Summit on Serials in the Digital Environment

Recent Developments in MARC with XML

Prepared by Sally McCallum

The MARC formats are the foundation of library automation and they continue to serve the community well for communication of bibliographic data. However, XML seems to be a computer format development with staying power and new programming possibilities. Therefore the Network Development and MARC Standards Office, which is responsible for development and maintenance of MARC 21 in collaboration with the library community, has taken several steps that will assist libraries in using the XML environment and perhaps eventually migrating to it. With a billion or so MARC records and thousands of MARC 21 based systems in institutions around the world, it was recognized that a path was needed that allowed experimentation and flexible movement of data.

MARCXML

MARCXML is an XML schema that allows the MARC record to be encoded into XML. It is a simple schema that encodes the MARC 21 tags and indicators as attributes of simple XML tags like "datafield". It provides for lossless conversion from MARC 21 to MARCXML, and near lossless from MARCXML to MARC 21 (the exception being that a record that originated in MARC 21 and contains non-roman data may not have the same coding for the "ASCII clone" characters when it returns but uses comparable encoding from the ASCII set). LC has made available from the MARC web site transformations that convert the MARC 21 record to MARCXML and vice versa, including the conversion of the character set in and out of Unicode (XML requires Unicode). Serial records and Holdings records in MARC 21 can thus be converted to and from MARCXML. LC also provides other transformations of MARCXML data in its MARC Toolkit. (LC originally experimented with a different more detailed tagging structure in a SGML DTD. It is still maintained on the MARC web site as an XML DTD, but MARCXML is recommended for current use.)

In recent years MARC has incorporated CONSER-stimulated changes to accommodate continuing and integrating resources. These changes, which had far reaching impacts on many systems, are just now being implemented by applications. As noted above, MARCXML supports all of the detail in MARC 21 including these new serial concepts.

MODS

The Metadata Object Description Schema (MODS) is a MARC compatible XML schema that accommodates bibliographic records including serials. It is maintained at LC by the community of users via a listserv. It uses word tags which are more user friendly for certain applications. This schema provides a simpler record, but does not support some of the detail accommodated in MARC 21, although it does have well developed tagging for recording citations to journal articles and it accommodates all the important serial linking entry fields. The schema is organized differently from MARC 21 in some areas. Transformations to and from MARCXML are provided on the web site which also enable taking records to and from MARC 21. It is especially useful for electronic resources and in instances where records are needed for transfer via XML protocols such as SRW, METS, or OAI. A corresponding authority format is also under development.

Web sites:

www.loc.gov/marc/
www.loc.gov/marcxml/
www.loc.gov/mods/