Campbell - Enhanced Tracking and Shared Data

Learner Interactions - I agree, SCORM has it backwards when it comes to data persistence. SCORM currently defaults to resetting data instead of persisting it. I can see no reason why. It is not what it expects and adds no value. The only argument that I have heard in favor of resetting data is that storing data imposes an undue burden on LMS's. I just can't buy that argument in an era where I can buy a terabyte for less than $200.

The issue of a smallest permitted maximum is also interesting to consider (the 250 interaction limit mentioned in the paper). From a content developer's perspective, I can see why this would be frustrating, but I also haven't seen any LMS that caps things right at the SPM (correct me if I'm wrong). It's a hard problem, you can't say that an LMS must be able to support infinite interactions and you can't say an LMS is conformant if it only accepts say 2 interactions...so how do you draw the line? I think the concept of an SPM is a good compromise.

The need to share data between SCOs is also quite real for reasons John and Don mention, although peer to peer communications is probably a lot tougher nut to crack than the rest of the stated reasons. 

Two very key points made in this insightful paper were:

(1) "interactions are not guaranteed to be persistent beyond the life-cycle of a SCO" and

(2) "however, in order to be truly interoperable, a very well-defined, consistent approach to data persistence is needed"

I would like to begin some discussion about this paper, by building on Mike's comments above, reiterating the idea that sharing data between SCOs is a real and valid problem for SCORM 2.0 to address and just might enable developers to create better eLearning (since interoperable sharing of data between learning objects COULD produce more engaging and maybe more effective learning).

Introduction of Global Objectives in existing SCORM reflects a natural desire of developers to have a Learner Profile - LMSdata- sharable by different SCO. That was a first step. Now it is already clear that it is not enough to support decision making on all levels of the Activity Tree. To be useful the Learner Profile should be defined on all levels of Activity tree, not just on its root level. 
Instead of one set of Global Objectives defined for all activities (each having its own local objectives), I would propose to consider local objectives of each parent activity as global objectives of its child. This provides a simple and clear pattern, which can be systematically applied across the entire tree. But definition of activity tree, which now is package oriented, should be extended as well to more open and regular construct from such simple and clear patterns.

This paper did a great job of bringing to light one of our two main problems with SCORM development.  Our lab works with curriculum development in higher education, and we need the ability not only to make learner interaction data accessible to other SCOs in the course (a progress report or map through the content), but also to make it accessible outside of a learner interaction (report scores to an instructor).

As Daniel wrote "Our lab works with curriculum development in higher education, and we need the ability not only to make learner interaction data accessible to other SCOs in the course (a progress report or map through the content), but also to make it accessible outside of a learner interaction (report scores to an instructor)." --

I have heard a large number of times from SCORM developers, that without this capability in SCORM, developers will simply point to external "non-standard" tracking files, which will help accomplish this same feat - inter-SCO communication and SCO to external reporting. Use case I hear the most is for interactive gaming efforts tracked via an LMS.