Keith Hazelton of the University of Wisconsin gave an overview of the MACE CourseID group.
The Major project goals for MACE-CourseID are to:
• Propose standard data element syntax to describe courses and hierarchical components of courses.
• Propose a schema describing courses and course components that:
1. Conforms to IMS standards or requirements for course description,
2. Maps readily from existing applications that utilize course descriptions such as administrative data systems, instructional management systems, etc,
3. Is Shibboleth compliant, to further leverage Shibboleth developments to enable authorization based on course enrollment, though Shibboleth is not intended to take on the functions of a CMS or LMS
4. Is valid for inter-institutional as well as international collaborations

The CourseID WG would name some agent to register as a namespace authority under the MACE URN, requesting that they be assigned the urn namespace urn:mace:courseid
• Institutions would be encouraged to identify courses under their DNS name, e.g. urn:mace:courseid:uchicago.edu…
• Local course offering identifiers could be formed by combining whatever the institution uses as the short name in the timetable of course offerings with some indicator of the particular session in question as well as the primary section, e.g. urn:mace:courseid:uchicago.edu:Physics-101:fall-2004:section-01

Choices ahead on formation of course-offering-section identifiers:
• More prescriptive, standardized vs. more local autonomy, local preferences
– Stipulate ISO start-end dates rather than idiomatic “fall-04”
• More opaque vs. more suggestive components
– :uchicago.edu:35433:A2334:3002-1 vs. earlier example
• More self-contained vs. more reliant on associated metadata
– :uchicago.edu:IPEDS-Physics-sequence-for-majors:first-semester-….section-lead:j-spencer01

Scope of CourseID work

How identifiers are resolved is separate from how they are assigned and resolved. The initial question is how to form the identifiers. The immediate focus of work is on a globally unique course identifier. We want to build on Tom's model and, as feasible, align with the IMS work on the enterprise spec.
The goal is flexible, platform independent, extensible, identifiers that follow a general form that can be read by any campus or vendor that implements the courseID model and that campuses can use as desired and as dictated by inter-campus agreements.

There are at least two general use cases for managing access to a course site:
Granting access to a course site to students, guests, auditor's etc.
Archival course sites. Managing access to a previous course site for students with incomplete's, reuse of a site, reference, etc. A good reference is Cliff Lynch ECAE paper on the afterlife of course sites at http://www.cni.org/staff/clifford_publications.html

Another area for consideration is: What should happen when one university asserts a course id for another? I.e. a student is enrolled in a course at a second campus.

Periodic re-synchronizing of courses with the course catalog is of concern as well.

It was generally agreed that several areas, while of future interest, are beyond the immediate scope of the MACE-CourseID work:
1. The articulation/equivalencies issue will not be addressed at this time, though the future utility of the courseID for such a use is recognized. There is a core competency specification within IMS that people interested in this area immediately can participate in.
2. A way to identify learning resources or objects is also outside the initial scope of MACE courseID work. The CMS is intended to handle fine-grained authentication/authorization issues such as access to tests, syllabi etc. The group is not trying to extend Shibboleth to handle fine-grained authentication / authorization issues for objects. The degree of learning object exposed isn't a Shib issue; it's a CMS issue of "Does the CMS include access control mechanisms that are mature enough to be able to meet the use cases described?"
3. How physical location and time of class are handled is left deliberately vague since there are traditional face-to-face model, hybrid online/face to face and strictly virtual models. Later work or a future group can determine what time and meeting data looks like.
4. The courseID work will not attempt to address business process issues such as how credit is granted, how non-credit, certificate, seminar, training, and professional development classes are tracked.

Discussion
Dirk Herr-Hoyman of the University of Wisconsin is involved with an e-learning systems project using the IMS enterprise specification to bring information from the PeopleSoft SIS into e-learning systems. They had to go through exercise of defining use cases for mapping what a course offering is and what the associated vocabulary for that is. There is a need to be cautious with assumptions about meeting time and physical location since models of what defines a class/session are changing. Tom Barton's eduCourse model appears to align with that. There will also be a need to deal with roles at some point.

Fred Beshears of UC – Berkeley explained that IMS uses the notion of a "group ID", which is abstracted away from course (it is not a course id) to try and bridge "training apps" and "college course." They want to extend to course catalogs in future versions of the IMS Enterprise specs. They tend to relate to "lecture sections" not courses. The IMS enterprise spec will get into how to define data exchange standards, API's, etc. for course catalogs.

IMS is looking more to UML and modeling architecture
Fred could be liaison to the IMS enterprise group for Internet2.

Another general case to consider is: What should happen when one university asserts a course id for another? I.e. a student is enrolled in a course at a second campus. If a student at Brown is enrolled in a class at Chicago what is the best practice recommendation for how to represent that? Brown has arrangement with RISD (Rhode Island School of Design) for Brown student to take Brown courses for credit delivered from RISD. The Brown student will need access to RISD resources for the duration of the course.

Issues:
How does student get access to materials while course is on?
How is credit determined afterwards? This can be informal or formalized. For the Brown example there is a formal agreement. For Fred's Simple Use Case (http://middleware.internet2.edu/ under Draft Documents), credit is given to the UCB students by the UCB instructor and to the UCLA students by the UCLA instructor to eliminate registrar issues with credit and cross-enrollment. The case is a collaborative teaching one versus a cross-enrollment one.

Since ownership in the URN is pegged to the section id who can accept assertions about ownership? Should this be incorporated into the object model?

Keith raised the issue of the URN being meaningful v. opaque. One possibility is allowing the institution to choose degree of opacity. Institutions can then agree between themselves on how to reveal the identifier.

Is the important property of the courseID is that it's unique? How to assign identifiers, and how to resolve them online is a bigger task that may be independent of unique question. Another tension is mandating how to form these ids versus letting local practice rule.

There is interest in concrete examples that can be used as an education tool for system administrators and other key parties on campuses. Examples that could be developed might include:
1. The Penn State project Using Shibbolized version of WebAssign. They want to take it to next level, and send one or more course identifiers from Penn State to WebAssign to match against existing courses in the WebAssign system.
2. U. Texas hopes to go live in January with a Blackboard project. The School pf Public Health in Houston teaches a course that includes students from several area campuses. The want the campuses to be able to assert membership in the Houston based course and Blackboard automatically create user objects within the Blackboard system the first time these people arrive.

Mike Grady stated that Illinois is going to replace their system and are looking at how to identify courses in their new system. They have a concern over continuing education courses such as those for veterinarian's that currently don't get a traditional course id. It is credit course but not degree credit, I.e. credit is given at the state licensing level but not institutional credit. Brendan Bellina has a similar issue at Notre Dame. The classes may count for degrees and certificates but not for Notre Dame degrees.

Outside the U.S. the UK and Australia are also active in this space.