• 01 - Introduction
•  
• A Look at Sharing
  • 02 - The Current Sharing Environment
  • 03 - Aggregators
  •  
• Defining Shareable Metadata
  • 04 - Foundations
  • 05 - Content
  • 06 - Coherence
  • 07 - Context
  • 08 - Activity 1
  • 09 - Communication
  • 10 - Consistency
  • 11 - Conformance to Standards
  • 12 - Activity 2
  •  
• Shareable Metadata in Practice
  • 13 - Crosswalking
  • 14 - Workflow
  •  
• 15 - Looking Ahead

Metadata for You & Me - Shareable Metadata in Practice: Crosswalking

Module Content

• Screencast
• Powerpoint Slides and Other Resources
• Module Text

Screencast

---

Powerpoint Slides and Other Resources

• Download PowerPoint
• XSLT
• Emory Metadata Migrator
• Godby, Carol Jean, Jeffrey A. Young, and Eric Childress. (December 2004) "A Repository of Metadata Crosswalks" D-Lib Magazine 10, no. 12.
• DLF/NSDL OAI Best Practices, Crosswalking Logic
• Introduction to Metadata: Pathways to Digital Information

---

Module Text

1. General advice on metadata mapping

The process of using a metadata record in one format to generate another in a second format is known as mapping or crosswalking. When sharing metadata, the record you share is likely to be a second copy of the metadata record, different in some ways than the metadata record you use to drive local discovery and display. This second shared record can be pre-generated or derived on the fly, depending on the system you're using for sharing. Again, it's not practical to create a shared record tailored towards every single aggregation. By following the principles outlined in this course, you should be able to create shared records that work for many instances of a general type of aggregation.

Some general principles apply when mapping metadata. These include:

• Mapping metadata from more robust formats to simpler ones is most effective; mapping from simpler to more robust formats generally results in inadequate records.
• Plan for both mapping values between fields and for transforming data values themselves. The former is fairly easy; the latter is where the real challenge lies.
• Repeat elements when your target metadata schema allows it.
• A standard crosswalk might exist for mapping between any two standard metadata formats. You will likely need to make some slight modifications to the standard crosswalk to accommodate local practice or to add in "boilerplate" elements common to all records in a given set.

2. Mapping technologies

XSLT (eXtensible Stylesheet Language Transformations) is the technology most commonly used for metadata mapping. XSLT is an XML language, written specifically for transforming of XML documents into other types of documents (including other XML documents). When the source XML document isn't overly large, XSLT transformations can be done on the fly, although many systems will use XSLT to pre-generate output documents.

Other technologies can be used to transform metadata records as well. Scripting languages with strengths in text processing, such as Perl and Python, are often used by programmers to perform metadata mapping tasks. Increasingly, other options are emerging requiring less technical skill, such as the Metadata Migrator from Emory University. Shared technical infrastructure for metadata mapping is also emerging, with machine-readable repositories of crosswalks such as that hosted at OCLC.

3. Developing a crosswalk

Developing crosswalks can be a time-consuming task, but you'll get better at it with practice. Here are a few tips to help you get started:

• Start with a conceptual mapping, a human-readable table matching elements in the source format to elements in the target format. Record notes about additional processing required beyond simply copying the value from one field to another.
• After the conceptual mapping is complete, you can begin implementing it in whatever transformation technology you've chosen. Like any kind of programming, don't try to do it all at once. Solve one small problem at a time, test to make sure it works the way you expected, then move on to the next one.
• Pay special attention to the order in which elements appear in the output format. The desired order is often different in the target format than in the source.
• Use a small sample of records to prepare the crosswalk, then spot-check the results with other records from your source set. Try to test with records using all possible fields in both the source and target formats before you consider the crosswalk complete.

There are some more useful hints specific to developing crosswalks for metadata to be shared via the OAI protocol in the Crosswalking Logic page of the DLF/NSDL OAI Best Practices. Some sample crosswalks are available in the online resource Introduction to Metadata: Pathways to Digital Information.

Next module: Shareable Metadata in Practice: Workflow

Metadata for You & Me is a collaboration between the University of Illinois Library and Indiana University.
This project is made possible by a grant from the U.S. Institute of Museum and Library Services (IMLS) through the Laura Bush 21st Century Librarian Program.