Building Metadata Infrastructure

1

• Building Metadata Infrastructure
• National Library of New ZealandNovember 26th
  2003
• Jon Mason jmason_at_educationau.edu.au

2
Presentation structure

• Who
• When
• What
• Why
• Where
• How

players stakeholders historical
context dimensions of metadata the bigger
picture case-studies the tricky details!
3
(No Transcript)
4
When Historical Context

• Growth of the information economy
• Organizing information on the Web
• directories, search engines, portals
• from META tags to metadata schema
• the XML revolution (dont forget HTML!)
• RDF the Semantic Web (next generation)
• Intersecting domains
• information services
• e-learning lifelong learning
• performance support
• human resource development

5
What Metadata Infrastructure

• Two characteristics
• Conceptual semantic
• Technical
• Information superhighway
• Metadata like the street signs
• Metadata standards like the road rules
• Accommodates content process
• Two main uses
• Machine-to-machine interoperability
• Human consumption

6
What Dimensions of Metadata

• Surface (e.g, web page last updated)
• Embedded
• as HTML META tags
• as suffixes (.doc, .html, .pdf, .ppt, jpeg, )
• as custom XML tags (metadata-rich markup)
• Detached
• in repositories of catalogued information
• RDF statements (relationship-rich)
• RSS channels
• Formally defined standards protocols
• Dublin Core (ANSI Z39.85)
• IEEE LOM (IEEE 1484.1-2002)
• METS (Metadata Encoding Transmission Standard)

7
What Dimensions of Metadata
an innovation in metadata handling page
ranking via citations (backlinks), anchors,
proximity algorithms
8
What Dimensions (METS)

• Descriptive (aboutness)
• Structural
• linking or organizing content chunks into
  aggregates(printed books are relatively
  self-organizing e-content can have infinitely
  assigned structural maps)
• Administrative
• e.g, managing catalogues
• metadata IP meta-metadata
• File Groups
• associating content chunks
• Behaviour
• rule-based
• METS - http//www.loc.gov/standards/mets/

9
Why

• Organizing managing information
• for subsequent discovery retrieval
• to enable data interchange
• separating content-structure-presentation
• Tool for the digital economy
• networked environment (diverse complex)
• economic value-chains based on value creation
  rather than value extraction
• Organizing knowledge

10
How Tasks

• Information modeling
• classification taxonomy development
• terminology vocabulary development
• data-typing
• referencing authoritative thesauri
• ontology development
• topic maps
• Description Abstract construction
• Keyword craft
• controlled vocabularies
• natural language
• Index creation
• Standard schema encodings

11
How Tool Development

• Many kinds of tools
• item level metadata assignment
• batch handling (assignment harvesting)
• retrospective assignment (Autonomy)
• topic map generation
• facilitate schema interoperability (SchemaLogic)
• taxonomy generators
• metadata extraction re-processing
• Content Management Systems
• RSS
• Principles
• automate where possible
• minimize human intervention

12
http//mapageweb.umontreal.ca/turner/meta/english/
13
Application Profiles

• Why
• - Local requirements are sometimes not fully
  accommodated by standard schemas
• How
• Extending or Abbreviating existing standards
• Mixing Matching standards namespaces
• However! Current practice suggests
  interoperability an issue
• Where
• SCHEMAS www.schemas-forum.org/registry/desire/app
  profile.php3
• CORESwww.cores-eu.net/registry/schema/TLFwww.th
  elearningfederation.edu.au/repo/cms2/tlf/published
  /8519/Metadata_Application_Profile_1_3.pdf

14
Where Case-Studies

• EdNA Onlinehttp//www.edna.edu.au/search
• MERLOThttp//www.merlot.org/
• The Learning Federationhttp//www.thelearningfede
  ration.edu.au
• RDNhttp//www.rdn.ac.uk/
• EOE Foundation http//www.eoe.org/

15
Conceptual Challenges

• What is content?
• new data-types object-types constantly
  appearing
• Boundaries
• Where are the boundaries between content,
  context, learning activities?
• What are digital objects what are digital
  collections?
• Where are the boundaries between data,
  information, knowledge?
• When is a discovery service a directory service
  ( vice versa)?
• Modeling knowledge
• extending the scope of metadata assignment use

16
Source Chris Blackall, ANU, 2003
17
Ongoing Questions

• How best to encode context?
• separate from content (eg., EML, EAD, EAC)
• embedded within descriptive metadata?
• include in Audience descriptions?
• include in Activity descriptions?
• Knowledge Management considerations
• Is completeness of description possible?
• How to achieve interoperability?
• DC, ERC, LOM HTML, XML, RDF,
• Semantic Web (semantics syntax)
• balancing generalist specific classification

18

• Questions?