Published on December 6, 2007
Maximising Utility, Minimising CostSKOS and OWLOntologies and XML, XMLUK, 20005: Maximising Utility, Minimising Cost SKOS and OWL Ontologies and XML, XMLUK, 20005 Alistair Miles CCLRC Rutherford Appleton Laboratory Semantic Web Best Practices and Deployment SKOS Core: SKOS Core What is it? Simple Knowledge Organisation System(s) Main Purpose: Subject Metadata & IR E.g. ‘this document is about romantic love’. E.g. ‘this document is about the cure of tuberculosis by x-ray in India in the 1950s’. Based on ‘traditional’ wisdom … … thesauri, classification schemes, subject heading systems, taxonomies … … applied to distributed information … … SKOS Core is built on RDF. SKOS Core: SKOS Core What’s it good for? Bring the power of a thesaurus, classification scheme, or taxonomy, as an information retrieval tool to your distributed metadata. Low-cost migration of existing thesauri/classification schemes/taxonomies/… to RDF. SKOS Core: SKOS Core What’s the status? SKOS Core is maintained by W3C Semantic Web Best Practices and Deployment WG Recently published 2nd Public Working Drafts SKOS Core Guide SKOS Core Vocabulary Specification W3C Recommendation Track work item next year? N.B. SKOS is still work in progress, all feedback most welcome, email: [email protected] OWL: OWL OWL Web Ontology Language W3C Recommendation SKOS and OWL: SKOS and OWL Why two ‘standards’? What are the strengths & weaknesses? How do they fit together? How can I use them together? … How can I get the best ROI? Right Tool for the Job?: Right Tool for the Job? N.B. SKOS and OWL are tools What sort of tools are they? What jobs are they good for? … consider ‘ontologies’ vs. ‘thesauri’ or ‘classification schemes’ … (warning: humour coming) Ontologies: Ontologies [picture of amazing power tool with lasers] Thesauri &c.: Thesauri &c. [picture of stone tools used by cavemen] Perspective: Perspective Perspective … Ontologies are modern, and therefore better, whereas thesauri/taxonomies/classification schemes are older, and therefore legacy. N.B. I think this wrong. Motivating Scenarios: Motivating Scenarios “Better Metadata” “Smart Content” “Better Metadata”: “Better Metadata” “Better Meta…” Through more meaningful metadata, we can increase the effectiveness and efficiency with which we organise, curate, locate, deliver … information resources. “Better Metadata” e.g. …: “Better Metadata” e.g. … E.g. Danbri’s Google Earth Photo Demo … Google Earth: Google Earth Google Earth: Google Earth Google Earth: Google Earth Google Earth: Google Earth Google Earth: Google Earth Browser: Browser “Smart Content”: “Smart Content” “Smart Content” By making information (content) available in a machine-understandable form, that information can be re-used, re-purposed, re-combined, harvested, aggregated … “Smart Content” e.g. …: “Smart Content” e.g. … E.g. “FOAF” ontology and social networking … FOAF: FOAF FOAF: FOAF FOAF: FOAF FOAF: FOAF Which tool for the job?: Which tool for the job? OWL is a great tool for … Rich metadata Smart content SKOS is a great tool for ‘Not-quite-so-rich’ metadata Esp. (but not only) describing subjects of information resources E.g. ‘this document is about the cure of tuberculosis by x-ray in India in the 1950s’ SKOS e.g. …: SKOS e.g. … E.g. Semantic Web Environmental Directory (SWED) www.swed.org.uk SWED: SWED SWED: SWED SWED: SWED SWED: SWED SWED: SWED SWED: SWED SWED: SWED SWED: SWED Basic ontology of environmental organisations, properties (‘facets’), defined using OWL. Within each ‘facet’ controlled vocabulary defined using SKOS … e.g. ‘topic of interest’ hierarchy e.g. ‘project type’ hierarchy e.g. ‘operational area’ hierarchy Metadata Tradeoff: Metadata Tradeoff Cost ‘Semantically rich’ metadata is expensive to create, and expensive to maintain. Benefit A little bit of ‘meaning’ goes a long way. Metadata Tradeoff: Metadata Tradeoff W.r.t. use of metadata to support I.R. … Generally, as semantic richness increases, ROI diminishes. Although there are exceptions: minimum requirements, specialist applications … SKOS and OWL: SKOS and OWL Layer on SKOS: Layer on SKOS Build on top of SKOS … Layer on OWL: Layer on OWL Build directly on top of OWL … Hybrid: Hybrid Build on both, as appropriate … Exploring the tradeoff …: Exploring the tradeoff … SKOS and OWL … allow you to explore the cost/benefit tradeoffs involved in investing in semantics Summary: Summary Tools RDF+SKOS RDF+OWL Tasks Better Metadata (Information Retrieval) Smart Content Which tool? OWL for smart content, rich metadata SKOS for ‘not-quite-so-rich’ metadata, especially describing subjects of docs Maximising ROI? Establish the business need (better meta/smart content)? If better meta, how much meaning is really needed? SKOS …> SKOS+OWL …> OWL Thanks: Thanks Please comment, suggest, critique … [email protected] [email protected] N.B. Look for a ‘guide to extensions’ appendix to SKOS Core Guide in the new year.