4.7 million objects, 250 million triples -> "moderately large" MPTStore
  • Projects included e.g. Wiki/Blog integration to existing software (!)

Similile

They basically showed a tool collection

Community aspects

There was much emphasis on community aspects and user generated content. The user generated content was in all cases kept "separate" from the "proper" metadata, since the librarians want to keep control over the cataloguing

Scripting workshop

  • Metric for "sematic performance", didn't seem too interesting
  • Debugger for rules, cool but not for Talia
  • Book mashup: Collecting data from Google and Amazon, books are identified in web data through isbn, crawlers collect data
  • Rapid prototyping toolkit
  • Integrated SIOC browser (online communities)
    • Presented by Eyal Oren
    • http://activerdf.org/sioc
    • Specialized interface for community content
    • Faceted browsing
    • Part of the scripting challenge
  • Lookup index
    • Eyal Oren, Giovanni
    • web crawler/search engine
    • a query is just for one resource
    • a query returns thelocations where a is mentioned, and nothing more
    • results are ranked
    • http://sindice.com/
    • Part of the scripting challenge
  • SPARQL enhancement with Javascript
  • Javascript framework
  • Semantic functional language
    • Part of the scripting challenge
  • Semantic Radar (with Firefox extension)
    • reports RDF data locations to it's server (kind of a meta-crawler)
    • Basically uses users to crawl the web
    • The user-function of the plugin is to alert the user to semantic content in web pages
  • DOAP browser (?)
    • Part of the scripting challenge
  • SPARQL to LDAP bridge
    • also a poster presentation
    • Part of the scripting contest (winner?)