<< previous article | table of contents | next article >>
A Meditation on Metadata
by Steven Forth
Editors’ note: Steven Forth’s primer on metadata--what it is and why to use it, with a focus on online learning--is an excellent introduction to the subject. For more about some metadata initiatives and nonprofit-specific taxonomy projects, be sure to read this issue’s “In Brief.”
You have probably heard the term metadata. You may even have encountered the verb metatagging. In fact, there has been so much noise about metadata and metatagging, it has begun to seem like a dark science, practiced by a mysterious group of people who use an arcane vocabulary for obscure purposes. A look at what metadata is, what it is good for, and how it is being used for digital learning resources should dispel a little the mystery that cloaks metadata.
Defining Metadata
In fact, metadata is a simple thing, and there is a much simpler word that can be substituted: description. Metadata is no more, and no less, than a description of the information or resource to which it is attached. A most common example of metadata is an index card in a card catalog; it gives a book’s title, its author, and its Library of Congress number. Another example is the label on a soup can; it tells you the kind of soup, the ingredients, and the nutritional value. When it comes to digital resources, some of these descriptions need to be readable by the computer, some by the people using them, and some by both.
Why We Need Metadata
Why do we need metadata? There are three main reasons.
- Metadata helps us to find resources. Think of the card catalog again, one of its main uses is to help us find books. The card catalog tells us what books are available on a topic and where to find them.
- Metadata also tells us what a resource can be used for and how to use it, much like the label on a soup can. With a digital learning resource, metadata on how to use it can include instructional issues, such as the learning objectives covered and the kind of learner it is meant for, and technical issues such as what kind of computer, network, and software are required to deliver it.
- Another use of metadata is that it can tell us how resources relate to one and other. Metadata can include information about prerequisites and corequisites and supporting examples and assessments.
Taxonomies and Controlled Vocabularies
When discussing metadata, taxonomies and controlled vocabularies usually also come up. Metadata, practically speaking, is a collection of data fields with rules about what information goes in the fields. The rules about what information that goes in the fields are often based on controlled vocabularies. For example, perhaps only certain people can be authors of learning resources, so the metadata field for “learning resource author” might be limited to a controlled vocabulary of approved author names.
Sometimes cataloging resources can be more complicated, and this is where taxonomies come in. A taxonomy is a hierarchically organized controlled vocabulary. A famous taxonomy is the one used to describe the animal kingdom: A beagle, is a dog, is a canine, is a mammal, and so on. Different taxonomies are used to categorize learning resources by subject matter, by learning style, by learning objectives, even by type of media used.
The Need for Standards
Metadata is used in so many different ways and is so important to effectively searching, out, organizing, and using learning resources that many different approaches have been developed. When the Internet made it possible to share and reuse learning resources widely, people discovered they had often used very different ways of describing their resources, which made them difficult to use, reuse, and share.
The result was a call for standards. The librarians were the first to heed this call, and at a meeting in Dublin, Ohio in 1995, they began to develop a metadata standard for library resources called the Dublin Core (http://dublincore.org/). But the needs of librarians are not the same as those of others, and the educational community soon realized that it needed its own standard.
Work on such a standard has been carried out in many places, but especially important has been the Institute of Electrical and Electronics Engineers Learning Technology Standards Committee Learning Object Metadata, or IEEE LTSC LOM (http://ltsc.ieee.org/wg12/). (LOM can be pronounced to rhyme with Tom.)
The LOM is meant to be all things to all people--well almost. It includes more than 80 different metadata fields that cover everything from all of the Dublin Core to semantic density--a relationship between the amount of content and the amount (e.g., size or duration); high semantic density means that a lot of information is concisely presented--and metametadata--you guessed it, metadata about metadata. The idea is not that every field in the LOM is completed for a single learning resource; organizations are meant to select from the LOM the metadata fields that are most relevant to them and, in some cases, to further constrain the possible values in the field using controlled vocabularies, taxonomies, or other rules. These selections from the LOM, with their additional constraints, rules, and guidelines, are often referred to as metadata application profiles. Most organizations will need their own metadata application profile. This will ensure that learning resources are described consistently and with the vocabulary used by members. Developing these profiles will be an important task for organizations that want to take full advantage of shared resources.
SCORM
SCORM, the shareable content object reference model, also uses a subset of the LOM. SCORM is a way of packaging learning resources so they can be shared across different learning management systems. SCORM is not one standard, but a collection of different standards and specifications together with rules on how these fit together. It uses the LOM as its metadata standard, and the required metadata is a subset of the LOM.
The LOM, and other metadata standards for that matter, are not meant to stand alone. Most of these standards can be written in XML (extensible mark-up language), and XML can combine different metadata standards. One might want to combine the LOM with a metadata standard for a specific field such as healthcare. Most organizations will need their own metadata application profile to ensure that learning resources are described consistently and with the vocabulary used by members. Developing these profiles will be an important task for organizations that want to take full advantage of shared resources.
The Promise of Metadata
The real promise of metadata is its ability to facilitate sharing across organizational boundaries and proprietary systems. As the technology and standards evolve, metadata will play an increasing role, becoming a powerful tool for description not only by the creators of the content, but also the consumers. In the future, learning resources will accumulate metadata as they are used; there will be records of which learning resources have been used together, how effective they have been, and comments from instructors and learners that have used them. Metadata will make it possible for learning resources to serve not only as artifacts of their designers’ original intent but also as living records, as the learning resources carry with them a history of their use.
Steven Forth (steven@recombo.com) is the CEO of Recombo Inc. (http://www.recombo.com/), a content distribution and integration company serving the learning and training industry. He lives in Vancouver where he enjoys describing the many different kinds of rain and the profusion of greens that the rain brings. He is happy to answer your questions about metadata.
<< previous article | table of contents | next article >>
© 2003 Isoph | sophist@isoph.com | subscribe to The Sophist