ISO/IEC JTC1/SC32 WG2 Interim Meeting Seoul, Korea
Study report on SMM process
2007. 12. 6 Tae-Hoon Lim and Tae-Sul Seo taehoon@dpc.or.kr tsseo@kisti.re.kr
Background
According to the resolution of SC32 New York meeting (SC32N1604a), the study on Semantic Harmonization of Metadata was performed.
Reference: SC32N1658
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
2
Summary
Title was changed. The procedures were modified.
Name of each step was changed The 2nd and 3rd steps can be replaced by each other.
A description system for mapping was established.
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
3
Title change
From “semantic harmonization of metadata” To “semantic metadata mapping (SMM) process” The later is more specific expression than the former.
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
4
Procedure modification
1st Surveying metadata sets 2nd Constructing common DECs based on 11179 3rd Grouping data elements by the DECs
1st Collecting metadata schema 2nd Grouping attributes 3rd Finding common DECs
4th Completing crosswalks
4th Mapping into a table
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
5
Semantic Metadata Mapping Process
Overall Process
1st Collecting metadata schema 2nd Grouping attributes 3rd Finding common DECs 4th Mapping into a table
The 2nd and 3rd steps can be replaced by each other.
2007-12-06 SC32 WG2 Interim Meeting, Seoul, Korea 6
Semantic Metadata Mapping Process
1st. Collecting metadata schema
Survey and identify candidate metadata schema in a domain. Surveying form includes:
Domain name, Service DB name, or an other equivalent name. Number of fields Sample data Value domains
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
7
Semantic Metadata Mapping Process
2nd. Grouping attributes
Selecting a metadata set as a primary metadata set.
The simplest or the highest level metadata set is desirable to be the primary one.
For all available metadata schema, attributes should be aggregated by the attributes of the primary metadata set.
There may exist attributes which aren’t fitted to any of them.
Some attributes, which are not important, may be removed.
The remaining are grouped separately.
Metadata experts should perform the work along with domain experts.
2007-12-06 SC32 WG2 Interim Meeting, Seoul, Korea 8
Semantic Metadata Mapping Process
3rd. Finding common DECs
Analyzing each attribute of the primary metadata set and find out an object class and a property hidden in and related to the attribute. Constructing common DECs based on ISO/IEC 11179 standard using the object classes and the properties. If there exists an attribute which isn’t fitted to any of the DECs, a new DEC may be constructed for them.
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
9
Semantic Metadata Mapping Process
4th. Mapping into a table
Finally, arranging all attributes into a table by the common DECs. Comments on the types of mapping can be included in the table as bellow.
Same, no difference: no description Level difference: upper/lower terms
Domain difference: generic/specific (book, technical report, article, …)
Term difference: synonym, antonym or preferred term Naming rule difference: Order or representation rules
A recommended set of metadata can be provided for guiding future standardization.
2007-12-06 SC32 WG2 Interim Meeting, Seoul, Korea 10
Application to e-Book
Domain: e-Book (1st) Available metadata sets: OpenEBPS, MODS and TEI primary metadata set: OpenEBPS
OpenEBPS Domain name Description of Electronic Book MODS Description of Library resources About 60 (top level: 20) no TEI header Encoding methods for machine-readable texts Over 20 yes
Number of fields 15 Sample data yes
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
11
Application to e-Book
(2nd) Grouping attributes
OpenEBPS title titleInfor:title titleInfor:subTitle titleInfor:partNumber titleInfor:partName titleInfor:nonSort creator(role) creator(file-as) name:role name:namePart name:displayForm name:affiliation name:discription subject subject:topic classification subject:catographics subject:occupation profileDesc:textClass:keyword profileDesc:textClass:classCode profileDesc:textClass:catRef fileDesc:titleStmt:author MODS TEI fileDesc:titleStmt:title fileDesc:seriesStmt:title fileDesc:seriesStmt:idno
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
12
Application to e-Book
(3rd) Constructing common DECs based on 11179: Object class: e-Book Properties: title, author, subject, abstract, publisher, distributor, authority, contributor, publication-date, genre, format, extent, identifier, language, coverage-geographic, coverage-temporal, right, location, edition DECs: ebookTitle, ebookAuthor, ebookSubject, ebookAbstract, ebookPublisher, ebookDistributor, ebookAuthority, ebookContributor, ebookPublication-date, ebookGenre, ebookFormat, ebookExtent, ebookIdentifier, ebookLanguage, ebookCoverage-geographic, ebookCoverage-temporal, ebookRight, ebookLocation, ebookEdition
2007-12-06 SC32 WG2 Interim Meeting, Seoul, Korea 13
Application to e-Book
DEC ebookTitle
(4th) Mapping into a table
OpenEBPS title MODS titleInfo:title TEI titleStmt:title Recommaned DE ebookTitle
titleInfo:subTitle
ebookAuthor creator(role) creator(file-as) T:pre name:role name:namePart D:gen
seriesStmt:title
T:pre
ebookSubtitle
titleStmt:author
N:rep
ebookAuthorName
ebookSubject
subject
N:rep
subject:topic
classification
T:pre
N:rep
textClass:keyword
textClass:classCode textClass:catRef
N:rep
N:rep T:pre
ebookSubjectWord
ebookSubject-classCode
L - up: upper term/lo: lower term D - generic: gen/… T - syn: synonym/ant: antonym/pre: preferred term N - ord: order/rep: representation
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
14
Future plan
The SMM process will be elaborated more in order to be proposed as a new work item in ISO/IEC JTC1/SC32 next year.
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
15
Thank you!
2007-12-06
SC32 WG2 Interim Meeting, Seoul, Korea
16