MACE CourseID Call August 16th, 2004

[AI] Keith will contact the IMS enterprise specification architect about the model driven architecture.
[AI] Tom will add a course name as an attribute for course in the courseID object structure and clarify references to IMS specifications in the courseID object structure.
[AI] Keith will e-mail Fred regarding Telcert and available IMS information.
[AI] Tom will forward to the list a matrix of standards of technical standards bodies obtained from the IEEE Learning technology standards committee.
[AI] Grace will forward links for MPEG 21.
[AI] Keith and Grace will incorporate discussion items into next version of the CourseID draft.

*Attendees*
Tom Barton, U. Chicago
Grace Agnew, Rutgers
Fred Breshears, UC - Berkeley
Michael Kruckenberg, Tufts
Keith Hazelton, U. Wisconsin – Madison
Scott Cantor, OSU
Jeanette Fielden, Internet2
Renee Frost, Internet2/Michigan
Steve Olshansky, Internet2

*Discussion*

Tom’s latest draft of the courseID object structure can be viewed at:
http://middleware.internet2.edu/courseid/docs/draft-internet2-courseID-eduCourse-01.html . The draft is an UML structured document to sketch a few objects but is not intended as a full characterization of these objects. The intent is to provide a clear enough basis for certain attribute values and identifiers. Over time changes are expected in the associations of the attributes. There are Two attributes: eduCourseoffering and eduCourseMember, which was previously eduCourseRoleType.

For IMS the role type values are the strings themselves and not the valuation of the strings though that was discussed at one point for a possible mapping of values. The original intent of the role type was regarding how registrars’ create identifiers on campus and pass information around. The examples in 2.4 are correct; the examples in 3.2 will be revised. The strings are case sensitive. Sub-roles are limited to 32 characters, which rules out URI’s. We should work with the IMS on the next revision to lengthen the string so a URI is supported. There is a handbook on globally unique identifiers though there is not consensus in the digital library world about globally unique identifiers. Cliff Lynch of CNI has been involved with it.

Colin Smythe, the IMS enterprise specification architect, is working on a model driven architecture to express UML as XMI where UML is serialized, standardized, and XSLT is used to generate WSDL automatically. UML can be used to generate test suites if used correctly.

In the library environment where the start and stop time of access to a resource is important, time elements in the attribute for offering and member of the courseID object structure would be very useful. If definitions for time are being used from the IMS that should be stated explicitly. Co-ordination with the active DRM space would be worthwhile as well. MPEG 21 was published in August of 2003. The rights expression language is XRML, which is heavily patented.

There is an ISO standard for data dictionaries. IMS is working towards a registry of their elements. IMS would like to structure things so that any changes to the registry will ripple out. Since there are more tools for UML than RDF, the desire is to use UML tools that would generate RDF as well.

One idea in the IMS is the notion of the application profile, where you take base specifications that are more abstract and then create an application profile, which is more specific. The basic specification is more abstract so different communities of interest can define the profile they need from it. For example, if eduCourse is for US higher education then it could be considered an application profile for that geographic area.

Telcert contains a section on “what you should do if you think you are doing an application profile.” In their world, you can only do compliance testing against application profiles not against a base specification, because the base spec is more abstract than specific.

Is it possible to create an application profile for learning management systems that would be acceptable to a worldwide community? IMS is in favor of the specifications they’ve created being globally used with certain concerns. They do not want things to be handed off to the standards bodies for development with standards bodies not collaborating with each other and creating incompatible deployments. After study and legal review IMS has opted to reserve the right of first refusal to implement changes to the specification by a standards body. IMS wants to ensure their specifications can be a mirror image of the formal standards. IMS is utilizing caution with standards groups in order to avoid issues that occurred with LOM and metadata. The end goal is to avoid competing differing standards based on IMS specifications.