• OntoMedia (OM) Ontology – Developed at ECS, University of Southampton by • • • • K Faith Lawrence Mike Jewell Mischa Tuffield [interest: description of online fiction] [interest: automated soundtrack generation] [interest: description of life narratives] OM Stories Ontology – Development lead by Mike Jewell • Intended to fully integrate with FOAF, Event and Timeline ontologies See: http://contextus.net/ Existing top-level ontologies did not deal well with: Non-real world assumptions Multiple, cross-linked conceptual worlds Inexact/Referential details Multiple interpretations < !‐‐ Events‐‐> <!‐‐Sleeping Beauty‐‐> <rdf:Descrip;on rdf:about="#SB_Kiss"> <rdfs:label>Sleeping Beauty.Kiss</rdfs:label> <rdf:type> <owl:Class rdf:about="&omea;Sex"/> </rdf:type> <ome:exists‐in rdf:resource="#SleepingBeauty"/> <ome:causes rdf:resource="#SB_Wake_Up"/> <ome:has‐object‐en;ty rdf:resource="#SB_Princess_lips"/> <ome:has‐subject‐en;ty rdf:resource="#SB_Prince_lips"/> </rdf:Descrip;on> <rdf:Descrip;onrdf:about="#SB_Wake_Up"> <rdfs:label>SleepingBeauty.PrincessAwakes</rdfs:label> <rdf:type> <owl:Classrdf:about="&ome;Transforma;on"/> </rdf:type> <ome:exists‐in rdf:resource="#SleepingBeauty"/> <ome:caused_by rdf:resource="#SB_Kiss"/> <ome:has‐subject‐en;ty rdf:resource="#SB_Princess"/> <ome:from rdf:resource="#Asleep"/> <ome:to rdf:resource="#Conscious"/> </rdf:Descrip;on> Sun, Moon and Talia Sleeping Beauty King sees Talia asleep/dead Prince sees sleeping woman King is overcome by beauty (Prince is overcome by beauty) King has sex with Talia King leaves Talia gives birth to twins Baby sucks splinter from finger Talia wakes up Prince kisses sleeping woman Sleeping woman wakes up 'Lord of the Ring: The Two Towers' "So that is the answer to all the riddles! The One Ring that was thought to have perished from the world. And Boromir tried to take it by force? And you escaped? And ran all the way – to me! And here in the wild I have you: two halflings, and a host of men at my call, and the Ring of Rings. A pretty stroke of fortune! A chance for Faramir, Captain of Gondor, to show his quality!" - Faramir Paul Rissen at the BBC Future Media & Technology BBC has long history of creative output Audio Video Text … and now the Web People Places Events Factors Michael O. Jewell, Goldsmith's University of London Annotation of media an issue Focus on screenplays Create automated metadata Sacrifice accuracy for speed ? Create base store of information which can be improved Performance Text Module Cast Lists Roles Speakers Stage Directions Established Existing tools and knowledge Hundreds of resources Dbpedia (from Wikipedia) Geographical information Songs and bands Species People Over 1000,000,000 triples available Why? Electronic texts: People, places, events Linked data: People, places, events We can augment electronic text with the wealth of knowledge on the Semantic Web How? RDFa (RDF in Attributes) A means to add machine-readable hints to XHTML via attributes in elements For example: <stage type="location"> <p property="dc:title" about="[ds:Control_Room]" typeof="[loc:Room]"> INTERIOR – CONTROL ROOM </p> </stage> Dark Star Dark Star dataset as TEI, N3, and XML (http://contextus.net/datastores) SPARQL endpoint (http://contextus.net/sparql/) Shakespeare Midsummer Night's Dream (other plays coming soon) TEI from Perseus Digital Library Has to have at least two (named) female characters Who talk to each other About something besides a man… 'The Rule' In Dykes to Watch Out For (Alison Bechdel, 1985) Crowd sourcing the corrections - many hands make light work Potentially vast dataset Find repeated structures within the data Find consensus interpretation Opportunity to study the audience response Find points of agreement/contention Can trends in interpretation by links to trends in the audience composition? For more information: Website: http://contextus.net/ Mailing List: http://groups.google.com/group/digital-narrative IRC: http://irc.freenode.net#ontomedia Questions?