Learn more about the metadata crosswalk repository

A crosswalk is used for translating between metadata formats, such as this example created by the Library of Congress. Many institutions have developed bibliographies about crosswalks, which contain links to crosswalks, executable code, and descriptions of metadata standards, as well as more general and theoretical discussion. These bibliographies are a good starting place for inquiry, but most of them can't be searched. Nor can they be bound into software applications.

What we are presenting here is a prototype for a repository of metadata crosswalks, called SchemaTrans. The repository improves the usability of the bibliographies by collecting all information required to document and execute a collection of crosswalks.

Crosswalks in the SchemaTrans repository are initially modeled using METS, a standard for encoding metadata about objects in digital libraries. The resulting records are deposited into an OAI repository that can be searched by human or machine using the SRW/U protocol. SchemaTrans translates XML documents from the specified source to the target by matching the dependencies in the documents against the records in the METS database.

Try it out

The demo is accessible here. For demonstration purposes, we have included a pointer to a collection of MARC records in the URL box.

  • Pressing Submit results in a page with several options. The drop-down menu on this page is an automatically generated list of options for translating the data. Since the sample input data is MARC, the menu lists only the crosswalks from the repository that can be applied to MARC records.
  • Selecting one of the options and pressing Execute Crosswalk translates the data.
  • Pressing the Crosswalk Metadata button retrieves the METS record for the crosswalk. It contains pointers to the three pieces of information required for a formally precise definition of crosswalks: the table of equivalences, the source metadata schema, and the target metadata schema, each of which may have machine-processable and human-readable files.

The SchemaTrans demo and the METS data model are described in more detail in "A Repository of Metadata Crosswalks" ( D-Lib Magazine, December 2004).

For a good introduction to and overview of SRW/U, see An Introduction to the Search/Retrieve URL Service (SRU), by Eric Lease Morgan, in Ariadne 40 (July 2004).