FRBR in RDF
Ian Davis (ian.davis@talis.com)
Technical Lead, Research Group
Talis Information Ltd.

What is the FRBR?
- For us non-librarians...
- "Functional Requirements for Bibliographic Records"
- Published by the IFLA
- Describes an entity-relation model for bibliographic information
- Provides a common model to enable interoperability
- http://www.ifla.org/VII/s13/frbr/frbr.htm
[any material that should appear in print but not on the slide]
How Did I Get Involved?
- Joined Research Group at Talis in June to work on Semantic Web
- Thought I'd better learn some of this library stuff
- Found a rich, well-thought out model - FRBR
- First thought - "this feels like a good fit with RDF"
- No published schema but several people thinking about it.
- Richard Newman (http://www.holygoat.co.uk/) had most complete version
[any material that should appear in print but not on the slide]
Richard's Original Schema
- Started in July 2004
- Based on "casting down" abstract notions from FRBR into specific domains like musical recordings
- FRBR has several duplicated relationships between entities
- e.g. reproduction can apply between manifestation and manifestations, manifestation and item, item and item
- Ended up with stupidly long names like
isAManifestationReproductionOfManifestation
- Work stalled
[any material that should appear in print but not on the slide]
Richard Newman (http://www.holygoat.co.uk/) had most complete version
Cleaning The Schema
- Richard had focussed on detailed modelling of constraints
- Properties like
isAReferentialSuccessorToExpression
- Very precise - lets domains and ranges infer concrete types
- Together we felt the need for a simpler version
- Split into two parts:
- Given a stable, persitant home at http://vocab.org
[any material that should appear in print but not on the slide]
FRBR Core Schema
- http://purl.org/vocab/frbr/core
- Models the core entities (Work, Expression, Place, Person etc.)
- General entity relationships (realization, complement, part etc.)
- Focus on simple expression of FRBR
- No entity attributes yet... but they belong in this schema
[any material that should appear in print but not on the slide]
FRBR Extended Schema
- http://purl.org/vocab/frbr/extended
- Models refinements of entities (ReferentialWork, AutonomousExpression etc)
- Specific typed entity relationships (complementsWork, hasPartWork etc.)
- Focus on typed expression of FRBR
- Still under heavy development
[any material that should appear in print but not on the slide]
Group 1 Entities
- Endeavour - Any of the products of artistic or creative endeavour.
- Not a FRBR entity but a superclass for the following group 1 entities:
- Work - An abstract notion of an artistic or intellectual creation.
- Expression - A realization of a single work usually in a physical form.
- Manifestation - The physical embodiment of one or more expressions.
- Item An exemplar of a single manifestation.
[any material that should appear in print but not on the slide]
Group 1 Entities
[any material that should appear in print but not on the slide]
Group 2 Entities
- ResponsibleEntity - One responsible for the creation, production, distribution or maintenance of artistic or creative content.
- Again not a FRBR entity but a superclass for the following group 2 entities:
- CorporateBody - An organization or group of individuals and/or other organizations.
- Person - A living or dead individual.
[any material that should appear in print but not on the slide]
Group 3 Entities
- Subject - Something that may serve as the subject of a work.
- Again not a FRBR entity but a superclass for the following group 3 entities:
- Concept - An abstract idea or notion
- Object - A tangible or material thing.
- Event - An action or occurrence.
- Place - A location.
[any material that should appear in print but not on the slide]
Properties
- Naming follows convention
- XXX has property with value YYY
- e.g. foo has realization with value bar
- Most properties are paired with an inverse
- realization/realizationOf - relate works and expressions
- embodiment/embodimentOf - relate expressions and manifestations
- exemplar/exemplarOf - relate manifestations and items
- Synthetic superclasses used for some property ranges and domains:
- creator/creatorOf - relate works and ResonsibleEntities
[any material that should appear in print but not on the slide]
What's Missing?
- Entity attributes (title, date, context, medium etc)
- Looking to Dublin Core, PRISM, MODS etc.
- Forms - what form a work has
- e.g. Biography, ArtisticWork, ScholarlyWork, LiteraryWork, LegalWork, News
[any material that should appear in print but not on the slide]
Abbreviated Example
@prefix frbr: <http://purl.org/vocab/frbr/core#> .
@prefix dc: <http://purl.org/dc/elements/1.1/> .
<http://www.imdb.com/title/tt0245712/>
a frbr:Work;
dc:title "Amores Perros"@es ;
dc:title "Love's a Bitch"@en ;
frbr:subject <#Mexico_City> ;
frbr:realization <#amores-perros> ;
<#amores-perros>
a rbib:MotionPicture ;
dc:creator "Alejandro González Iñárritu" ;
frbr:embodiment <#amores-perros-dvd> .
<#amores-perros-dvd>
a rbib:DVD ;
dc:date "2000" .
[any material that should appear in print but not on the slide]
Schema Design
- Lucky - started with a good model!
- Use of OWL-DL to provide constraints
- e.g. Works cannot be Expressions
- Grounding with respect to other schemas
- SKOS to document schema
[any material that should appear in print but not on the slide]
Documentation
- Generated by XSLT! - no RDF templating language
- Requires RDF/XML to follow regular patterns
- Permission from IFLA to reuse content from report - but were commercial constraints
- Experimenting with paraphrasing OWL constraints
- A class:
A frbr:Manifestation is something that, amongst other things, is a frbr:Endeavour but is not a frbr:Work and is not a frbr:Expression and is not a frbr:Item
- A property:
Having a frbr:complement implies being something that, amongst other things, is a frbr:Work or a frbr:Expression
[any material that should appear in print but not on the slide]
Paraphrase Generation
- <rdfs:subClassOf rdf:resource="http://purl.org/vocab/frbr/core#Endeavour"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Work"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Expression"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Item"/>
A frbr:Manifestation is something that...
[any material that should appear in print but not on the slide]
Paraphrase Generation
- <rdfs:subClassOf rdf:resource="http://purl.org/vocab/frbr/core#Endeavour"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Work"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Expression"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Item"/>
A frbr:Manifestation is something that, amongst other things, ...
- Remember the open world assumption!
[any material that should appear in print but not on the slide]
Paraphrase Generation
- <rdfs:subClassOf rdf:resource="http://purl.org/vocab/frbr/core#Endeavour"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Work"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Expression"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Item"/>
A frbr:Manifestation is something that, amongst other things, is a frbr:Endeavour
[any material that should appear in print but not on the slide]
Paraphrase Generation
- <rdfs:subClassOf rdf:resource="http://purl.org/vocab/frbr/core#Endeavour"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Work"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Expression"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Item"/>
A frbr:Manifestation is something that, amongst other things, is a frbr:Endeavour but is not a frbr:Work
[any material that should appear in print but not on the slide]
Paraphrase Generation
- <rdfs:subClassOf rdf:resource="http://purl.org/vocab/frbr/core#Endeavour"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Work"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Expression"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Item"/>
A frbr:Manifestation is something that, amongst other things, is a frbr:Endeavour but is not a frbr:Work and is not a frbr:Expression
[any material that should appear in print but not on the slide]
Paraphrase Generation
- <rdfs:subClassOf rdf:resource="http://purl.org/vocab/frbr/core#Endeavour"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Work"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Expression"/>
- <owl:disjointWith rdf:resource="http://purl.org/vocab/frbr/core#Item"/>
A frbr:Manifestation is something that, amongst other things, is a frbr:Endeavour but is not a frbr:Work and is not a frbr:Expression and is not a frbr:Item
[any material that should appear in print but not on the slide]
Grounding Schemas
- Learnt that it is very important to relate ones work to others
- Puts work in context
- Makes the schema part of the "web" of schemas
- RDF enables this naturally...
subClassOf/subPropertyOf - is-a relationships
equivalentClass/equivalentProperty - stronger equivalence
[any material that should appear in print but not on the slide]
Specific Groundings
- FOAF
- frbr:Person equivalentClass to foaf:Person
- CorporateBody equivalentClass to union of foaf:Organization and foaf:Group
- Geo
- frbr:Object subClassOf geo:SpatialThing
- frbr:Place subClassOf geo:SpatialThing
- OWL Time
- frbr:Event subClassOf owltime:Event
- SKOS
- frbr:Concept subClassOf skos:Concept
[any material that should appear in print but not on the slide]
Relevancy Beyond Bibliographic Domain
- Model appears to fit with other domains
- Web Architecture
- Notion of resource identified by a URI, corresponds to FRBR Work
- Resources have representations, corresponding to Expressions
- Web browser negotiates encodings etc, corresponds to Manifestation
- Actual bits sent correspond to Item
- Known FRBR issue with electronic Items...
[any material that should appear in print but not on the slide]
Dublin Core
- New activity to assign domains and ranges to core terms
- FRBR fits well in many cases:
| Term | Domain | Range |
| dc:title - | Endeavour | - |
| dc:creator - | Endeavour | ResponsibleEntity |
| dc:subject - | Endeavour | - |
| dc:description - | Endeavour | - |
| dc:publisher - | Endeavour | ResponsibleEntity |
| dc:contributor - | Endeavour | ResponsibleEntity |
| dc:date - | Endeavour | - |
| dc:type | - | - |
[any material that should appear in print but not on the slide]
Dublin Core 2
| Term | Domain | Range |
| dc:format - | Manifestation | - |
| dc:identifier | - | - |
| dc:source - | Endeavour | Endeavour |
| dc:language - | Endeavour | - |
| dc:relation - | Endeavour | Endeavour |
| dc:coverage - | Endeavour | - |
| dc:rights - | Endeavour | - |
[any material that should appear in print but not on the slide]
Issues
- Most issues are with abstract nature of FRBR
- What constitutes a different Work? Different creators? What about collaborative works?
- Is a periodical a Work?
- Quite far removed from citation requirements, e.g. chapters, articles etc
- Issue whether Items can be digital.
- FRBR report suggests they must be physical
- Complexity of modelling typed relationships in RDF + OWL
- Remember
isAManifestationReproductionOfManifestation
- Needed so domain and range can be set explicitly to Manifestation rather than more general Endeavour
- Alternate is multiple namespaces: frbrman:reproduction, frbritem:reproduction!
[any material that should appear in print but not on the slide]
Thanks!
[any material that should appear in print but not on the slide]