CMI Improvements From Mike Rustici

[Mike Rustici]

Here are the comments I presented on the CMI data model to ADL and
IEEE back in January:

For a better formatted version, see: http://docs.google.com/View?id=dcb7m9mj_118d6sgd7c4


IEEE LTSC CMI Standards Feedback

Mike Rustici
January 8, 2010
1484.11.1 - Data Model for Content Object Communication Comments


Remove the abigiuity on the use / meaning of interactions. In other
words, disambiguate between journaling and state.
Add a notion of attempts. Content should be able to report on the
status achieved on each of the learner's attempts. With monolithic
SCOs, multiple attempts are often delivered within a single "SCORM/
sequencing attempt". An attempt identifier could be quite useful on
interactions as well as part of the disambiguation between journaling
and state.
[Might be a SCORM issue] Separate the notion of persisting data from
the beginning/end of an attempt. This is especially important for
saved state data like suspend data and location.
Add shared state across SCOs (like SSP or ADL shared data stores)
Add the ability for students to take notes and have them stored.
People will use comments from learner for this purpose, but that's not
really what comments from learner is intended for.
Comments From LMS doesn't seem to have much use or implementation.
Interactions need to allow for more description of the question. To do
an effective report on test results. An LMS needs to know what each
possible response was. Currently correct responses and learner
response only include an identifier for the response. A human readable
description is also needed on both of these data model elements.
Furthermore, a SCO should be able to report on what the other
responses were that the used selected from in the event that the list
of possible responses is finite (as in a choice interaction).
People in the simulation community might be able to benefit from more
interaction types. They often have a need to log each of the actions
that a learner takes. I don't know if the current model adequately
meets their needs.
Provide more clarity around the use of interaction objectives. Either
they should be required to tie into an actual objective. Or, they
should be more robust and able to contain a description.
A numeric interactions result is nice in theory, but without context
of the scale of the number it isn't very useful.
Extensions to the data model to allow for more information about the
learner to be sent to the SCO could be useful. I don't know how to do
this in a generic way, but I can see a lot of benefit to being able to
vary content based on demographic information about the learner (job
title, department, etc)
Learner preferences should be more global. Persistent across attempts
and SCOs.
Provide guidance on the reporting of scores. For instance, it might be
nice to know when a score is "final" or "submitted" vs when the score
is just something that is accumulating as the learner progresses
through the SCO. The "test progress" might be different from the
course progress.
Might be nice to see total time broken down by each reported session.

1484.11.2 - ECMAScript Application Programming Interface for Content
to Runtime Services Comments


Create a web services version

1484.11.3: Extensible Markup Language (XML) Schema Binding for Data
Model for Content Object Communication Comments

Need to be able to be run through code generators when used in WSDL on
all major platforms. The current version doesn't work on Axis2 (Java).
Chuck Allen can provide more details of how to fix it. Also doesn't
work in .Net 1.1 (not sure if we know the fix for that).
As part of the LETSI Run-Time Services For Learning project, I created
a simplified version of the XML binding. It is still a work in
progress, but I have attached a list of what I did. Probably have more
official feedback on this to come through LETSI.