Re: [dita-learningspec] Design of Learning Map Topicrefs: Why NoSubordinate Refs?

[john_hunt@us.ibm.com]

Hi Eliot,

A few responses back to your comments.

>>Allowing learningContentRef to nest recursively
I think satisfies my
requirement to be able to compose learning objects from subordinate topics.
<<

JH: Excellent!

>>I think that all the currently leaf-level topicrefs should allow
nested
learningContentRef so that, for example, I can have a pre- or
post-assessment that is composed of subordinate topics. Not sure how that
plays with SCORM but it certainly makes sense for printed tests, where
groups of questions can be organized into arbitrary hierarchies.
<<

JH: Are you suggesting the same model use learningContentRef?


That is:

a) set default of chunk="to-content" and
 
b) allow zero or more child elements of same specialize ref type
 
So, for example, for learningPreAssessmentRef:
       <!ELEMENT learningPreAssessmentRef    
        ((%topicmeta;)?, (%learningPreAssessmentRef;)*)
>

and so on?

And after a wee bit more thought during the day about
SCORM processing.

a) if chunk="to-content", then we get the
SCORM leaf nodes into the chunked topic
b) if no chunk="to-content", then we need
to get SCORM leaf nodes, so I suggest we simply flatten all of the child
topicrefs and make them siblings of the parent in the SCO. Open to other
thoughts.

>>And a reminder that DITA 1.2 constraint modules can re-impose constraints
that are more relaxed in the base module, so that you can provide a default
or example constraint module that reflects the original design.
<<

Thanks for this reminder! (And you very kind to call
it a reminder, too.)

This sounds like an excellent approach. Perhaps you
can help with such sample module.

Thanks for the good input.

Looking forward to comments from others.

___________________________________
John Hunt
Structured Content Architect / Lotus Information Development Center
IBM Software Group/Lotus Software
phone: 617.245.8053; t/l 268.9401
john_hunt@us.ibm.com