Svante Schubert presented on metadata and the new metadata model for OpenDocument Format 1.2. The new model addresses limitations of the current ODF metadata by making it more extensible and descriptive. It uses RDF and OWL to annotate content in a common way, aligning with semantic web standards. Metadata is stored in RDF files and linked to content elements via IDs. This allows software to more easily find, combine and share information. OpenOffice.org 3 will provide APIs to access and extend the new metadata capabilities.

1. A RDF Metadata Model for OpenDocument Format 1.2 Svante Schubert Software Engineer Sun Microsystems Inc.

2. About the Speaker Since 1999 working for Sun Microsystems on StarOffice Since 2006 co-lead of OpenOffice XML project Responsible for the XML based filters Co-editor of the OASIS Metadata Specification

3. Agenda Metadata basics / existing standards Metadata model of ODF 1.2 Metadata support in OpenOffice.org 3

4. Metadata Basics “ The solution to the overabundance of information is more information.” David Weinberger Everything Is Miscellaneous: The Power of the New Digital Disorder

5. Metadata Basics What is metadata? “Metadata is data about data” 1) Why do I need metadata? Enhanced search Workflow Accessibility Citation Bridge the semantic gap (e.g. zip vs. post code, native languages) ... 1) http://en.wikipedia.org/wiki/Metadata

6. Metadata Basics More Precise Metadata Definition What is metadata? Labels to identify/categorize your data Related data Why do I need metadata? Metadata makes your data interpretable by other applications

7. Metadata Basics Extending current ODF Metadata Support Why a new metadata model? Current ODF metadata related to document Not extensible Content tagging by styles is not enough! Styles are not descriptive Styles are not interchangeable with other applications

8. Metadata Basics The Idea of the Semantic Web What is the Semantic Web / Data Web? A web, where software can find/combine/share information more easily Requirements of Semantic Web / Data Web Data annotated in a common way using metadata Web applications acting upon standardized metadata

9. Existing Metadata Standards Resource Description Framework (RDF) RDF/XML is W3C Recommendation (2004) Resources Unique identification by IRI Described by RDF statements RDF Statements - Triple: subject + predicate + object uri://sun/employees/svante foaf:name “Svante Schubert” .

10. Existing Metadata Standards RDF Graphs... Based on http://www.w3.org/2000/Talks/1206-xml2k-tbl/slide6-0.html

11. Existing Metadata Standards ...superimpose Based on http://www.w3.org/2000/Talks/1206-xml2k-tbl/slide7-0.html

12. Existing Metadata Standards Web Ontology Language (OWL) Ontology (from Greek) Onto (being) Logia (written/spoken discourse) Description of entities/concepts and their relations Like OO language using classes, properties, etc. ODF 1.2 includes OWL package description OWL Classes of pkg:Package, pkg:File, odf:Element

13. Metadata Model of ODF 1.2 Metadata Files in the Package RDF file e.g. “/meta/data.rdf” Content file e.g. “/styles.xml” Meta data manifest “ /manifest.rdf” RDF file e.g. “/meta/cita.rdf” Content file e.g. “/content.xml”

14. Metadata Model of ODF 1.2 Metadata Files in the Package Content files (e.g. content.xml, styles.xml) About 50 ODF elements with xml:id attribute Metadata manifest (manifest.rdf) Heart of metadata model Mapping from content's xml:id to RDF IRIs User RDF/XML files Metadata file possibly from an office extension

15. Metadata Model of ODF 1.2 Metadata Files in the Package <table:table xml:id=” someID ”> ... Hospital Doctor Duty List ... </table:table> RDF files e.g.“/meta/data.rdf” <odf:ContentFile pkg:path=&quot; content.xml &quot;> <pkg:hasPart> <odf:Element rdf:about=&quot; uri:someIRI &quot; pkg:idref=&quot; someID &quot;/> Content files e.g. “/content.xml” <odf:Element rdf:about=&quot; uri:someIRI &quot;> <ex:workingHoursOf> <med:Doctors rdf:about=&quot; http://hospital-DB/doctors/ID116 &quot;> <med:fieldName xml:lang=&quot;en&quot;>Neurologist</med:fieldN. Meta data manifest “ /manifest.rdf”

16. Metadata Model of ODF 1.2 In Content Metadata If metadata is equal text/visual data Reason: No data duplication Used by ODF 5 elements: Bookmark start - <text:bookmark-start> Heading - <text:h> Metadata text - <text:meta> Paragraph - <text:p> Table cell - <table:table-cell>

17. Metadata Model of ODF 1.2 In Content Metadata <text:p>The doctor's name was <text:meta m:about=” http://hospital-DB/doctors/ID116 ” m:property=” http://xmlns.com/foaf/0.1/name ”> Dr. J. Franklin</text:meta> RDF files e.g.“/meta/data.rdf” Content files e.g. “/content.xml” <med:Doctor rdf:about=&quot; http://hospital-DB/doctors/ID116 &quot;> <med:hasPatient> <med:Patient rdf:about=”http://hospital-DB/patients/IDA1”>

18. Metadata Model of ODF 1.2 Metadata Text Field ODF field “text:meta-field” based on metadata Appears within paragraph Holds any paragraph content Citation example: “According to [2]” <text:p>According to <text:meta-field xml:id=” someID ”> <text:style text:style-name=”s1”>[2]

19. Existing Metadata Standards Semantic Web Architecture Based on http://www.w3.org/2000/Talks/1206-xml2k-tbl/slide10-0.html

20. Existing Metadata Standards ODF in the Semantic Web Based on http://www.w3.org/2000/Talks/1206-xml2k-tbl/slide10-0.html

21. Metadata support in OOo 3 Support of metadata in the ODF 1.2 package API for metadata extension developers Providing UNO APIs to access metadata Wrapping existing opensource tools Some possible choices: Jena, Sesame, librdf, RDF Twig Possibility of a generic metadata extension Import / Create your own RDF vocabulary Relate vocabulary to ODF content by GUI

22. More Information Contact me here at OOoCon 2007 Download latest docs: http://www.oasis-open.org/committees/ documents.php?wg_abbrev=office-metadata RDF N3 Tutorial: http://www.w3.org/2000/10/swap/Primer RDF Converter: http://www.mindswap.org/2002/rdfconvert/ http://blogs.sun.com/GullFOSS

23. Thank you – Questions & Answers Svante Schubert [email_address]