Upload
leslie-hopkins
View
213
Download
1
Embed Size (px)
Citation preview
Metadata Vocabularies in Subject Gateways for Japanese Regional
Public Libraries
Shigeo SugimotoResearch Center for Knowledge CommunitiesGraduate School of Library, Information and
Media StudiesUniversity of Tsukuba
Some Japanese Governmental Activities, Metadata Vocabularies, and
a Model for Sharing Metadata Schemas
Shigeo SugimotoResearch Center for Knowledge CommunitiesGraduate School of Library, Information and
Media StudiesUniversity of Tsukuba
DC in Japanese Industrial Standard• Simple Dublin Core has been accepted as a
National Standard in March 2005 by JIS. JIS X 0836
• Translation of ISO 15836• Key issue for translation
– Choice of Japanese terms for Labels, which should be domain neutral
– Example: Subject element• 主題、件名
– Straightforward translation of “subject” – Library oriented terms, possibility of confusion with the meaning
of Title
• キーワード( keyword)– Domain Neural– chosen for JIS X 0836
Some Activities in Japan• Preservation
– Web Archiving: National Diet Library (NDL)– Governmental Resources: NDL and National
Archives of Japan• NDL
– National Deposit Library– Appointed as the central organization for national
Web archiving• Revision of Law for legal deposit
• National Archives of Japan– Archiving of government resources
• Born digital resources• Definition of “Governmental Resource”
– Central repository for government documents
Digital Okayama Dai-Hyakka (DODH)
• DODH– Regional portal by the Okayama Prefectural Library– Metadata creation by librarians and non-professional
s, e.g. school teachers, students, and volunteers.
• A Key Issue: Subject Classification– Choice of subject vocabularies– Small set of subject terms usable for the non-professi
onals designed in accordance with regional needs.
Basic Ideas in DODH• Basic Question:
Is comprehensive/conventional subject vocabulary really useful?
• Use domain specific vocabularies in addition to a comprehensive and widely used by libraries– Three vocabularies
• NDC (Nippon Decimal Classification)– Very widely used by Japanese libraries
• Kid’s Vocabulary– Designed for children and child resources
– Multiple labels in accordance with user ages
• Prefectural Resource Vocabulary – Developed by the government of Okayama prefecture for their r
esources
Basic Ideas in DODH• Maintenance of subject vocabularies
– Librarians’ major concern on domain specific vocabularies
• This is still an open question.
– Need software tools to maintain vocabularies• Metadata Schema Registry• Semantic Web technology
Some Issues on Subject Vocabularies
• Comprehensive and conventional subject vocabularies are not always useful for domain-specific resources.– Regional governments’ vocabularies for
classifying resources
• Comparison between the three subject vocabularies– Mappings between all pairs of these three
vocabularies show the characteristics of vocabularies and requirements for classifying resources from different perspectives.
Classification Vocabularies of Four Prefectural Governments in Japan
Ibaraki Kagawa Kanagawa Okayama
Number of Top Level terms 23 11 14 16
Max Depth 2 2 2 3
Total Number of Terms 115+ 82 88 280+
“+” means that terms to express names of regions or organizations
are excluded.
Okayama
Ibaraki
Kanagawa
Kagawa
Tokyo
NDC 000 100 200 300 400 500 600 700 800 900 total
#Kids
terms17 8 8 196 58 54 28 62 6 6 443
#NDC in KV
7 7 3 44 27 27 20 26 4 4 169
#Pref
terms15 2 12 171 30 56 44 17 1 1 349
#NDC in PV
4 2 5 34 11 18 25 15 1 1 116
Distribution of Terms in the NDC term space - Okayama’s Case -
NDC: 000=Generalities, 100=Philosophy, 200=History, 300=Social Sciences, 400=Natural Sciences, 500=Technology, 600=Industry, 700=The Arts, 800=Language, 900=Literature #NDC in KV/PV: the number of NDC terms in x00 used in the KV/PV mapping
Subject Vocabularies - Okayama’s Case -
NDC
PV KV
Social ScienceTechnology
Natural ScienceArtsIndustries
Comprehensive
Educational and Learning Resources
Resources for Government and Social Activities
General Resources
Discussions • Interoperability vs. Domain Specificity
– Reasonably small and domain-specific vocabulary is advantageous for domain-specific applications.
– Domain-specificity is, in general, disadvantageous for interoperability
• Maintainability of Vocabularies– Long-term maintenance of vocabularies is expensive.
• Reusability and customizability– Neighboring communities would need to share a
vocabulary in part.
• Need a good model to solve these issues
A Model for Reusing Metadata Schemas
termA: Mandatory
termC: Optional
Repeatable
termX: Mandatory
Repeatable
termZ: Mandatory
if applicable
termX termY termZ
Metadata Vocabulary 2(Metadata Element Set)
Metadata Vocabulary 1(Metadata Element Set)
termA termB termC
A structural view of application profile
A Model for Reusing Metadata Schemas
termA: Mandatory
termC: Optional
Repeatable
termX: Mandatory
Repeatable
termZ: Mandatory
if applicable
Application Profile: Terms used in an application and structural constraints
termX termY termZ
Metadata Vocabulary 2(Metadata Element Set)
Metadata Vocabulary 1(Metadata Element Set)
termA termB termC
<rdf:Description about=”foo”> <mv1:A>an example.</mv1:A> <mv2:X>bar</mv2:X> ...
<meta name=”mv1:A” content=”an example”><meta name=”mv2:X” content=”bar”>
...
Description in a syntax defined in an application
Application Profile
Abstract Syntax and Concrete Syntax
A Model for Reusing Metadata Schemas
Layered Model of Metadata Schema
Layer 1Semantics
Layer 2Abstract Syntax
Layer 3Concrete Syntax
DCMES (Elements and
Qualifiers)IEEE-LOM ULIS element
extension
DCMI Library Application
Profile
Open Archives Initiative Schema
IPL Asia Schema
ULIS Core Schema
XML implementation
in an XML Schema
An Oracle implementation
RDF implementation
A Layered Modelsplit semantics and syntax into layers
A Model for Reusing Metadata Schemas
Layer 1
Layer 2
Layer 3
DCMI Registry
DCMES Terms
(Elements and Qualifiers)
ULIS element extension
ULIS-DL Subject
Vocabulary
Tsukuba Registry
Application Profile A Application Profile B
Layered Model and Metadata Schema Registry
A Model for Reusing Metadata Schemas
Layer 1
Layer 2
Layer 3
DCMI Registry
DCMES Terms
(Elements and Qualifiers)
ULIS element extension
ULIS-DL Subject
Vocabulary
Tsukuba Registry
XML Schema for A XML Schema for B
Layered Model and Metadata Schema Registry
A Model for Reusing Metadata Schemas
Layer 1
Layer 2
Layer 3
DCMI Registry
1. Use registries to share, reuse, customize and maintain vocabularies
2. Develop software tools to support these functions.
Tsukuba Registry
Layered Model and Metadata Schema Registry
Discussions • Need good model and software tools
– DCAM provides the basic architecture of metadata elements and metadata descriptions
– The layered model presented is designed in order to enhance reusability/customizability of metadata schemas by splitting semantic and syntactic features of metadata schemas.
• Metadata Schema Registry as a key component to share metadata schemas– Collaborating registries to collect community-specific
schemas and to share the schemas among communities– Software tools attached to a registry, e.g.
• Vocabulary maintenance tool attached to a registry,• Metadata application software generator