Building a better LOM? ISO Metadata for Learning Resources

By: Norm Friesen; January 24, 2006

As we've reported earlier, the ISO sub-committee on "Information Technology for Learning, Education and Training" (ISO/IEC JTC1 SC36; http://jtc1sc36.org) is developing a metadata standard for learning resources. (For our earlier report, see: http://www.cancore.ca/elearndublin.html/.)

Now, this work has resulted in a committee draft for a standard: http://mdlet.jtc1sc36.org/doc/SC36_WG4_N0145.pdf.   This short document describes this work, and explains its underlying architecture and rationale.

When Learning Objects and their Metadata were originally conceived (way back in the mid-1990's) the idea was that their popular up-take could revolutionize e-learning and education itself (e.g. Merkow, 2002).  It was envisioned that learning objects, metadata and other standards-based approaches would form an essential part of an infrastructure built on top of the Internet specifically for e-learning.   Some imagined that such an infrastructure and corresponding applications could achieve the kind of popularity and notoriety enjoyed by Napster (hence the brief appearance of terms like "LOMster" and "Learnster"; e.g. UNB, 2001). In the meantime, conceptions of learning objects, their systematic description, and architecture initially had significant impact in white papers and design documents (e.g. Barritt, et al, 1999). In addition, a number of exemplary implementations were undertaken, with many still underway.

However, it seems that these exemplary implementations are most readily found in jurisdictions where there is centralized policy control and large-scale procurement of digital resources. In such highly coordinated implementations, application profiles are frequently created (often at considerable expense) in order to refine the LOM data elements and vocabulary values, and to designate LOM elements according to their functional importance (e.g. as "required," "conditional" or "optional"). Also, some of the more complex aspects of the LOM standard (e.g. classification elements) have been utilized with great precision and effectiveness.  

However, the traction of the LOM and learning objects generally in smaller, localized and less centralized educational contexts has been a very different story. In these settings, comparable but simpler approaches to metadata and resource sharing such as RSS feeds, folksonomies and social bookmarking have arguably have enjoyed more success.

Combined, the successes and challenges faced by the LOM have resulted in apparently contradictory requirements for metadata:  

1. That they present a highly simplified, "kernel" for those wishing to work with only very basic metadata records.
2. That they also accommodate a high level of complexity in terms of element structure and semantics --but that they enable, but not impose, this complexity.

The proposal for an improved LOM described here attempts to address these requirements by presenting 6 core elements. All are top-level, mandatory or required: Instantiation, Description, Creation, Contextualization, Access, and Record. Each of these elements is also an element category, containing further sub-elements. These top-level, mandatory elements can be used and defined in two ways:

• As containers for more specific sub-elements. In this case, the top-level element is defined exclusively in terms of its function as a category. Generally, this definition will begin with the phrasing "This category describes [or identifies] those characteristics of the resource that..."
• As a data element that can be assigned a particular value that describes the learning resource. In this case, the definition of the element is taken directly from the first the first, "leaf" data element beneath it -from the first element occurring after it that asks for a value.

This draft ISO document also addresses the need to provide multilingual capabilties and equivalencies at the most basic levels of its design. It does so, perhaps most prominently, through the function and structure of the "Instantiation" category, which allows for particular linguistic versions (as well as technical formats) of a single resource to be designated as separate "instantiations" of that resource within a single record.

This draft standard document also reflects a number of very specific lessons learned through what has been nearly a decade of LOM implementation. These have been the subject of significant disucssion and study (and are reported, for example, in http://jtc1sc36.org/doc/36N0871.pdf and http://www.cancore.ca/ppt/LOMLessons.ppt ). Among these specific lessons are the following:  

• The complex, hierarchical and iterative data structures of the LOM can be difficult to accommodate in common, relational database technologies.
• vCard semantics and its encoding have presented challenges for implementers and record creators. As one illustration, the example for vCard encoding in the LOM document is incorrect.
• That the elements utilized in implementations do not indicate that the items described are understood as "objects" or software modules.   Instead, these data elements suggest that these "learning objects" are described more generically as "resources." (Elements corresponding to the Dublin Core are widely used; those describing the "aggregation level" "version" and "semantic density" of the resource, for example, are employed very infrequently.)

These issues have all been addressed in the proposed metadata standard:   The standard, in both title and substance, addresses the attributes of educational "resources" generally, not of "objects" or software components; this standard presents a relatively "flattened" data structure; and vCard is omitted, and replaced with a small number of data elements.

For background and context associated with of this draft, please see the documents listed at: http://mdlet.jtc1sc36.org/

Norm Friesen (norm_friesen@sfu.ca) and Pierre-Julien Guay (pjg@ntic.org) are the co-editors of this draft.

References:

Barritt, C., Lewis, C. Weiseler, C. (1999). Cisco Systems Reusable Information Object Strategy: Definition, Creation Overview and Guidelines. http://www.cisco.com/warp/public/779/ibs/solutions/learning/whitepapers/el_cisco_rio.pdf

Merkow, M. S. (2002). Learning Objects Spark an E-learning Revolution. Techlearning. http://www.techlearning.com/db_area/archives/WCE/archives/mmerkow.htm

UNB. (2001). Learnster bound to make a difference. UNB Perspectives 27(12).   http://www.lib.unb.ca/Texts/Perspectives/Mar12-01/learnster.htm