Indexing and Cataloging

1
Indexing and Cataloging

• Similarities and Differences
• Diane Boehr
• Head of Cataloging, National Library of Medicine
• for the ALCTS Continuing Resources Update Forum
• July 13, 2009
• Chicago, IL

2
Why would a library want to provide article level
access to material?

• The most up-to-date information is in the journal
  literature
• Want to fully expose the richness of our
  collections
• We believe users want one-stop shopping and may
  not know whether the material they are looking
  for is in an article or a book

3
So why dont more cataloging agencies index
material?
4
Volume!
5
Historical Development at NLM

• Original workflows at NLM did combine indexing
  and cataloging

6
Historical Development

• Index-Catalogue of the Library of The Surgeon
  Generals Office began 1880
• Intended to detail the actual holdings of NLM
• Set forth in dictionary order both books and
  periodical articles. Books listed by author and
  subject, articles by subject only
• Volumes were released alphabetically and it took
  15 years for the first full set to be issued

7
Historical Development

• Index Medicus began in 1879 as a monthly
  publication, focusing on new articles from
  selected journals arranged topically
• Each quarter, one person arranged and
  alphabetized 15,000-20,000 index cards in
  preparation for printing
• In 1921 it changed to a quarterly publication,
  published by AMA and also went to an alphabetical
  arrangement for subjects

8
Historical Development

• Large overlap in material in Index-Catalogue
  (books and serials) and Index Medicus (chiefly
  articles only)
• Index-Catalogues lack of currency was chiefly
  because of its multivolume alphabetical
  arrangementtook 20 years to publish each full
  set of volumes and begin alphabet again

9
Historical Development

• A 1943 survey of the library performed by an ALA
  team was critical of Index-Catalogue
  (inconsistent completeness of coverage, lack of
  uniformity in subjects, use of little known
  headings, lack of x-refs), and said that the
  library had focused on it to the detriment of
  reader service, acquisitioning, and cataloging
• In 1950 indexing of journals ceased for Index
  Catalogue, by 1961 publishing of Index Catalogue
  ceased
• In 1954 the first official list of subject
  headings was published by NLM

10
Current Practices

• The same vocabulary, MeSH, is used to describe
  both cataloging and indexing material
• The adoption of a single subject authority list
  for both books and periodical articles is a
  departure from traditional practice. We take the
  view that subject cataloging and periodical
  indexing, as exemplified in the Index Medicus and
  in the NLM Catalog, are identical processes in
  their major dimensions. A single list can and
  should be used for both purposes. This has two
  major virtues simplicity for users, in requiring
  familiarity with only a single scheme and
  economy to the Library in the development and
  maintenance of a single scheme. Frank Rogers,
  1960
• The Index and Cataloging Sections are in two
  separate divisions of the library

11
Numbers

• Indexing FY2008 figures
• 671,904 citations were indexed from 5,319
  journals
• Cataloging FY2008 figures
• Cataloged a total of 21,507 items
• 18,834 monographs and integrating resources
• 1,640 serials
• 1,033 AVs

12
Staffing

• Indexing has an internal staff of ca. 42 FTE and
  ca. 100 contractors
• Cataloging has an internal staff of ca. 32 FTE
  and ca. 10 contractors

13
Automation

• Because of the volume of work indexers must
  handle, the indexing process has been highly
  automated
• NLM has developed its own input, tracking, and
  indexing system (DCMS-Data Creation and
  Maintenance System)

14
Automation

• NLM has developed a Journal Publishing DTD that
  has become a de facto standard for STM publishing
• The Journal Publishing Tag Set defines elements
  and attributes that describe the content and
  metadata of journal articles. The Tag Set allows
  for descriptions of the full article content or
  just the article header metadata

15
Automation

• Because of consistency in article presentation,
  the DTD is widely used by publishers
• NLM is able to machine load the descriptive
  metadata for over 85 of the articles it indexes
  from XML supplied by the publishers
• Descriptive metadata for the remaining 15 is
  loaded by OCR scanning (done under contract)

16
Automation

• Cataloging, on the other hand, has to deal with
  material in a wide variety of formats, with no
  standard presentations
• Even publishers of print material have not been
  able to fully agree on or widely use a standard
  XML format for descriptive info, so catalogers
  are still forced to input this data manually

17
Automation

• Because the DCMS is an internally developed
  system, NLM has complete control and can design
  it to best meet the needs of indexers and tightly
  couple it with other in-house products, such as
  MeSH and MTI (Medical Text Indexer)
• Cataloging is using a commercial product
  (Voyager) that makes it difficult to integrate
  other tools and take advantage of many of the
  efficiencies developed for indexers

18
Subject Expertise

• Indexers assign many more subject headings and to
  a greater degree of specificity than catalogers
  usually do
• Our indexers generally have an MD or a Ph. D. in
  their area of expertise
• Catalogers tend to be generalists and have an MLS

19
What Indexers Dont Do

• Indexers only supply subject headings, they do
  not create any descriptive data
• No assignment of classification numbers
• No authority work for names

20
Are the Jobs Interchangeable?

• Both require
• Intelligence and good judgment
• Fluency in foreign languages
• Subject knowledge

21
Are the Jobs Interchangeable?

• A good cataloger could probably adjust to
  indexing more easily than an indexer could adapt
  to cataloging, because the primary indexing task
  is subject analysis
• Indexers have no equivalent to complex
  descriptive rules like AACR2, since records
  traditionally have not been shared among AI
  providers

22
Considerations Before Making a Final Decision

• Clarify your goals and priorities
• What added value are you providing?
• Is the title already indexed by an AI service?
• Explore other options for opening up your
  existing citations, such as a federated search
  engine or arranging a license with an AI service
  to load their citations in your catalog

23
Considerations

• Does staff have expertise in the subject area to
  be covered?
• Will you use the same staff and the same
  procedures for both types of material?
• If the procedures differ, what impact will that
  have on display and retrieval?

24
Considerations

• Do users really want one-stop shopping?
• Figures from NLM databases
• Medline (articles only) 776 million annual
  searches
• NLM Catalog (books, journals, AVs) 4.7 million
  annual searches
• NLM Gateway (federated search) 998,000 annual
  searches

25
Good Starting Points

• Special collections in fields where the primary
  journals are not covered by existing AI services
• Material in your institutional repository

26
Credits and citations

• Drowning in the paper sea. Sept. 10, 2008.
  Picture courtesy of Megan Mayer.
• Greenberg, Stephen. The great contribution
  Index Medicus, Index-Catalogue and IndexCat. J
  Med Libr Assoc 97 (2) April 2009.
• Miles, Wyndham Davies. A history of the National
  Library of Medicine. Bethesda, MD, US Dept. of
  Health and Human Services, 1982.