Motivating the Semantic Web

1
MARCXTM Topic Maps Modeling of MARC
Bibliographic Information
2005.10.07
Hyun-Sil Lee, Yang-Seung Jeon, Sung-Kook
Han Semantic Web Services Research Group Won
Kwang University, Korea
2
Agenda
3
Overview MARC

• MARC Machine-Readable Cataloging
• standards used for the representation of
  bibliographic and related information for books
  and other library materials in machine-readable
  form and their communication to and from other
  computers.
• All MARC Standards conform to ISO 27091996
  Information and documentation - Format for
  Information Exchange.
• MARC was originally designed in the late 1960s
  to aid in the transfer of bibliographic data onto
  magnetic tape, and also to replace the printed
  catalog cards with electronic forms.
• There are a number of implementation of MARC,
  including USMARC used in US, CAN/MARC used in
  Canada, and UKMARC used in Britain.
• After discussions and minor changes to USMARC and
  CAN/MARC, MARC21 was evolved to harmonize both
  formats and to cover diverse types of resources
  including digital materials and Internet
  resources.
• MARC accommodates extensive data elements
  describing all forms of materials susceptible to
  bibliographic description, as well as related
  information.

4
Family of MARC Formats

• Bibliographic
• a carrier for bibliographic information about
  printed and manuscript textual materials,
  computer files, maps, music, serials, visual
  materials and mixed materials.
• Authorities
• a carrier for information concerning the
  authorized forms of names, titles,subjects, and
  subject subdivisions to be used in constructing
  access points in MARC records, the forms of these
  names, subjects and subdivisions that should be
  used as references to the authorized form, and
  the relationships among these forms
• Holdings
• a carrier for holdings information for three
  types of bibliographic items single-part
  multipart serial and may include copy-specific
  information information peculiar to the holding
  institution information needed for local
  processing, maintenance or preservation version
  information.
• Classification
• a carrier for information about classification
  numbers and the captions associated with them
  that are formulated according to a specified
  authoritative classification scheme
• Community Information
• a carrier for descriptions of non-bibliographic
  resources that fulfil the information needs of a
  community.

5
Supporting Documentation of MARC

• MARC 21 Specification for Record Structure,
  Character Sets, and Exchange Media
• Character sets
• MARC-8 (8-bit encoding)
• UCS/UNICODE UTF-8 (8/16 bit encoding)
• Repertoire of 15,000 characters
• Latin Cyrillic Hebrew Arabic CJK
• Code lists
• Countries, Geographical Languages Sources
  Relators

6
MARC Record Format
Leader the first 24 characters of the record defining parameters for processing the record data elements that contain coded values and are identified by relative character position the first 24 characters of the record defining parameters for processing the record data elements that contain coded values and are identified by relative character position
Directory directory entries that contain the tag used in variable fields, starting location, and length of each field within the record constructed by computer from the bibliographic record, and can be reconstructed in the same way if any of the cataloging information is altered directory entries that contain the tag used in variable fields, starting location, and length of each field within the record constructed by computer from the bibliographic record, and can be reconstructed in the same way if any of the cataloging information is altered
Variable Field Variable Control Field 00X fields in the MARC 21 formats are variable control fields. either a single data element or a series of fixed-length data elements identified by relative character position
Variable Field Variable Data Field Indicators The first two characters which interpret or supplement the data found in the field. Subfield codes Two characters that precede each data element within a field that requires separate manipulation
7
MARC Record Format Example
8
MARC Record Format Example
Sign Post
9
Formalization of MARC
ltMARC21RecordgtltLeadergtltDirectorygtltVariableFiel
dgt ltDirectorygtltDirectoryElementgt
ltDirectoryElementgtltTaggtltLengthgtltPositiongt
ltVariableFieldgtltControlFieldgtltDataFieldgt
ltControlFieldgtltControlNumbergtltControlFieldEleme
ntgt ltDataFieldgtltTaggtltIndicatorgtltSubFieldgt
ltIndicatorgtltFirstIndicatorgtltSecondIndicatorgt ltS
ubFieldgtltSubFieldCodegtltSubFieldValuegt
10
Problems with MARC

• Lack of expandability due to rigorous record
  formats, since it was originally intended for the
  production of printed catalogue cards in 1960s
• Difficulties in representing bibliographic
  relationships
• Ambiguities in describing MARC records
• Incompatibilities between other MARC formats
  since the various library systems have invented
  their own non-standard peculiarities in order to
  handle local bibliographic materials
• Weaknesses in describing bibliographic attributes
  of digitized resources

11
MARCXML
MARC21 (2709)Records
MARC21 (XML) Records
Tagging Transformations
Character Set Conversion
Dublin Core Records
MODS Records
Other XML Formats
HTML Output
MARC Validation
12
MARCXML

• MARCXML a framework for working with MARC data
  in a XML environment
• Design Considerations and Features
• Simple and Flexible MARC XML Schema for
  representing a complete MARC record in XML
• Supports all MARC encoded data regardless of
  format
• Lossless Conversion of MARC to XML
• Roundtrip ability from XML back to MARC
• Data Presentation and Data Conversion
• Extensibility
• A component-oriented, extensible architecture
  allowing users to plug and play different
  software pieces to build custom solutions

13
MARCXML Example
14
MODS

• MODS Metadata Objects Description Schema
• XML-based descriptive metadata standard that
  includes a subset of data elements derived from
  MARC21
• Features
• MODS is intended to complement other metadata
  formats. MODS provides a richer bibliographic
  element set than Dublin Core.
• MODS has a high level of compatibility with MARC
  records because it inherits the semantics of the
  equivalent data elements in the MARC21
  bibliographic format.
• In MODS some elements that appear in various
  fields in MARC have been repackaged into one. So
  MODS can define 19 upper metadata elements.
• MODS takes advantage of the XML environment. It
  uses language-based tags rather than the numeric
  tags traditional to MARC.
• MODS also has flexible linking mechanisms by
  providing for all the top-level elements with
  attributes such as xlink and ID.
• MODS accommodates special requirements for
  digital resources.

15
MODS Example
16
Topic Maps Modeling of MARC 21

• Requirements for MARC Modeling
• A model should be able to support the full set of
  data elements in MARC21 to achieve seamless
  compatibility with MARC formats.
• This is a practical requirement in order to
  embrace the current circumstances even though it
  is awkward.
• It should have the same expressive power as
  metadata.
• This implies that the model should be realized
  with semantic descriptors to be used in an XML
  environment instead of obsolete alphanumeric
  codes.
• The use of attributes should be minimized to
  maintain consistency and increase readability.
• It should be able to maintain the structure of
  MARC record format
• A model does not intend to develop bibliographic
  metadata system based on MARC.
• A model can be handled without expertise in MARC
  to achieve the usability of the model.
• A model should be simple and lightweight for
  system implementation and harmonization with
  other models.

17
UML diagram of MARC Modeling
18
MARCXTM Implementation
19
XTM Realization of MARC Specification

• DataField ltassociationgt of data item,
  indicators, and subfield codes

20
XTM Realization of MARC Specification

• Hiding the real data value by topic abstraction

lttopic id"TypeOfPersonalNameEntryElement"gt
ltbaseNamegt ltbaseNameStringgt Type of
personal name entry element lt/baseNameStringgt
lt/baseNamegt ltoccurrencegt ltinstanceOfgt
lttopicRef xlinkhref"Forename"/gt lt/instanceOfgt
ltresourceDatagt 0 lt/resourceDatagt
lt/occurrencegt ltoccurrencegt
ltinstanceOfgt lttopicRef xlinkhref"Surname"/gt
lt/instanceOfgt ltresourceDatagt 1
lt/resourceDatagt lt/occurrencegt
ltoccurrencegt ltinstanceOfgt lttopicRef
xlinkhref"FamilyName"/gt lt/instanceOfgt
ltresourceDatagt 3 lt/resourceDatagt lt/occurrencegt lt/t
opicgt
21
MARCXTM for MARC Specification
22
XTM Realization of MARC Records

• MARC Records
• Complex to maintain MARC structure due to its
  idiosyncratic dependency between indicators and
  subfield code
• Difficult to realize the seamless compatible with
  MARC records
• Repeatability of subfield elements are
  individually defined in MARC specification.
• XTM supports for MARC modeling
• XTM does not provide multiple instances for
  ltoccurrencegt.
• Difficult to define record schema with
  ltassociationgt.

23
XTM Realization of MARC Records
24
MARCXTM for MARC Records
25
Conclusions

• MARCXTM Topic Maps-based implementation of MARC
  21
• MARCXTM for MARC Specification
• MARCXTM for MARC Records
• Application of Topic Maps paradigm to
  bibliographic information system
• Seamless compatible with MARC 21
• expressive power as metadata
• XTM is inappropriate to represent MARC format due
  to its idiosyncratic structure and dependency
  between data elements.
• Metadata models similar to Dubline Core or MODS
  are necessary for XTM modeling of MARC.
• FRBR (Functional Requirements for Bibliographic
  Records) framework is an attractive model for XTM
  modeling of bibliographic information system.

26
Thank you!!!