Repesenting Learning Object Behavior

Mark J. Norton
Phillip Dodds

May 3, 2001

Given that the IMS Content Packaging spec provides a means to transport content and other data from one learning envionrment to another, how will the target LMS know what the behavior of this aggregated content will be? How are behaviors represented and transferred between environments?

Traditionally, this was handled by separating the content from the player, and including some indicator in the content a reference to which player should be used. There is a lot to be said for this mechanism, since content delivery software can be complex. Perhaps we should re-think that. If the code wasn't too large, it could be packaged in with the content. This increases the likelihood that a collection of content objects can be properly delivered from its new environment.

Behavior (again, traditionally) is captured directly in software. It was the easiest way to do it. It is not, however, the only way to do it. It may be possible to describe learning behaviors in such a manner that the behaviors could be separated from the engine that implements them. We could create a behavior representation document. Examples of this already exist (Cascading Style Sheets, etc.).

Delivery behavior of aggregated learning objects falls into a few categories: presentation, interaction, and navigation. Presentation is a description of how the learning object is shown to the learner. It is not necessarily the same for all users, allowing for accessibility and personal preferences. Interaction is the function of the learning object itself. Many are passive, but others require inputs and response from the learner. Navigation is a system level behavior. It provides the means to traverse from one learning object to the next either manually or in some automated way. Each of these behaviors is different from the others and needs to be considered separately.


Behavior Representation

Representing presentation and inteaction behavior has long been the goal of traditional CBT standards (AICC, etc.). Presentation and layout representation can be handled by a number of well defined mechanisms including various templating mechanisms, HTML, and DHTML. Representing user interaction has been done by decomposing interaction into a set of well understood gadgets and controls, flagging the content elements as to the kind of interaction desired, and leaving the actual execution to a run time user interface (such as a web browser). Some of these representation systems are very flexible and extensible.

This leaves navigational behavior remaining to represented. The papers written by Mark Norton on Learning Models and Learning Structures can serve as a starting point for codifying navigation models. Graph theory forms the basis of this approach. There are specific ways to represent the ordering and organization of learning objects using directed graphs restricted in various ways (looping, branching limits, etc.). Graph representations cover almost all possible methods of navigation. The sole exception is random presentation, which is limited in its usefulness.

Taken together, the representation of presentation, interaction, and navigation behaviors creates a means to represent pedagogy in technology based learning system. A full representation of a pedagogy may require addition information such as learning objectives.