The AMMTF is developing metadata based on IEEE Learning Object Metadata (LOM) and embodied in IMS metadata. The latest releases of these documents are LOM Working Document 3 and IMS Specifications Version 1. LOM/IMS metadata is structured metadata. Please refer to the IMS Best Practices document, A Mathematician's Guide to IMS Metadata, and to a paper on Pedagogic Metadata for further explanation. The LOM/IMS specifications consist of a base scheme that defines a large but general set of elements that can be used to describe a learning object. In some cases, the vocabulary allowed is restricted; in other cases, the vocabulary refers to a "best practice". The base scheme can also be extended to accommodate new elements.

Although LOM/IMS give criteria for testing conformance to their standards, they do not prescribe specific ways of encoding metadata records. These are called bindings. The IMS Specification gives an XML binding and it is anticipated that RDF bindings will appear in the future.

The work of AMMTF can be categorized as follows:

1. Defining or identifying "best practice" lists to use for existing LOM/IMS elements.
2. Possibly defining extensions needed for mathematical resources. At this point in time the route being taken is to provide input to the organizations developing the standards since they are seeking feedback from communities of practice such as ours.
3. Creating taxonomies (subject classifications) for pre-professional mathematics. This only fills one slot in LOM/IMS metadata but may well require more effort and maintenance than all the other slots combined.
4. Assisting in the concrete implementation and application of 1 - 3.
5. Putting into place mechanisms for maintaining, publicizing, and updating the work done in 1 - 3.

Much of this work is just starting. The following outlines what has been done so far.

The AMMTF started with the concrete case of records in the SMETE digital library being developed at Berkeley. The SMETE project stores many of the LOM elements and has a Web-based interface for creating records. This interface was originally designed for engineering, so the AMMTF has identified extensions or changes needed to accommodate mathematics. These, as well as suggestions forwarded to the IEEE Learning Technology Task Force, are documented in the documents section of this site.

The AMMTF has divided mathematics into three levels. Level I can be described as pre-variable mathematics and roughly encompasses subjects taught in grades K - 8 in the American school system. Level II might be described as pre-proof-oriented mathematics and encompasses subjects taught in high school and the first two years of college. Level III is professional mathematics, including that taught to undergraduate mathematics majors.

One reason for creating this trichotomy is to accommodate differences in meanings for the same terms. Another reason is that the pedagogic foci of these three levels are different. We feel that the division into three levels is the minimal division that will work.

Level III will be handled by MSC, which has the simple structure of a depth-two tree. The structures proposed for Levels I and II are more complex and are complicated by the fact that they must accommodate not only the internal logical structure of mathematical subjects but also treat semantics derived from cultural and natural language sources. These taxonomies consist of of terms (controlled vocabulary) and relations among the terms. The relations are broadening, narrowing, similarity, and attribute. There is no transitivity, similarity is one-sided (not symmetric), and attributes are used as a means to limit the size and granularity of the taxonomies while sill allowing some highly constraining search strings to be constructed. Details are given in the document Taxonomies for School and College Mathematics. Examples can be seen in the Level I and Level II draft taxonomies.

The AMMTF includes the directors of the Math Forum, ENC, and SMETE. As a result of the AMMTF work, records from the Math Forum and ENC have been successfully transferred to SMETE. The next cooperative project will move towards federated searching.