Learning Object Standards
There is a veritable alphabet soup of specifications, standards and profiles related to learning objects. In the previous column, we discussed in passing the Sharable Content Object Reference Model (SCORM), an application profile designed by Advanced Distributed Learning (ADL) for use by the U.S. military and related learning providers. Other well known acronyms include the Instructional Management System (IMS) specification, the IEEE â Learning Object Metadata (LOM) standard, and the CanCore application profile.
The purpose of these standards is to fulfill the promise of learning objects. As Friesen (2001) and others note, learning objects are supposed to be modular, interoperable, and discoverable. The idea is that learning object standards identify the properties, and in particular, the pedagogical properties, of learning objects, much in the way the label on a soup can identifies the contents, and in particular the nutritional contents, of the can.
Not all standards initiatives are the same. Some initiatives, referred to as specifications, are intended to capture a rough consensus among practitioners. They are descriptive of current practice in the field, often incomplete, and often the work of an ad hoc consortium. (Lim 2001) By contrast, standards are regulatory principles formally endorsed by a standards body such as IEEE or the World Wide Web Consortium (W3C). Once a standard is created, it may be customized for a particular use. This customized schema is called an application profile (Heery and Patel, 2000). Application profiles typically specify permitted values and they can refine standard definitions.
The only standard for learning object metadata is the IEEE Learning Object Metadata standard. Though variations continue to exists, other metadata standards initiatives are beginning to coalesce around the IEEE standard. The final draft is available at the IEEE website (the formal standard is subject to copyright restrictions and might no be available for general viewing without charge).
The IEEE-LOM describes learning object metadata in nine different categories:
The learning object metadata is expected to be encoded in XML, according to the schema provided by IEEE, and contained in a single file. This file may then be attached to the learning object or made available as a stand-along file for the purpose of locating and retrieving learning objects. Groups of learning objects may be bundled together; this bundle is described using the IMS content packaging specification.
The major criticism of the IEEE-LOM is that the standard contains too much metadata. There are some 67 elements, many of which take some time and consideration to complete (what is, for example, the correct classification for a picture of an oak tree). As Hatala and Richards comment, "the business and educational communities have been slow to adopt the full IMS standard mainly due to the high number of the fields and vagueness with which the values for these fields have been defined." (Garrido, 2003) Filling the many fields would in many cases take longer than the creation of the object itself.
IEEE-LOM also overlaps with many other standards. The clearest case of this is the overlap between LOM and the Dublin Core metadata initiative. This has led Nilsson to use Dublin Core, rather than IEEE-LOM, metadata for many elements. (Nilsson, 2003) Additionally, IEEE-LOM provides only partial information which could be provided in more detail by other standards. If the object is an image, for example, the IEEE metadata does not even indicate whether the picture is colour or black and white, and while the size (in bytes) is indicated, the dimension (in pixels) is not. Something like the Digital Still Image Metadata (DSIM), released in draft form by the National Information Standards Organization, might be much more appropriate.
Finally, what happens when different people have different points of view about what a learning object's metadata might be? While the author may classify a picture of an oak tree under âtree,' a more precise designer may prefer âhardwood' while yet another may prefer âflora and fauna of England.' The more subjective the metadata â and in IEEE-LOM, there are many areas of subjective assessment, including educational metadata, annotation and classification, the more points of view may be offered.
Though it has been adopted as a standard, IEEE Learning Object Metadata will evolve. It has to evolve. It is based on the wrong set of assumptions about both metadata in general and about learning objects in particular.
The use of metadata in general to describe any sort of object can and will be over time a case of mixing and matching appropriate metadata. Thus, an image, for example, will be described using image-specific elements from DSIM, publication data elements from Dublin Core, Rights information using the Open Digital Rights Language (ODRL), and pedagogical elements from IEEE-LOM or its successor. It makes no sense for metadata in one domain to overlap into other domains, especially when such overlap creates vagueness and cross-categorization.
The description of learning objects in particular, meanwhile, will over time come to be seen not as reflecting the inherent properties of an object, but rather, how that object has been used in education. As argued last month, this is the only way to describe and create learning objects in a way that preserves their pedagogical value without impacting their reusability.
What is important here is that any given object may have many uses. A picture of an oak tree may be used as a specimen in a botany course, as an identification aid in a carpentry course, and as illustration in a geography course. Thus, if we are describing the educational use of this object, we obtain three separate descriptions. Contemporary learning object metadata doesn't really envision this possibility, and it doesn't really address the question of how metadata from different sources can be combined to form a single description of the object.
IEEE. 2002. Draft Standard for Learning Object Metadata
Friesen, Norm. (2001) What are Educational Objects? Interactive Learning Environments, Vol. 9, No. 3, Dec. 2001.
Garrido, Javier Sarsa. (2003) Two Scenarios Using Metadata. Learning Technology: IEEE-LTTF.
National Information Standards Organization. (2002) NISO Releases Digital Still Image Metadata Draft Standard.
Nilsson, Mikael. (2003) Semantic issues with the LOM RDF binding. Centre for User Oriented Design.
SUBSCRIBE TO OLDAILY DONATE TO DOWNES.CA
Web - Today's OLDaily
Web - This Week's OLWeekly
Email - Subscribe
RSS - Individual Posts
RSS - Combined version
JSON - OLDaily
National Research Council Canada
All My Articles
About Stephen Downes
About Stephen's Web
Subscribe to Newsletters
Privacy and Security Policy
Stephen's Web and OLDaily
Half an Hour Blog
Google Plus Page
Huffington Post Blog