More Feedback/Questions on 8 March Design

[Eliot Kimber <ekimber@reallysi.com>]

Here are more questions on the design:

- The ID attribute is required on all the interaction question types but 
this is not indicated in the formal docs. Should ID be required? I would 
think not (it certainly breaks with the DITA pattern where only topic 
has a required ID and there would be many use cases where interactions 
would never be referenced). Is this driven by the IMS QTI requirement 
that each assessment item have an ID? If so it seems unnecessarily 
restrictive if we assume that to go from DITA learning content to actual 
QTI markup one would have to go through a transform in any case. That 
transform could of course assign IDs to any items that did not already 
have them.

- With regard to my earlier comments about the need for questions to 
allow multiple paragraphs I see that one of the goals of the learning SC 
is to be compatible with the IMS QTI design. If I understand that 
specification, it explicitly limits the prompt content of a question to 
a single block. This seems inappropriately limited. In any case, I have 
a requirement for multi-paragraph questions. Here are several real 
examples from published test prep documents:

Question 1:

   <p>Mrs. Daley is placing students into the fourth-grade classes at 
her school. There can be no more than 25 students in each class.</p>
   <p>Which of the expressions below best represents the number of 
students, n, in a fourth-grade class?</p>

Question 2:

<p>The spinner below is divided into 8 equal sections. Xavier spins the 
pointer.</p>
<art>
   <Image href="art/93OH_CR6_02.eps"/>
</art>
<p>What is the probability that the pointer will land on yellow?</p>

These questions are typical.

Certainly the naive approach to this content is to say that the sequence 
of paragraphs or paragraphs and assets represents the question as a unit.

But I could also see an argument that says that everything up to but not 
including the final paragraph is some sort of intro or preamble to the 
assessment item and only the last paragraph is the actual prompt, 
leading to a model like:

<lcSingleSelect>
  <lcPreamble>
    <p>Mrs. Daley is placing students into the fourth-grade classes at 
her school. There can be no more than 25 students in each class.</p>
  </lcPreamble>
  <lcQuestion>
    <p>Which of the expressions below best represents the number of 
students, n, in a fourth-grade class?</p>
  </lcQuestion>
  ...
</lcSingleSelect>

As far as I can tell from my initial reading of the QTI spec the 
foregoing cannot be accommodated e.g. by <choiceInteraction>, which 
simply allows <prompt> followed by <simpleChoice>, unless you accept 
that the paragraph breaks in the original source would become <br/> in a 
QTI result (yuck).

Another approach that occurs to me here is to put each assessment item 
into a topic where the topic represents the question as a whole from a 
presentation standpoint (that is, as the question will be rendered in 
the print publication) but the assessment item is just the final 
paragraph plus the choice items.

This last approach would also handle the case where you have a multipart 
question, such as this real example:

7. The Carr Memorial Aquatic Center has a diving pool and a swimming 
pool, where people can swim laps. The diving pool is 20 yards long, 20 
yards wide, and 12 feet deep. The swimming  pool is 50 yards long by 25 
yards by 6 feet deep.

a. What is the volume of the diving pool?

________________________

b. The swimming pool is being repaired. It needs to be painted. What is 
the surface area of
the combination of the swimming pool and the cover for the pool? Show 
your work.

________________________

c. What is the volume of the swimming pool?

________________________


I'll have to think about this more.

Cheers,

Eliot

-- 
Eliot Kimber
Senior Solutions Architect
"Bringing Strategy, Content, and Technology Together"
Main: 610.631.6770
www.reallysi.com
www.rsuitecms.com