FAST (Faceted Application of Subject Terminology): A Vocabulary to Facilitate Faceted Browsing

1
FAST (Faceted Application of Subject
Terminology) A Vocabulary to Facilitate Faceted
Browsing
OCLC / ISKO-NA Preconference
Everything Need Not Be Miscellaneous Controlled
Vocabularies And Classification In A Web
World Sponsored by OCLC, ISKO-NA and Université
de Montréal Université de Montréal 5 August 2008

• Eric Childress
• Consulting Project Manager
• OCLC Programs Research

2
Outline

• A Changing World
• About FAST
• Illustrative uses of FAST

3
A Changing World

• Search Retrieval evolves
• Interface expectations are set by others
• Presentation and navigation patterns flow
• Data/Metadata production sources expand
• Machine services are increasingly important

4
Search Retrieval evolves
Library catalogs

19th Century Author, Title, Subject Card
Catalogs Manual Complex, Non-intuitive Controlled
Vocabulary
high unit costs
5
Interface expectations are set by others
Top Web properties by visitor June 2008 (SEW)
6
Presentation and navigation patterns flow

• Losing favor
• Text-heavy interfaces
• All options, all the time
• Jargon-laden labels

• Gaining favor
• Single-search box simplicity
• White-space and eye-catching graphics
• Side-bar navigation
• Tagclouds

7
(No Transcript)
8
Closer to, but not libraries
9
Traditional library interfaces
10
Newer library system interfaces
11
Tagclouds
12
Data/Metadata production sources expand

• Institution-based
• Sophisticated rules
• Expert-built
• Authority
• Formal distribution channels
• Institutional networks

cost accounting
13
Machine services are increasingly important
14
About FAST

• Arose from expert recommendations
• Project of OCLC Research in consultation with LC
• Complement of faceted vocabularies based on LCSH
• Suitable for non-expert application
• Machine-to-machine friendly

15
A New Frontier for Controlled Vocabularies

• Expert group studied issues related to controlled
  vocabularies and the Web environment
• American Library Associations ALCTS/SAC/Subcommit
  tee on Metadata and Subject Analysis (1997-2001)
• Conclusion For certain circumstances a
  controlled vocabulary was needed which was
• Web-friendly
• Low-learning curve/non-expert user-friendly
• Existing major vocabularies were not ready as-is
• Adapting an existing vocabulary might be a
  pragmatic option

16
Requirements for a New Vocabulary

• Simple in structure and syntax
• Usable by non-catalogers and in non-library
  environments
• Compatible with MARC, Dublin Core, and other
  popular metadata schemas
• Easy maintainability
• Machine-compatible

17
Launching FAST

• OCLC Research project launched in 1998
• Advisory group ALA/ALCTS/SAC Subcommittee on
  FAST (Faceted Application of Subject Terminology)

• Team
• OCLC Eric Childress, Kay Clapton, Becky Dean,
  Anya Dyer, Kerre Kammerer, Ed ONeill (Lead),
  Diane Vizine-Goetz
• LC CSPO Lynn El-Hoshy (now retired), Janice
  Young
• Consultant Lois Mai Chan (University of Kentucky)

18
Why Adapt LCSH?

• Rich vocabulary covering all subject areas
• Synonym and homograph control
• Extensive hierarchical and associative
  references among terms
• De facto standard controlled vocabulary,
  extensively used by libraries, contained in
  millions of bibliographic records
• Long and well-documented history
• Strong institutional support of the Library of
  Congress

19
What is FAST?

• OCLC FAST (Faceted Application of Subject
  Terminology)
• A faceted vocabulary based on LCSH
• Modular each facet may be used independently
• Supports post-coordinate search retrieval
• Designed for use by non-expert assigners
• Machine-friendly controlled vocabulary

20
Authority Control LCSH vs. FAST
LCSH FAST
Very large number (billions plus) of possible headings Faceting limits the number of possible headings to a few million
Common headings are established most assigned headings are synthesized by catalogers based on rules All headings (except chronological) are established
Most headings are distinct (based on NACO normalization rules) some conflicts occur particularly with x v All normalized headings are distinct tagging and subfield coding provides no unique information (with the exception of forms)
http\\www.loc.gov/catdir/pcc/naco/normrule.html
21
FAST - Eight Facets
22
How is FAST built?

• OCLC Research-built software processes LCSH
  authority file and LCSH present in WorldCat
  bibliographic records to automatically build
• FAST MARC authority file covering 8 facets
• FAST MARC authority reference records to assist
  with LCSH conversion

23
Sample Authority Record - Geographic
001    2130675 003    OCoLC 005
20040512160245.0 008    040512nneanzbabn n ana
d 040    OCoLC   b eng   c OCoLC   f fast 043
   n-us-ak 151    Pacific Ocean z Rowan Bay 670
   GNIS, Feb. 10, 2004   b (Rowan Bay bay 7
mi. N of Tebenkof Bay, on W coast of Kuiu I.,
Alex. Arch. Wrangell-Petersburg Census Area,
Alaska 56º40'02" N, 134º14'34" W another Rowan
Bay, pop. place in Wrangell-Petersburg Census
Area) 688    LC subject usage 0 (2006) 688    WC
subject usage 2 (2006) 751  0 Rowan Bay (Alaska
Bay)0 (DLC)sh2004005090
24
Sources of FAST Headings

• Library of Congress Subject Authority File
• LC headings that combine different facets are
  deconstructed into discrete headings, each
  containing only one facet.
• Headings assigned to bibliographic records in
  OCLCs WorldCat
• Many complex headings, i.e., those containing
  more than one element in the heading string, are
  based on literary warrant. They are derived
  from subject fields in the records in OCLCs
  WorldCat.
• Headings created for FAST
• In some cases, faceting has required FAST
  headings to be created when no LCSH equivalents
  exists.

25
Example of a FAST-only heading Events Facet

• In LCSH, it is common to established events as a
  combination of a geographic heading and a
  chronological (y) subdivision
• Buffalo (N.Y.) x History y Civil War,
  1861-1865
• Grenada x History y American Invasion, 1983

For these subdivisions, a FAST topical heading is
also created American Civil War,
1861-1865 American Invasion of Grenada, 1983
26
LCSH to FAST Comparison
27
Current FAST Authority File
Personal name headings 699,200
Corporate name headings 351,494
Topical headings 407,772
Geographic name headings 148,952
Chronological headings 676
Event headings 12,225
Title headings 48,245
Form headings 711
Total FAST authorities 1,669,375
July 2008
28
Future Development Plans

• Update and resynchronize all FAST headings with
  LCSH (In process)
• Improve the LCSH to FAST conversion (In process)
• Complete the FAST manual (In process)
• Expand the geographic names based on usage data
  and add information from the Geographic Names
  Information System (GNIS)
• Revise and expand the form (genre) facet

29
Sample OCLC applications of FAST

• WorldCat Identities
• FictionFinder

30
FAST headings
31
FAST headings
32
http//fictionfinder.oclc.org/
33
More information

• OCLC FAST
• Project page http//www.oclc.org/research/project
  s/fast/
• Search interface http//fast.oclc.org/
• OCLC WorldCat Identities
• http//orlabs.oclc.org/Identities/
• OCLC FictionFinder
• http//fictionfinder.oclc.org/