Stuart L' Weibel

1
Dublin Core Adventures in international
standardization

• Stuart L. Weibel
• Senior Research Scientist
• OCLC Research
• Scholar in Residence,
• University of Washington Libraries and
• the Information School

Bangalore 2009-03-26 2008-02-07
2
IN SOOTH

• The distance between theory and practice is
  always smaller in theory, than in practice
• Marshall T. Rose
• International standards are an exercise in
  closing this gap finding compromises across
  domains, sectors, cultures, languages, time
  zones, and egos all the while, adding value
• Best accomplished with people of good faith,
  imbued with passion and with the aid of a laptop
  in bars

3
Dublin Core The Beginning

• A casual discussion at WWW-2 in Chicago, October
  of 1994
• Joseph Hardin, Yuri Rubinski, Stuart Weibel,
  Terry Noreault, Eric Miller
• How can we help make things on the Web easier to
  find?
• OCLC NCSA co-sponsored an invitational workshop
  in March of 1995

4
Dublin Core Starting Assumptions and Essential
Features

• Simple
• true to a point the elements are simple, the
  underlying model is not
• Consensus-based
• Crucial to early success, both in attracting
  expertise and deployment.
• Bottom up
• Based on the experience of practitioners, but
  hard to capture and capitalize on lessons learned
• Cross-disciplinary
• Central success factor

5
Essential Features (continued)

• The Web is the strategic application
• On the mark
• International
• Also central success factor, but hard (25
  languages in the Registry)
• Lego-like modularity extensibility
• Partially realized promise
• Application Profiles are the means
• Syntax independence
• An ongoing nightmare
• Authors will describe their own works
• Laughably naïve

6
Modularity and extensibility the Lego metaphor

• DC is a beginning, not an end
• An architecture for modular, extensible metadata
• The simplest common denominator
• Add stuff you need for
• Local requirements
• Domain specific functionality
• Other dimensions of description
• Eg cloud cover management structural metadata.

7
Refinement

• In its simplest form, DC metadata is free-text in
  crude buckets
• Refinement is often desirable
• Two flavors
• Sub-properties (varieties of date, for example)
• Value spaces (controlled vocabularies or encoding
  schemes)

8
Refinements (continued)

• Creator
• Illustrator, author, composer, sculptor.
• Refine the meaning of the base element
• Must refine the meaning of the element, not
  broaden it.
• dumb down to the base element

9
Refinement (continued)

• Value spaces
• Controlled vocabularies and classification
  schemes
• LCSH
• DDC
• Many alternatives
• Implications for interoperability
• Encoding schemes
• W3C Date Time Format
• Subset of ISO 8306
• Disambiguate dates like 03-06-2006

10
Application Profiles

• Mix and Match metadata element sets
• snap together community-specific modules with a
  common architecture
• An AP should include
• A hybrid schema incorporating elements of some
  number of other schemas
• Further specification of cardinality rules,
  refinements, value spaces.
• Guidance on application (cataloging rules)

11
Syntax Semantics

• Form and meaning oil and water
• Thou shalt not discuss the two in the same room
  (ha ha)
• Try separating MARC and AACR2

12
Some Hard Questions

• How much metadata is enough?
• Your users probably need less than you think
• What is the right balance between human expertise
  and automated metadata creation?
• How can we coherently merge metadata from a
  diversity of sources created using many standards?

13
Inevitable Tensions in Metadata Deployment

• Simple and broad versus richly specific
• Cross-domain versus community-specific
• Generic (one size fits all) versus
  customized
• Readable by humans or processable by machines

14
The Process

• How we got it done

15
How we went about it

• Invitational workshops at the start
• Later, open workshops, and conferences
• Conscious development of diversity
  (cross-disciplinary and international)
• Openness (in participation and process)
• The DC World Tour
• Dublin, Warwick, Dublin, Canberra, Helsinki,
  Washington, Frankfurt, Ottawa, Tokyo, Florence,
  Seattle, Shanghai, Madrid, Colima, Mexico,
  Berlin, (Seoul)
• Many secondary events workshops, symposia,
  conference presentations on 5 continents

16
Means and Methods

• Face to face, Email, teleconferencing,
  publications
• Communities organized around topics or tasks
• A community Website (DublinCore.org)
• Process slow evolution, borrowed from others,
  leavened with (sometimes rancorous) experience
• (IETF, W3C, MARBI)
• Standardization (simple to more difficult)
• Standardization is often a politicized process
• IETF ? NISO ? ISO
• National derivative standards (Australia,
  Finland)

17
Dublin Core Metadata Initiative The maintenance
organization

• Original Host (OCLC) an accident of history
• Now incorporated organization hosted at Singapore
  National Library
• Directorate Executive management team
• Board of Trustees (9 seats) strategic direction
  and oversight
• Usage Board the editors and arbiters of the
  standard
• Affiliate Program Governance and sustenance
• Annual Conference (200 to 300 attendees)

18
Affiliate Program

• Affiliates are national entities the DCMI local
  franchise
• Staircase fees based on UN national GDP formula
• Distribute costs, and share governance
• Roughly half of the Board of Trustees seats
• Must be committed owners (not just funders) for
  this to succeed

19
Return on investment

• DCMI has framed the international metadata
  discussion for 15 years
• Many standards have been built on top of DC
• 7 countries formally, others informally
• The EU and the UN
• The US Government
• Agrivoc
• ANSI Standard for Description of Standards
• BBC
• PBS
• Reuters, Siemens, many others
• Knowledge management systems
• VRA, LOM, others

20
Connecting Value and Funding

• Network value is proportional to adoption
• Adoption is a function of value and ease of
  access and deployment
• Open participation and free accessibility were
  (and are) central features of Web deployment
• Those who harvest the value are generally not the
  ones who create it (an ongoing problem)
• But Pay to Play inhibits participation and uptake

21
Community

• The DC Makers the core group of people with
  the expertise, and opportunity to contribute
• Theorists and researchers juicy problems
• Practitioners rules and procedures
• DC Takers the people who just want to use the
  stuff. Just tell us how to do it.
• Communicating effectively to the community is a
  constant challenge
• Makers takers have different needs
• Innovators and maintainers have different skills
• Practitioners and theoreticians have different
  interests
• Languages and community disparities are barriers

22
The Big Picture

• DCMI has always been under-capitalized (it could
  scarcely be otherwise)
• Leadership, methodologies, and supporting
  processes were all midwifed amidst constant,
  hyper-change of the Web Decade
• Constantly inventing for the past what DCMI
  needs to do to add value in 2009 is different
  than 2003, and unrecognizable compared to 1998.
• And there be dragons politics, NIH syndrome,
  conflicting value propositions.

23
If I were King of the Forest(wed have)

• A formal data model (we sort of do finally Andy
  Powells Abstract Model)
• A coherent syntactic environment (pipe dream)
• Reference applications (DC in a box)
• Professional attention to documentation
  (especially application guides)
• Professional attention to organizational
  development (PR, Strategy, Community ownership)
• A successful approach to securing sustaining
  funds. Agencies love to support innovation, but
  rarely care to support standards maintenance

24
And the point is?

• DCMI is a dynamic, international community of
  theoreticians and practitioners who work
  collaboratively within a formal process on
  practical problems in the metadata domain.
• Does it matter if we stop, and how do we know?
• If we continue, will The Commons approach sustain
  us?

25
Many thanks!

• http//weibel-lines.typepad.com
• http//flickr.com/photos/weibel-lines
• http//twitter.com/stuartweibel
• (yeahFacebook, too)

26
(No Transcript)