1
Convergence is the GoalActivity Reportof the
IFLA FRBR/CIDOC CRM Harmonization Group

• Patrick Le Bœuf
• FRBR in 21st century catalogues
• an invitational workshop
• Dublin, Ohio, May 2-4, 2005

2
FRBR/CRM Harmonization Group

• formed 2003
• gathers representatives for corresponding
  members of
• the IFLA FRBR Review Group
• the CRM Special Interest Group (CRM-SIG)
• chaired by Martin Doerr, Institute of Computer
  Science of the FOundation for Research
  Technology Hellas ICS-FORTH (assisted by
  Patrick Le Bœuf)

3
Persons involved

• For the FRBR community
• Trond Aalberg
• Ketil Albertsen
• Allyson Carlyle
• Beth Dulabahn
• Tom Delsey
• Stefan Gradmann
• Mauro Guerrini
• Patrick Le Bœuf
• Dan Matei
• Glenn Patton
• Gerhard Riesthuis
• Richard Smiraglia
• Barbara Tillett
• Maja Žumer

• For the CRM community
• Gerhard Budin
• Nicholas Crofts
• Martin Doerr (chair)
• Tony Gill
• Dolores Iorizzo
• Stephen Stead
• Matthew Stiff
• Manfred Thaller
• Günter Waibel

4
Why harmonize FRBR CRM?

• They represent similar efforts in close fields
  (cultural heritage)
• It is in line with the mainstream of ALM
  convergence
• It will facilitate mediation systems between
  library catalogs museum inventories
• It will facilitate reuse of tools designed for
  CRM (e.g., RDF Schema expression of CRM)
• It will facilitate cross-domain projects (e.g.,
  the SCULPTEUR/PICTEUR Project)

5
Comparative chart

• FRBR
• E-R
• focuses on multiple objects (ideally
  identical copies of publications)
• scalable and incomplete FRBR for descriptive
  aspects, FRAR for authorities,
• static, non-event-aware model
• few entities, many attributes

• CRM
• OO
• focuses on unique objects (that can be grouped
  by type, e.g. specimen/species)
• integrated seamless coverage of descriptive
  aspects authorities
• dynamic, event-aware model
• no attribute as such, only relationships (?
  many classes)

6
What is CIDOC CRM?

• Developed from 1996 on by ICOM CIDOC
  (International Council of Museums International
  Committee for Documentation)
• Maintained by CRM-SIG
• About to be validated as ISO 21127
• Builds upon the CIDOC Information Categories
• Covers fine arts, archaeology, natural history

7
Role of CRM Dig Meaning Out of Flat Statements
Author
has an
INV4884 Artist Théodore Géricault Date 1818-1819
Something
and a
Date
actually means
8
ActorAppellation
known as
Théodore Géricault
was carried out by
Actor
was produced by
ProductionEvent
Something
lasted for a given
Time-Span
known as
can be approximated with reference to
Time Primitive(date range)
ObjectIdentifier
1818-1819
INV4884
9
Event-Centered Structure of CRM
Involving whom?
Involving what?
ActorAppellation
Appellation
Actor
PhysicalStuff
What happened?
ConceptualObject
Event
Time-Span
Place
TimePrimitive
PlaceAppellation
When?
Where?
10
Methodology (1)

• 3 meetings so far
• Meeting 1 2003, Nov. 12-14
• Getting to know each other talks and debates
• Collection vs. Sets vs. Multipart Objects
• Richard Smiraglia on Work notion
• Allyson Carlyle on Expression notion
• Manfred Thaller on Manuscripts
• Tom Delsey on Subjects
• Meeting 2 2004, March 22-25
• Work Expression Attributes
• Meeting 3 2005, Feb. 14-16
• Manifestation Item Attributes
• Detailed reports have not been made publicly
  available

11
Methodology (2)

• Examine each Attribute
• What does it mean?
• Is there any implicit assumption about its
  meaning?
• How do non-librarians understand its definition?
• How to express the same meaning in a CRM-like
  structure?
• Whats on a librarians mind?
• Cataloging processes sometimes important to model
  too

12
Methodology (3)

• Too many Attributes? ? Split the entity!
• A given Attribute actually refers to an Event? ?
  Make the Event explicit!
• How do catalogers acquire knowledge about merely
  abstract entities? ? Through concrete entities
  that are deemed to be representative for abstract
  entities

13
Some Principles

• The idea of the Work is only known through a
  representative Expression
• i.e., my idea of Hamlet it is an English text
  a Japanese version an Expression too, but not
  representative of what the Work is
• An Expression is only grasped through a
  representative Manifestation
• i.e., an edition of Hamlet titled Something
  rotten in Denmark but otherwise with correct
  text a Manifestation, but not deemed
  representative as to its title
• It is the Bibliographic Agency ( the cataloger)
  who determines what is representative and what
  is not

14
The old debate again!New Work, or just another
Expression?
has constraining super-type
has representativeexpression
Work
Expression
has type
Type
BibliographicAgency
RepresentativeAssignment
performed

• A cataloger says I think of Hamlet as a Work
  that is best represented by an Expression of
  Text type
• If something called Hamlet is of another type
  than that constraining super-type (e.g., a
  movie), then it is another Work

15
Other Constraining Super-Type
Same Constraining Super-Type
16
Three basic distinctions that were absent from
E-R FRBR
Work
is a
is a
has distinct parts, either in its conception
(e.g., a trilogy), or over time (revisions,
translations)
ComplexWork
Self-ContainedExpression
is expressed in
is a
SerialWork
17
Expression
is fragment of
is a
is part of
is a
FragmentExpression
Self-ContainedExpression
shows something conceived as a whole by its
creator
not intended as a whole by its creator
18
InformationCarrier
Type
2 classes from CRM
is a
is a
is a
ManifestationSingleton
ManifestationProduct Type
Items
e.g., a publication(abstract notion)
e.g., a manuscript (physical object)
19
The nature of a title-page
Expression
ManifestationProduct Type
comprises carriers of
has publisher content
(instance authors text)
PublicationExpression
is composed of
Title
(instance title-pageTOC publishers logo)
(proper)

• The info found on title-pages does not belong to
  the embodied Expression
• It is a peculiar kind of Expression created by
  the publisher we called it Publication
  Expression

20
What next?

• Group 2, Group 3, FRAR attributes
• FRBR FRAR relationships
• Polish the overall picture (some attributes were
  postponed, some new concepts need clarification)
• Check the robustness
• Draft deliverables scope notes and examples for
  each class property, tutorials, explanatory
  documents
• 2 years of work??

21
Pros Cons

• Pros
• Goal only 1 conceptual model for museums
  libraries ? unified field
• OO formalism more appropriate for Semantic Web
  activities
• Opens ways to revolutionary OPACs

• Cons
• Sounds too complicated for catalogers?
• Appeals more to information computer scientists
  than to librarians?
• Do we need a unified field at all?

22
By way of conclusion

• 2007 Annus Mirabilis
• International Cataloguing Principles
• AACR3
• and hopefully OO-FRBR??

23
Thanks for your patience!
Special thanks to Martin Doerr and Anila Angjeli
for re-reading this presentation and help me
correct it