Dedicated to the research, development, implementation, and standardization of metadata for educational and research mathematics.
AMS Panel discussion: Wednesday, January 19. Ballroom Balcony A, Marriott Wardman Park Hotel. 2:15 - 5:15. Immediately followed by American Mathematics Metadata Task Force Meeting.
The Metadata Scene. The final portion of this paper is devoted to reporting on existing standards, on what work has been done, and on what work needs to be done.It will not come as a surprise that a number of separate organizations have developed pedagogic metadata and that they have to some extent taken separate paths.The Dublin Core Metadata Initiative (OCLC, 1999b), which applies to more than pedagogic metadata, held its first workshop in March, 1995, sponsored by the Online Computer Library Center and the National Center for Supercomputing Applications. Subsequent to that the most significant academically oriented organizations proposing metadata standards have been The IEEE Learning Technology Standards Committee (IEEE, 1999), which held its first metadata working group meeting in December, 1996. The Instructional Management Systems project (IMS, 1999), which was started by Educom (now Educause) and held its first metadata meeting in May, 1997. The Alliance of Remote Instructional Authoring and Distribution Networks for Europe (ARIADNE, 1999) which started in January of 1996. Getting Educational Systems Talking Across Leading-Edge Technologies (GESTALT, 1999) which was started in September of 1998. On the industrial side, the most active organization has been the Aviation Industry CBT Committee (AICC, 1999) and on the government side the biggest project with a stake in pedagogical metadata is the Department of Defense and Whitehouse Advanced Distributed Learning network (ADL, 1999). What is perhaps surprising is that there has been significant convergence of standards.As of August, 1999, the IMS project, ARIADNE, ADL (which uses IMS metadata), and AICC (where applicable) are compatible with a single standard known as Learning Object Metadata (LTSC, 1999) produced by the IEEE Learning Technology Task Force.This standard, in turn, is largely compliant with the more general Dublin Core standards. Some Details. This is not the place to treat the structure of Learning Object Metadata in great detail but it seems appropriate to discuss its general outline.This will help elucidate the ability of metadata to support the uses discussed above. Learning Objects Metadata, as defined by the IEEE, is as a tree structure.The â€œrootâ€ of the tree (or better yet, the â€œtrunkâ€)is the resource being described.The next level in the tree consists of nine main branches that are called categories: General.Context-independent features of the resource. Lifecycle. Authorship, ownership, etc. Meta-metadata. Describes what metadata scheme(s) are being used. Technical.Describes the format and the technical requirements needed to use the resource. Educational.Educational and pedagogical features of the resource. Rights. Refers to intellectual property rights. Relation. Describes the relation of the given resource to other resources. Annotation.Allows for comments on the educational use of the resource. Classification.Taxonomic classification of the resource. (Could be subject matter, educational objective, accessibility requirements, etc.) Each category may in turn have its own branches and each branch may branch again, thereby providing increasingly specific information about the resource being described.Figure I shows a part of this tree (going left to right) for a hypothetical document.
The Elements of Learning Object Metadata. Both the nature and the form of the various metadata elements are specified by the Learning Object Metadata standard (LTSC, 1999).For example, the difficulty of a resource is an element (denoted by Educational.Difficulty) whose values can be very low, low, medium, high, and very high.Other attributes are described using unrestricted lists of keywords while others use numerical values.Some (like MIME type) use words with format restrictions and some use lists of keywords that are envisioned as generated by best practices. The most complex and open-ended element is called a taxonomy and goes under the classification category.As in the case of the Library of Congress Classification, a taxonomy can be used to describe the subject matter of a hypermedia resource, but a taxonomy could equally well be used to classify a different attribute such as pedagogic approach. Disciplinary Metadata.Although the existing elements of Learning Object Metadata are powerful enough to describe the subject matter, level, difficulty, intended audience, pedagogical approach, and many other educationally oriented attributes of a resource, it must be recognized that no standard is ever complete or perfect.Learning Object Metadata takes this into consideration by allowing for the judicious extension of the standards as dictated by need.Moreover, a general standard cannot and should not specify anything other than a framework. The details of which lists of keywords or which taxonomies to use, as well as the interpretation of terms like â€œvery difficultâ€, must be answered by specialized communities of users.In the case of pedagogic metadata, these communities are naturally formed by academic disciplines. A number of academic disciplines are in the process of specifying the details of metadata for their own use.For example, the author of this article is chairing a national effort to define and implement pedagogic metadata standards for the discipline of mathematics (Robson, 1999).Each such attempt potentially reveals strengths and flaws in the structure Learning Object Metadata and creates needs for extensions.Each such attempt potentially also adds to the understanding of metadata itself. To illustrate the challenges, consider the problem of defining a subject classification taxonomy (in any discipline) that is of use to naïve and non-expert users.This has proved to be a most challenging task.One reason is that searches should be able to move from a given search term to more general and more specific terms, but the notions of â€œmore generalâ€ and â€œmore specificâ€ can simultaneously refer to natural language interpretations, technical interpretations, and definitional interpretations of the same terms.This makes it very hard (some would argue impossible) to structure taxonomies in traditional ways, and new ways have yet to be tested in the field.
The intent of this article was to give a non-technical overview of metadata and its pedagogical uses.The ones stressed were those involving information retrieval tailored to the user (adaptive searching) and using hypermedia from diverse sources in the same online educational environment (re-usability and interoperability).Metadata has powerful applications in these and other areas of import to online learning. Pedagogic metadata standards are just becoming what might truly be called standards.In the course of implementing these standards, mistakes will undoubtedly be discovered and corrected.This should not, however, serve as a deterrent to including metadata and planning for metadata in pedagogic material developed from this point in time hence.There is great value in doing so