dita-learningspec message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]
Subject: Re: [dita-learningspec] Implications of Subcomponents of lcBase Not BeingTopics
- From: john_hunt@us.ibm.com
- To: Eliot Kimber <ekimber@reallysi.com>
- Date: Thu, 13 Mar 2008 14:52:04 -0400
Elliot,
It's so good to have you in these discussions!
>>Has
anyone else thought about the question-level re-use use case where
each question should be (or may need to be) managed independently of all
other questions?
JH: Yes, this notion of a question bank came up during
our discussions. And the
samples Bill Blackmon (cc'd here) shared use a question-bank
approach.
>>The immediate challenge I see is that if questions
are not topics in
their own right there is no good DITA-conforming way to manage
individual questions as individual storage objects--they have to be
stored within a topic even if the intent is to use conref to get them.
<<
>>This suggests that in my case the only solution would be to create
a
topic type whose only purpose is to hold a single question. That doesn't
seem like a necessarily bad thing, if for no other reason than that it
provides a way to bind a title and specialized metadata to the question.
<<
JH: Of course - the only thing available for holding
DITA content is a topic
(unless you consider using a map, that doesn't seem
appropriate).
However, with the learning specializations, we defined
the questions interactions
as a DITA domain specializatoin, thus making them
available for use in other topic
types, including generic topic, by including the learningDomain.
And each of the
7 interactions are specialized from topic/fig, making
them available pretty much
anywhere in a topic body.
I just now quickly added the learningDomain to DITA
topic, and can now get a create compact little
DITA generic topic with a simple open question like
this:
<topic id="asdf">
<title></title>
<body>
<lcOpenQuestion id="asdf">
<lcQuestion>What time is it now?</lcQuestion>
</lcOpenQuestion>
</body>
</topic>
Or this true-false question:
<topic id="asdf">
<title></title>
<body>
<lcTrueFalse id="onetf">
<lcQuestion>The earth is flat.</lcQuestion>
<lcAnswerOptionGroup>
<lcAnswerOption>
<lcAnswerContent>True</lcAnswerContent>
</lcAnswerOption>
<lcAnswerOption>
<lcAnswerContent>False</lcAnswerContent>
<lcCorrectResponse/>
</lcAnswerOption>
</lcAnswerOptionGroup>
</lcTrueFalse>
</body>
</topic>
And, as you way, you could include metadata, here,
too, including I might add, lcLom metadata
defined in the DITA learning specialization, if you
add the learning metadata domain to topic, as well.
Like you, I'm not sure what to put in the title.
Finally, in putting together some processing samples,
I'd also like to point out that the
learning map domain is quite flexible in that it allows
for the use of *any* topic type in
any of the specialized learning topic references.
(It does place strict structural restrictions
on the learningGroup and learningObject, and recommends
using the specialized learning types
for the content, but does not require use of those
types.)
So, for example, the above interactions from these
generic DITA topics could be included as a preassessment
or postassessment in a learning group or learning
object - see the map below - those assessment topics
could be one of the above generic topics above.
<learningGroup
navtitle="Group 1">
<learningPreAssessmentRef href=""g1-preassess.dita""
/>
<learningOverviewRef href=""g1-overview.dita""
/>
<learningObject navtitle="g1-lo1-ref">
<learningPreAssessmentRef
href=""g1-lo1-pre.dita"" />
<learningOverviewRef
href=""g1-lo1-over.dita"" />
<learningContentRef
href=""g1-lo1-content.dita"" />
<learningSummaryRef
href=""g1-lo1-summary.dita"" />
<learningPostAssessmentRef
href=""g1-lo1-post.dita"" />
</learningObject>
<learningObject navtitle="g1-lo2-ref">
<learningPreAssessmentRef
href=""g1-lo2-pre.dita"" />
<learningOverviewRef
href=""g1-lo2-over.dita"" />
<learningContentRef
href=""g1-lo2-content.dita"" />
<learningSummaryRef
href=""g1-lo2-summary.dita"" />
<learningPostAssessmentRef
href=""g1-lo2-post.dita"" />
</learningObject>
<learningSummaryRef href=""g1-summary.dita""
/>
<learningPostAssessmentRef href=""g1-post.dita""
/>
</learningGroup>
And you can include any number of learningPreAssessmentRef's
or learningPostAssessmentRef's, so
you could conceive of using the map to gather the
pool of possible questions, and then apply
processing against the map to choose a random sub-set
(or a sub-set based on metadata, etc.) from that pool
to construct any given test.
I'd be interested in what others think.
And be sure to check out the material about the SCORM
processing templates under development
as a "starter kit" of DITA processing to
support the learning specialization in the wiki pages here -
http://wiki.oasis-open.org/dita/LearningContent/map2manifest_templates
My best.
John
___________________________________
John Hunt
DITA Architect / Lotus Information Development Center
Chair, OASIS DITA learning and training content sub-committee
IBM Software Group/Lotus Software
john_hunt@us.ibm.com
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]