Bibliographic Control Alphabet Soup: AACR to RDA and the Evolution of MARC; a webinar report

These are my raw notes and will be updated and cleaned out....

Wednesday. October 14, 2009
1:00 PM to 2:30 PM
Sponsored by NISO: the National Information Standards Organization
Speakers:
Barbara Tillett, btil@loc.gov
Diane L. Hillmann, metadata.maven@gmail.com
William E Moen,william.moen@int.edu
some acronyms you'll encounter
IFLA: the International Federation of Library Associations
FRBR: Functional Requirements for Bibliographic Records
RDA: Resource and Description Access
ISBD: International Standard Bibliographic Description
FRBR was first published in 1998

• users tasks
• new view of universe
• groups and things
• networks
• data elements and attributes

 RDA was born from IFLA's cataloging principles and FRBR
In 2002 the draft for AACR3 was started: this evolved into RDA.
ICP, International Cataloging Principles, was developed between 2002 and 2003.  It was posted in February of 2009 and has been translated into 20 languages.
The world wanted cataloging rules that were defensible, not arbitrary.
Designed to build catalogers judgement of decsription and access.
The metadata for publishers is called ONIX.
Also incolved:
DublinCore and IEEE/LOM
RDA/MARC working group (MARBI)
RDA is web based
FRAD: Functional Requirements for Authority Data
Work, Expression, Manifestation and Item: that's what FRBR works with
Transcription: principle of Representation
No More Abbreviating!!!
no more "et al"
no more "s.l." or "s.n."
No more GMD (General material Description)
instead, Content, Media and Carrier
new MARC fields: 336, 337 and 338
RDA is not a display standard (MARC is)
No more "polyglot"
more data in the authority records
FRAD helped RDA a lot

In the past we had a very small view, now it is broader.  We have an expanded bibliographic universe.  We need to show more connections, relationships and pathways.

RDA is a change in technology, in focus and in view.
Metadata created with RDA can be used beyond the library.
Catalogs will no longer be isolated because they will all be online.  There will be global access to data.
Coding and mapping tables
registered controlled vocabulary
more open source
LCSH uses SKOS as their controlled vocabulary
(Library of Congress Subject Headings use Simple Knowledge Organization Schema)

We are moving beyond MARC into a linked data environment.  We will be able to re-use metadata.
Westarted out with a card catalog, written by had, on paper cards.
We moved to MARC inside buildings, but not online (because there wasn't "online" yet)
Now all data can be linked on the web.

VIAF: Virtual International Authority File
It's a mashup of data from 15 International institutions
gives us a view of published aaes
variants of names
send a comment
uses MARC21 and UNIMARC and OCLC
Diane Hillmann says we are moving away from MARC and towards information ecosystems.  MARC is only used in libraries and no one else wants to use it.
Libraries need to be participating in universal descriptive metadata.
Dbpedia:  a community effort to extract structured information from Wikipedia and to make this information available on the Web. DBpedia allows you to ask sophisticated queries against Wikipedia, and to link other data sets on the Web to Wikipedia data
Libraries and our information is not being utilized.
Volunteers helped create the RDA vocabulary.
entity-relationship diagrams


RDF and XML vocabularies are encoding standards.
Semantic Web Community helped with RDA.
NSDL (National Science Digital Library) registry supports metadata interoperability.
Aggregated Statements
The library community can support innovation and change with the rest of the world.
RDA elements and vocabularies will help in the migration from MARC.
MARC is a "lossy output format" and can't convey as rich a description.
metadataregistry.org/rdabrowse.htm


William Moen did an analysis of record use from 2005 to 2007.  
MARC was supposed to be just a transfer format but it turned into a  metadata scheme.


data driven evidence for a core bibliographic metadata record.  We need to express and represent bibliographic data in XML and RDF.
MARC has 200 fields and 800 subfields.
Only 4% of fields account for 80% of all records.
He examined 56 million records from OCLC World Cat.  Only 62 records of non-Library of Congress-cataloged realia out of the 73,000 realia records.
Only 7 MARC tags were used in all records.  There was only ONE occurrence of 656 field out of 56 million records!!!
In an LC created record typically 7 fields and 10 field/subfields combinations were used.
In a non-LC created record, 10 fields and 18 field/subfield combinations were recorded.
Most used fields were:
8, 10, 43, 245, 246, 260, 300, 500, 650, 700.


empirical evidence indicates catalogers utilization of MARC fields is not directly aligned with the fields prescribed by BIBCO, CONSER and National Bibliographic records.
We need to look at what catalogers are doing before we change what we are telling them to do.
Some questions:
What does the low occurrence of fields suggest?
What is needed from FRBR for a bib record?
Should there be higher value data in fewer fields?
Can we argue persuasively for the cost/benefit of exisiting rpactices?
Do we know the extent of and which content designation structures are needed to support a usre atsk?
What constitutes a core record?


www.mcdu.unt.edu/wp-content
Amy Eklund
core elements analysis 3 May 2006