A Metadata Crosswalk, or a crosswalk for short, is a way of bridging two different metadata schemes. Another way of looking at it is that a crosswalk is used to transform one metadata scheme to another.
OCLC Crosswalks
There is significant work that OCLC Online Computer Library Center Research is doing with crosswalk repositories and software that performs metadata schema transformation.
A paper describing the beginnings of this work, "Two Paths to Interoperable Metadata," is attached. A brief summary of the paper is below.
The goal of this work is to develop software that performs metadata schema transformations.
Design Goals
- develop a self-contained metadata translation service
- clean separation of the document data model, the schema translations, and the machinery of the application
- automation of routine processes and a well-defined place for human input
- support for current metadata creation practice and for foreseeable innovation
The Design
- Creating a "loop" of transformations. The beginning of the loop is getting input in some format, the end is the output in a different format.
- A shortcut from the beginning to the end of the loop can be taken by using a simple XSLT transform between the two formats. For some, simpler, transformations, this may work. But, XSLT is a structural, syntactical transform only. It does not handle the semantics of the structures it is dealing with. Is the description the same as an abstract? Using XSLT only also limits in that formats cannot be chained together - you lose a significant amount of meaning when going from MARC to DC to LOM to GEM in a chain.
- By going through the loop, some structural transforms are taken as well as semantic transforms within the 'interoperable core'.
- The interoperable core is a set of "semantic maps" that acts as a middle-man for all external formats to map to. Once an external format, such as MARCXML, is mapped to the interoperable core, the data can be output in any form compatible with the interoperable core [img, p.4].