Upstream Metadata

1
Upstream Metadata

• Library needs and workflows
• Diane Boehr
• Head of Cataloging
• National Library of Medicine, NIH, DHHS
• boehrd_at_mail.nlm.nih.gov

2
Role of the cataloger

• Descriptive metadata
• Authority work
• Subject analysis

3
Past Practices

• In the past, there was no way for catalogers to
  get information into the catalog without manually
  transcribing the data from the piece

4
Whats changed?

• Even for print products, there is now metadata
  available in electronic format early in the
  creation process, particularly for the basic
  descriptive information

5
LC Working Group on the Future of Bibliographic
Control

• The Library of Congress commissioned a Working
  Group in the fall of 2006 charged to
• Present findings on how bibliographic control
  and other descriptive practices can effectively
  support management of and access to library
  materials in the evolving information and
  technology environment
• Recommend ways in which the library community
  can collectively move toward achieving this
  vision
• Advise the Library of Congress on its role and
  priorities

6
LC Working Group

• Their final report was issued in Jan. 2008,
  entitled On the Record http//www.loc.gov/biblio
  graphic-future/news/lcwg-ontherecord-jan08-final.p
  df

7
On the Record

• Recommendation 1
• Increase the efficiency of bibliographic record
  production and maintenance
• 1.1 Eliminate redundancies
• Libraries have so far taken minimal advantage of
  descriptive data being created in other venues.
  Given the explosion of material requiring
  bibliographic control, the model of item-by-item
  full manual transcription can no longer be
  sustained

8
On the Record

• 1.1.1 Make Use of More Bibliographic Data
  Available Earlier in the Supply Chain
• 1.1.1.1 Be more flexible in accepting
  bibliographic data from others that do not
  conform precisely to U.S. library standards
• 1.1.1.3 Develop standard crosswalks for the
  conversion of publisher/vendor data to library
  system formats
• 1.1.1.4 Develop managed processes for creating
  and sharing conversion programs so that
  programming is not done redundantly at multiple
  institutions

9
National Library Needs

• A great deal of original cataloging done by the
  national libraries is for CIP (cataloging-in-publi
  cation) titles
• Recommendation 1.1.3 Fully Automate the CIP
  Process

10
Potential CIP data workflow

• Publishers participating in the CIP program would
  submit their descriptive metadata to LC in ONIX
  (or similar XML format that could be easily
  mapped to a MARC record) and then be ingested
  directly into the library ILS using a standard
  crosswalk
• Catalogers would no longer have to highlight and
  paste data from a title page image and could
  focus on the tasks of authority work and subject
  analysis to create the complete CIP record

11
Advantages

• Streamlined workflow
• Publishers get their CIP data back promptly
• Catalogers time is freed up to devote to
  material lacking upstream metadata, often the
  unique and hidden treasures of their
  collections (another LC report recommendation)
• Users get richer records, with summaries and
  tables of contents if the publisher provides that
  data

12
Non-CIP Data Workflows

• Publishers could store their descriptive metadata
  openly on their websites in ONIX (or similar XML
  format easily converted to MARC)
• Vendors would be able to supply this data to
  libraries along with the books
• Libraries could harvest that data and import
  directly into their catalogs using readily
  available conversion programs

13
Related Needs

• Development of conversion programs from ONIX to
  MARC and a central repository for these (a
  possible role for LC or OCLC or NISO)
• Even betterrevise ONIX so that the data coming
  from publishers is more consistent and only one
  conversion program is needed

14
How OCLC Might Help

• For CIPs, publishers could continue to submit the
  galleys and CIP data forms to LC, while
  simultaneously submitting the descriptive
  metadata to OCLC
• OCLC creates the basic preliminary MARC record.
  The national library imports this record into
  their ILS to complete the catalogingenhancing
  the record with controlled name access points,
  subject headings and classification numbers
• The completed catalog record is uploaded to OCLC
  and sent to the publisher for printing in the
  book

15
Other Potential Enhancements

• OCLC searches the WorldCat database for possible
  author matches and supplies likely name access
  points where possible
• OCLC searches the WorldCat database for other
  editions of the work and supplies suggested
  subjects and classification
• Catalogers work is greatly streamlined

16
Potential Enhancements Outside of OCLC

• NLM already has the ability to run the
  preliminary MARC record through its Medical Text
  Indexing software and get suggested MeSH
• LC could do similar types of automated analysis
  on their database to suggest subjects and
  classification for new manifestations
• Mappings could be developed between MeSH and LCSH
  to streamline subject analysis on shared records

17
Author Identifiers

• Work would be greatly streamlined if the
  community developed an author identifier standard
  
• This numeric ID would be used by publishers to
  allow libraries to correctly identify and
  disambiguate authors

18
Publishers Metadata

• For upstream metadata to work efficiently
  libraries need complete, accurate, consistent
  data from the publishers
• It does not need to be in MARC format, nor does
  it need to follow ISBD styles of capitalization
  and punctuation
• Publishers must be aware of what constitutes a
  chief source of information, e.g. book title
  must come from the title page, not the cover or
  spine

19
End