QTI Extensions for Mathematics and Science

This page has been prepared to support the meeting held at Heriot-Watt University on 27th August 2003. This meeting was called to bring together people trying to use or considering using the IMS Question and Test Interoperability (QTI) specification to deliver computer based assessment in mathematics and other numerical disciplines.

The primary aims of the meeting were to:

• Foster the establishment of a focus group of practitioners,
• Reach a consensus on key needs of the community and the key issues for debate,
• Provide a summary of these key issues for presentation to IMS as they consider updating the QTI specification.

Notes from the meeting are available as a Rich text Format document: QTIM.rtf. Some notes by Steve Lay (which he intends to table as part of the working group for the new QTI 2.0 specification) are also available as an rtf document: slnotes.rtf. The next section links to resources identified on the day and collected since then.

Please email Colin Milligan (c.d.milligan@ma.hw.ac.uk) if you have any further links which you think would be relevant to the group - or if you have any comments on the meeting document (e.g. if I have failed to understand some Mathematical nuances).

Resources
The following resources are of relevance:

• The QTI-IMPLEMENT JiscMail List home page is at: http://www.jiscmail.ac.uk/lists/QTI-IMPLEMENT.html
  This list is managed by Dick Bacon. It is hoped that it will become the main mode of communication for the group.
• Dick Bacon's draft proposal covering "QTI support for question variables, expressions and graphs" has now been revised to Version 1.3 and is available here as a.doc file: stmp_qvarv1p3.doc
• The OpenMath home page is at: http://www.openmath.org/
  OpenMath was described during the meeting as being semanitically clearer (and richer) than MathML. Ina recent jiscmail post, MM remarked that for algebra and calculus, MathML would be accurate,but MathML may not be robust enough if support for proofs, logic and set theory were required.
• The MathML home page is at: http://www.w3.org/Math/
  MathML is the W3C standard for representation of Mathematical Expressions on the web. MathML can be either presentational (with no inherent mathematical meaning) or Content, where the XML description language also captures the algebraic meaning of the expression.
• The QTI homepage at IMS is at: http://www.imsglobal.org/question/index.cfm
• The Online QTI documentation supplied by Niall Barr is available at: http://ford.ces.strath.ac.uk/QTI/
  This site allows you to navigate through the elements in the specification and understand its' structure. The use of commenting provides a useful way of building up a structured knowledge base. The site above also provides a static link to the QTI test engine site which Graham Smith maintains, and a link to a set of examples which show interoperability between the QTI extensions code written by Dick and implemented within the QTI test engine by Graham. At present, Graham Smith's online QTI engine supports up to v1.2 of the QTI extensions doc.
• The paper presented at the 7th CAA Conference by Dick Bacon is available, entitled: Assessing the use of a new QTI tool within Physics. Another paper which considered similar isssues at the same conference was presented by Joan Parellada and was entitled Standardisation of Physics Exercises, Dynamical Generation of Data.
• Manolis Mavrikis demonstrated the WALLIS system which can be seen at: http://www.maths.ed.ac.uk/~wallis and recently wrote a paper about WALLIS for the LTSN MathSTOR site entitled: Incorporating Assessment into an Interactive Learning Environment for Mathematics
• The AIM System: http://aimmath.sourceforge.net/ is an Open Source testing system which utilises Maple for its expression handling etc. There is currently some work being done within the WALLIS project to allow interchange of questions with WALLIS (which utilise a more QTI-like format).
• There was a question in the meeting about MathML support for units. A quick search of Google returned the following archived discussion from 2001. Further investigation would be neeed to provide a definitive answer, but the linked discussion identified some possible solutions.

Examples
One of the outcomes of the August 27th meeting was that we should investigate the potential of MathML and OpenMath by creating examples within the group. It was also proposed that we should generate examples to define our view on expression simplification. It is hoped that this site will record the outcomes of this activity.