http://www.imsglobal.org/community/forum/index.cfm?forumid=11 FuseTalk Standard Edition http://www.imsglobal.org/community/forum/messageview.cfm?catid=52&threadid=490 2010-09-09T08:47:48 -05.00 mmckell 1
The demonstrations successfully showed that existing QTI v2.1 implementations could ingest the CC-QTI Profile test cartridges and "play" the QTI content. With this broad support, and because some of the tools can be embedded in existing learning platforms, the CC-QTI Profile is ready to become a widely accepted profile for sharing assessments. Due to significant changes in the XML structure of QTI 2.1, the new CC-QTI Profile allows much more consistent interoperability than the QTI v1.2 profile included in CC v1.0. As a result, simple tests can be easily exchanged from relatively simple LMS modules to much more complex specialized test engines. The CC-QTI Profile is designed for easy implementation and interested parties should contact the QTI project group for questions or assistance.

Onyx, MathAssessEngine, and Jassess applications also successfully previewed another profile that will be included as part of the Final v2.1 specification. This profile, still being developed, includes a significant portion of the full specification and is tentatively being called the Main Profile.

As work continues on the completion of these profiles, the ability to test the conformance of systems, applications, and content will be a core component of the Final v2.1 specification. The QTI project group anticipates finalizing the documentation and schemas for the CC-QTI Profile in the coming months and work on the Main Profile to be completed by early 2011, with the Final specification to be released soon after. The group would like to thank those who participated in this interoperability demonstration.]]> http://www.imsglobal.org/community/forum/messageview.cfm?catid=52&threadid=155 2010-08-11T10:18:00 -05.00 mmckell 7
thank you very much.]]> http://www.imsglobal.org/community/forum/messageview.cfm?catid=52&threadid=155 2010-08-10T11:08:53 -05.00 mmckell 7
I wasn't directly involved in QTI at the time, but I think the main difference between QTI 2.0 and 2.1 is that 2.0 updated the question item datamodel and interactions as well as the packaging piece. 2.1 updated everything else: the test model, results reporting etc.

At the time, the team were aware that all of the specification needed to move from 1.x to 2.x, but wanted to do the most difficult and critical pieces first. They then found that a lot of people were keen on the remainder as well, so they immediately continued with 2.1 after completing 2.0.

This means that items that were written to 2.0 are pretty much valid to 2.1 too- not much changed there.

Regards,

Wilbert]]> http://www.imsglobal.org/community/forum/messageview.cfm?catid=52&threadid=155 2010-08-05T05:02:48 -05.00 mmckell 7
thank you for the list. Why is it that there is nothing mentioned about QTI 2.0?

Marjan]]> http://www.imsglobal.org/community/forum/messageview.cfm?catid=52&threadid=421 2010-06-24T05:26:32 -05.00 FrancoiseLeCalvez 1 parameter declarations in template processing and the constraints between parameters that are our main concern, MathAssess (Fetlar) has chosen to create ScriptRules in a customOperator which contains Maxima commands.
Our proposition was to extend QTI with templateConstraint (see below the proposal we made in 2007). For the implementation, our templateConstraints are translated into a java program and a php program in order to be compatible respectively with tomcat and apache servers. As we have said before, these programs are created by our QTI-Editor and used by our Web-Page-Generator in order to create PHP and JSP pages to be run on many not-dedicated servers.
CustomOperator to integrate extern commands is a very powerful tool to express specific aspects. Thus, we think it will be useful for us to add our java and php programs using this operator in the QTI file, so that there would be only one file for QTI expression of the item and for the constraint implementation. In this way, it would be possible to have in the same file several possibilities of implementations (as for example, Maxima and Java ...) in order to calculate values for the template variables which are defined by interdependent constraints.

But, in order for item definitions to be interoperable and not dependant of the implementations and thus for QTI to be a standard, we think it is necessary to have constraint definitions independent of the implementation. It is why we think that the class templateConstraint would be added in the profile.

Such an extension will allow having QTI files, which can be read by several players and according to the particularities of the implementation of these players, would be run using Maxima or Java or php or any other CAS.
Moreover using these extensions would allow working with the items at a meta-level and would make QTI a more general specification.

Odette Auzende, Hélène Giroire, Françoise Le Calvez (UPMC)

templateConstraint Definition:
Class: templateProcessing
Associated classes: assessmentItem
Contains: templateRule [1..*]
Contains: templateConstraint[0..*] to declare a constraint binding several templateVariables

Class:templateConstraint
Contains: expression[1..*]]]> http://www.imsglobal.org/community/forum/messageview.cfm?catid=52&threadid=415 2010-06-22T09:47:55 -05.00 mmckell 1