IMS QTI Working Group, Ottowa, 20-21 August 2001
Notes from N.Sclater.

This meeting concentrated on reviewing the draft v1.2 results interoperability, outcomes processing and selection and ordering specifications before they are submitted to the Technical Board for approval.

We had an initial look at contextres; there was a query on the use of the Genidentifier element.  The GUID is inherited from the Enterprise and LIP specs.  The options were to just have a basic GUID or to keep the three sub fields: typname, source & ID.  See page 97 of ASI.  It was decided that source is removed and that we will just include typename & id.

Datetime uses an ISO standard - 8601

Summaryres:  the status attribute refers to whether this is the final score, after human checking for example.  Statusdescription was changed to statusvalue.  Duration is also defined under ISO8601.  Scorereliability is to be added ie if you took the test more than once what is the reliability of it. Ammendments were made to the various attributes of score.

Assessmentres: the original asi didn't have the concept of grade.  There was discussion as to how you calculate grade from score - this is not obvious.  Grade contains gradevalue and gradecut.  Gradeset is being removed but gradecut being kept.

There was a discussion on metadata and how to incorporate ims metadata into results.

Response: responseidentifier was moved from responseform into an attribute of response.  A problem is how to identify whether someone didn't reply or not.

SCORM: the mapping of results has been done by ADL and is straightforward - it will be posted to the forum imminently.  SCORM - IMS QTI integration will not be achieved before the end of this year. ADL is particularly interested in the implementation experience of QTI and is keen to do a joint plugfest at some stage. Agreed to try and compile list of people who are using the spec and they can choose if they want to make their entry public or private.

It was noted that people are complaining that the element names are often too cryptic in QTI.  You're not actually saving anything by leaving out full element and attribute names.  The EML people started off with cryptic ones and now have huge but comprehensible ones.   Before moving to schemas it would be worth making element names easier to understand.  It was also mentioned that IMS policy will be to use underscores when there are two words.  We then started rewriting all names in results.

Outcomes processing: This allows you to define your own scoring algorithms by defining the differences between inputs and outputs.  We spent much of the second morning going through the various outcomes processing use cases.

Can Studios, a new member of the group had a query.  They are using Shockwave to render qti.  They need to position text with xy coordinates using mattext.  There was concern about ammending the spec from an interoperability perspective.  Precise text positioning might not be appropriate for participants requiring large text for instance.  However other elements such as images already have positioning data.  There is a need to apply a context for the xy coordinates - does the position relate to the screen, the window or what?  It was agreed, after much discussion, that coordinates for text would be incorporated but it should be ensured that there is something in the spec saying that the rendering system can ignore these coordinates if necessary.