Writing in the Sand Metadata Object Description Schema (MODS) in Digital Libraries

The Metadata Object Description Schema was developed as a descriptive metadata scheme oriented toward digital objects, and drawing from the MAchine Readable Cataloging (MARC 21) Format. Most catalogers are already familiar with MARC, as that is the standard library cataloging format in the United States. By regrouping the fields, adding a few new ones, and translating the numeric codes to readable English in XML, the scheme becomes a reasonably usable, and fairly refined, descriptive metadata for digital objects.

MODS is under intense development right now; version 3.1 was released 27 July 2005, and version 3.2 was released 1 June 2006. In addition, MODS has been adopted by the Digital Library Federation (DLF) for their Aquifer Project which is seeking to develop the best possible methods and services for federated access to digital resources. You may search their test portal here. Their recommendations and profile for MODS are still under development. DLF intends for MODS to replace Dublin Core for descriptive metadata for digital objects in the digital library world, for MODS allows more specification of contents and better clarification of the various elements than does Dublin Core. And by promoting cross-institutional agreement to a specified profile (subset; more on this below), the metadata can be functionally usable for federated search capabilities and software.

An example MODS file can be seen here and its display online via software can be seen here. This sample is still in development, so the URL or identifier in the record may change, but the digital item referred to is currently viewable here. ( This display is offered on a development server. When the collection is ready to go live, I will assign a "persistent" URL to each object's primary display for reference in all other records. Until then, the identifier/URL in these records is a dummy. )

As you can see from the example, there are attributes for authority control used, and attributes for format used, and attrbutes (as well as elements) that clarify what the contents of the field are. All of these are huge improvements over Dublin Core. While in Dublin Core the use of the Date field is ambiguous, here it is clarified, allowing multiple dates with enough information to make them useful, and a specified machine-readable encoding (yyyy-mm-dd):

<mods:originInfo>
<mods:dateCreated keyDate="yes" encoding="w3cdtf">1862-02-21<mods:dateCreated>
</mods:originInfo>
<mods:recordInfo>
<mods:recordCreationDate encoding="w3cdtf">2006-06-12</mods:recordCreationDate>
</mods:recordInfo>

In addition, other date values are allowed, such as birth and death dates of the author/creator. As you can see, the cataloging can be much more extensive and specific using MODS. Using the "relatedItem" element, which allows all other elements within it, a MODS record can become recursive, and can be used to describe very complex objects. Because MODS is a fairly complex scheme, you will need to adopt a "profile" (a local subset of acceptable elements, attributes, authority lists, formats, usage, and ordering of elements from the schema). Software can only be written for specified sets of tags and ordering of tags, and the simpler, the better. By choosing a profile that works for your purposes, you ensure consistency within your institution. By publishing this profile online, you enable others to adapt their software to use your records, and you also enable others to follow your lead. The University of Tennessee Digital Library Center MODS profile is still in development, and will be available online when complete.

Girl writing in the sand

This information is provided without guarantees as to validity or completeness, particularly in light of the fact that the world of metadata in digital libraries is a world of shifting sands, constantly changing.