Deep-Indexing the OPAC: Integrating Contents Information into Search Results

1
Deep-Indexing the OPAC Integrating Contents
Information into Search Results

• Mary M. Strouse, CUA DuFour Law Library
• 7th MAIUG Annual Conference, October 2005.

2
Functions of Contents Information (TOC Data)

• Evaluation
• Does this resource suit my purpose?
• Navigation
• Which volume(s), pages do I need?
• Identification/Collocation
• What does the library have about?
• Written by?
• Containing known title?

3
Local Priorities for TOC Inclusion

• Edited collections on broad themes
• Diverse geographic treatments
• Conferences, symposia, anthologies
• Local interest (our faculty, etc.)
• Unifying themes multiple authorship and/or
  non-predictable content

4
Where, When and How

• Choice of vendors
• Blackwell
• Syndetic Solutions
• Marcive (using Syndetics data)
• Scanning/local input
• Loading mechanisms
• III loading services (Blackwell)
• Matching on demand
• Choice of formats

5
505 Enhanced Contents Note

• Standards-compliant
• Existing tools (macros, spell-check)
• Includes volume and chapter, but not page
• Keyword access to titles and authors
• Titles indexed (all or nothing)
• Cant index authors (not in inverted form).
• Can be difficult to read

6
Vendor TOC Format (97x)

• Displays as a table of contents
• Includes page numbers
• Indexing flexibility authors and titles as well
  as keyword
• Can exclude generic titles (Introduction
  Preface, etc.) from indexing
• Space for both transcribed and authorized forms

7
97x TOC Format Detail

• 970 field (one per chapter or section title)
• Indicator 1 title indexing
• 0 Non-distinctive title (dont index)
• 1 General chapter title or heading
• 2 Citable title No longer used
• Indicator 2 hierarchy, degree of indentation
• l Section or chapter label
• t Section or chapter title
• c Personal author
• f Personal author in inverted form
• d Non-personal author
• e Editor
• p Starting page number
• by default, authors of non-distinctive titles
  are not indexed. May be indexed on request.

8
Disadvantages of TOC Format

• Table-like format takes up screen, adds
  significantly to printing
• Limitations on use of vendor data (TOC blocked in
  exported results)
• Implicit burden on library staff
• Commingling of library and vendor data

9
Display bug Corporate author doesnt display
10
Effect on Keyword Search

• Adds significantly to retrieval in keyword
  searches authors and titles
• Display element in keyword results is always the
  book title also true of sorted/limited
  results.
• Search terms are highlighted in full record
  display

11
Effect on Title Search

• Library determines which titles to exclude (970
  first indicator)
• Chapter titles will appear in unsorted results
  browse
• Chapter titles not identified as such
• English initial articles automatically excluded
• Search terms highlighted in full record display

12
Effect on Author Search

• Individual authors are linked in record (as
  transcribed) and appear in browse list (indexed
  form)
• Authority work often needed to match with
  existing names
• Corporate authors from 970 d and editors from
  970 e not indexed
• Display of titles in extended browse follows same
  rules as title search

13
Keyword-only Indexing Option

• Includes authors and titles
• Must specify inclusion in author and title
  segments
• Avoids collocation issue/authority work
• Avoids noise retrieval, confusion between
  chapters and books
• Limits access to documents and reports (distinct
  works)
• Limits effectiveness of known-author and
  known-title searches

14
Formatting Controls

• BIB_TOC_HEADER WWWoption
• Places a caption at head of TOC display
• Default no caption
• Accepts HTML for formatting or link to a help
  file
• TABLEPARAM_BIB_TOC
• Stylesheet Class bibTOC
• No link in brief citation

15
Search result display options

• DISPLAY_245 does not apply to chapter titles
• EXTENDED_TU will not force a book title to
  display
• Beware confusion from forcing extended display
  (INDEX_EXTta)

16
BROWSE WWWoption

• Controls first line of index browse
• BROWSE_T controls first line of record browse
  (in absence of briefcit.htm )
• If no 970 subfields are specified, all subfields
  will display
• If specify default subfields for non-245 titles,
  must include subfield t

17
Example 1 BROWSE_T245/abnp/c or
BROWSE_T245/abnp/c 970/t/cd/a/c (ALL TOC
subfields display)
18
Example 2 BROWSE_T245/abnp/c 970/t/cd
and BROWSE_T245/abnp/c 970/t/cd /at/c
19
Briefcit Format

• ltspan class"briefcitTitle"gt
• lt!--linkfieldspecVbT--gt
• lt/spangt
• All record browse screens show book titles
  (includes limited and keyword results)
• All index browse screens show chapter titles
  (includes sorted results)
• Use BROWSE_T (define 970 t to avoid no title
  display)

20
Briefcit Format

• ltspan class"briefcitTitle"gt
• lt!--linkfieldspecVbt245abnp--gt
• lt/spangt
• All record browse screens show book titles
• (includes sorted, limited and keyword)
• Only system-sorted index browse shows chapter
  titles

21
Loading and Workflow Issues

• False adds monographic series w/ ISSNs,
• False drops CIP and other title discrepancies
• Coding consistency
• Authority control
• Volume of work
• Lack of tools
• No mechanism to identify/protect library-added
  data

22
Coding issues vendor- supplied data

• Titles and names transcribed from TOC, not from
  fullest form available
• No space for formal titles of included works
  -- we add 7xx
• Inconsistent coding of index-worthiness
• Is Appendix a title or a number?

23
Non-personal Authors d

• Used for corporate author
• 970 11 l9 tRedefining Discrimination
  'Disparate Impact' and the Institutionalization
  of Affirmative Action d United States Department
  of Justice Office of Legal Policy p121
• Also used for personal authors in direct order
  (but sometimes not)
• 970 12 tExcerpts from Antigone d Sophocles p11
  
• 970 11 tReith Lecture 2000 d The Prince of
  Wales p11

24
Non-personal Authors d

• Also used for other transcribed phrases and et
  al.
• 970 21 tWorkshop Discussion Civil Litigation
  Against Terrorism d Workshop Participants p185
• 970 21 tPublic Support for Access to Government
  Records A National Survey cPaul D. Driscoll
  fDriscoll, Paul D. cSigman L. Splichal
  fSplichal, Sigman L. cMichael B.
  SalwenfSalwen, Michael B. d et al. p23
• Library can add index link in f (not
  vendor-provided)

25
Recap User Issues

• Cost in screen space, added printing
• Multiple forms of author entry (split files)
• Cant distinguish between chapter and book-length
  treatment (increased noise)
• License limitations on data use

26
Wish List

• Fix corporate author display bug
• Identify chapter titles in search results
• Option to force display of both chapter title and
  book title in extended browse
• Link to full TOC display from brief citation
  format (briefcit.html)
• Allow limited data export for legitimate
  scholarly use

27
Recap Workflow Issues

• Vendor-dependent format
• Staff burden need coding regularization
• Co-mingling of vendor and library data
• False positives (multiple ISBN)

28
Wish List

• Additional Subfields/Codes
• Indexed/authorized form for corporate author
• Data source and ownership
• Authority history
• Subfield code(s) to identify library-added TOC
  data, overlay-protect library-added authority
  work
• Better coding conventions, transparency

29
References

• CSDirect TOC Data FAQ (password required)
• http//csdirect.iii.com/faq/tocfaq.shtml
• Blackwell TOC Enrichment brochure
• http//www.blackwell.com/pdf/TOCEnrichment.pdf
• Vendors
• http//www.blackwell.com/level2/TOC.asp
• http//www.syndetics.com/index.htm
• http//www.marcive.com/HOMEPAGE/MARCres.htm
• (Marcive uses Syndetics data)

30
Contact

• Mary M. Strouse
• Head of Technical Services
• Judge Kathryn J. DuFour Law Library
• Catholic University of America
• strouse at law.cua.edu

Thank You!