QTI-SIG People,

Following the meeting in Loughborough of the QTI working group the
following two issues which were discussed in Leeds were raised:

Min/Max in response processing.

The issue raised was what happens if multi-response multi-choice response
processing results in a number beyond the range of the min/max declared in
the associated decvar.  The position was clarified that outcome variables
must be constrained to their min/max values by implementations.  In other
words, if score is declared with min=0 and, during response processing, an
implementation attempts to subtract 1 from score when its current value is
0 its value will not change.  This discussion was not minuted though it was
agreed that we would examine the wording in the 1.1 spec to see if we can
help clarify this point in 1.2

Where is the stem?

We had some discussion about this as it was considered to be related to the
concept of section (or even assessment) level material which we have agreed
to support.  The upshot is that scenarios where you have some common
material followed by a set of questions each comprising stem and response
which you would like to manipulate individually will probably best be
implemented by placing the common material in the section and treating each
stem/response pair as an item.  The material for the stem is therefore the
material which sits in the item's presentation (but outside the response_*
tag) and can be distinguished.

This differs from the alternative approach which was to squeeze the stem
material between the response_* tag and the render_* tag and leave the
common material in the space between the item and the start of the response
and then manipulate the response_* tags.

Hope that makes sense.

Steve