Hello,
I'm searching for some information how moodle implements the qti 2.1 standard. Does someone know if there are some documentation documents?
Many thanks in advance for help.
It doesn't.
Does any system actually implement QTI 2.1?
What do you want to do that requires QTI 2.1?
Does any system actually implement QTI 2.1?
What do you want to do that requires QTI 2.1?
ok, moodle doesn't implement qti 2.1, but can export code that looks like qti 2.0. Do you have any information about this?
>>Does any system actually implement QTI 2.1?
Yes, for example onyx, equiz, asdel.
>>Does any system actually implement QTI 2.1?
Yes, for example onyx, equiz, asdel.
You are more likely to get informed answers about Moodle's abilites to handle Quizzes and Question types if you post in the quiz forum: http://moodle.org/mod/forum/view.php?f=121
Your three examples are not very convincing - the Onyx project web site only seems to mention QTI1.2, although my inability to read German may be getting in the way. (By the way, couldn't you have given me the URLs of the project web site to save me googling?)
equiz - I can't find any project page for - google only picks up a few emails on what looks like a Plone project email list.
asdel is a JISC funded research project trying to write a QTI 2.1 question engine which is about half way through. That is hardly a working implementation of QTI 2.1.
I do know one working example of a QTI 2 question system for Physics, by a semi-retired professor from Southampton. Like most such efforts, it was done be first reading and understanding the QTI specification (a remarkable achievement, I have a PhD in maths and I find them almost incomprehensible) then designing the entire question system around how the QTI specification works.
To my mind, that is totally backwards. QTI is meant to stand for Question and Test Interoperability. For it to be impossible to support unless you write your entire test system to support their model does not promote interoperability of existing testing systems. Moodle has had its own quiz code since 2002, long before QTI got anywhere.
Fortunately there is a way out of this. I am pleased to see the asdel project says it is going to "build an assessment delivery engine ... that can be deployed ... as part of a SOA enabled VLE". If they want to do this in a way that is Moodle-friendly, they should implement Opaque, a de-facto standard invented by the OU which does appear to be truly interoperable. It started out as the unnamed protocol used to connect the OpenMark test navigator to the OpenMark question engine. Then we needed to make OpenMark and Moodle interoperate, so gave the protocol a name, published it on the Moodle wiki, and made Moodle talk the test navigator end of it. Then the STACK developers wanted to make their question engine interoperate with Moodle, so they taught STACK to speak the question engine end of the protocol. And all that will work in Moodle 1.9 (in beta, will be released very soon), OpenMark 1.4 (already released) and STACK 2 (to be relesaed by Christmas). W3C only allow their standards to be published when there are two interoperable implementations, and Opaque has reached that point now.
To answer you specific question. I am not really sure of the current state of the QTI 2 export. I think it was a project that someone who was interested started. But no one has done any real work on it since at least Feb 2005 (Revision 1.3 in http://moodle.cvs.sourceforge.net/moodle/moodle/mod/quiz/format/qti/format.php?hideattic=0&view=log), and I don't know of anyone using it.
Finally, could I ask that you fill in some details in your user profile. It is considered polite here, and I woudl like to know who I am talking to.
(Oh, and in case you were wondering, the Opaque protocol is named in honour of my opinion of the QTI standard.)
Your three examples are not very convincing - the Onyx project web site only seems to mention QTI1.2, although my inability to read German may be getting in the way. (By the way, couldn't you have given me the URLs of the project web site to save me googling?)
equiz - I can't find any project page for - google only picks up a few emails on what looks like a Plone project email list.
asdel is a JISC funded research project trying to write a QTI 2.1 question engine which is about half way through. That is hardly a working implementation of QTI 2.1.
I do know one working example of a QTI 2 question system for Physics, by a semi-retired professor from Southampton. Like most such efforts, it was done be first reading and understanding the QTI specification (a remarkable achievement, I have a PhD in maths and I find them almost incomprehensible) then designing the entire question system around how the QTI specification works.
To my mind, that is totally backwards. QTI is meant to stand for Question and Test Interoperability. For it to be impossible to support unless you write your entire test system to support their model does not promote interoperability of existing testing systems. Moodle has had its own quiz code since 2002, long before QTI got anywhere.
Fortunately there is a way out of this. I am pleased to see the asdel project says it is going to "build an assessment delivery engine ... that can be deployed ... as part of a SOA enabled VLE". If they want to do this in a way that is Moodle-friendly, they should implement Opaque, a de-facto standard invented by the OU which does appear to be truly interoperable. It started out as the unnamed protocol used to connect the OpenMark test navigator to the OpenMark question engine. Then we needed to make OpenMark and Moodle interoperate, so gave the protocol a name, published it on the Moodle wiki, and made Moodle talk the test navigator end of it. Then the STACK developers wanted to make their question engine interoperate with Moodle, so they taught STACK to speak the question engine end of the protocol. And all that will work in Moodle 1.9 (in beta, will be released very soon), OpenMark 1.4 (already released) and STACK 2 (to be relesaed by Christmas). W3C only allow their standards to be published when there are two interoperable implementations, and Opaque has reached that point now.
To answer you specific question. I am not really sure of the current state of the QTI 2 export. I think it was a project that someone who was interested started. But no one has done any real work on it since at least Feb 2005 (Revision 1.3 in http://moodle.cvs.sourceforge.net/moodle/moodle/mod/quiz/format/qti/format.php?hideattic=0&view=log), and I don't know of anyone using it.
Finally, could I ask that you fill in some details in your user profile. It is considered polite here, and I woudl like to know who I am talking to.
(Oh, and in case you were wondering, the Opaque protocol is named in honour of my opinion of the QTI standard.)
Hello Tim,
first of all this is not a discussion of the advantages and disadvantages of qti. I only want to exampine which applications supports qti sufficiently. In fact it don't speak for the qti specifications that there are no applications that implements qti completely at the moment.
about onyx: (http://onyx.bps-system.de/). On the start page they write that their system is based on qti 2.1. I have created a qti content package and had tested it with the demo programme from the homepage. There are some errors, but the main components of the questions worked.
about ecquiz: (http://plone.org/products/ecquiz). It can handle only few qti questions types, multiple choice and textentryinteraction
>Finally, could I ask that you fill in some details in your user profile. It is >considered polite here, and I woudl like to know who I am talking to.
Some people like to show personal details in the internet, others don't like.
first of all this is not a discussion of the advantages and disadvantages of qti. I only want to exampine which applications supports qti sufficiently. In fact it don't speak for the qti specifications that there are no applications that implements qti completely at the moment.
about onyx: (http://onyx.bps-system.de/). On the start page they write that their system is based on qti 2.1. I have created a qti content package and had tested it with the demo programme from the homepage. There are some errors, but the main components of the questions worked.
about ecquiz: (http://plone.org/products/ecquiz). It can handle only few qti questions types, multiple choice and textentryinteraction
>Finally, could I ask that you fill in some details in your user profile. It is >considered polite here, and I woudl like to know who I am talking to.
Some people like to show personal details in the internet, others don't like.
(Oh, and in case you were wondering, the Opaque protocol is named in honour of my opinion of the QTI standard.)
Great stuff - I concur on the opinion, QTI isn't the most straightforward standard.
BTW, I'm taking part of an IMS-CC workshop where I'll have a chance to give IMS some feedback on CC and related standards. One of the "fun" things is that CC is a rollup of several standards, including the latest and greatest QTI.
Have you had a chance to look at it? Is there anything we should be telling the IMS crowd?
My worries with CC are mentioned here http://moodle.org/mod/forum/discuss.php?d=56017
To be honest, I have not looked at Common Cartridge. Another developer here is looking at it, and I have been busy with other things. At least that is my excuse.
QTI incorporates some clever ideas, but therefore it does not work like most question systems out there work, so I don't think it is very helpful for supporting Question and Test Interoperability.
On the other hand, what they are trying to do - give people a format that you can encode most learning content in, so that most other tools can use it - is a worthwhile goal.
So, I think I would make a few comments:
1. If IMS want their specifications implemented, then they need to start writing in language that developers can understand - and understand not by struggling for hours, by just by reading it.
2. They should fix their web site. Last time I wanted to ask some questions about one of their specs, I went to their forums (they were empty) and tried to register an account. The email sign-up system just gave an error. I email some support address, and never got a response. Compare that with the Moodle community. So if they want their specs implemented, they should try to get a supportive community started.
QTI incorporates some clever ideas, but therefore it does not work like most question systems out there work, so I don't think it is very helpful for supporting Question and Test Interoperability.
On the other hand, what they are trying to do - give people a format that you can encode most learning content in, so that most other tools can use it - is a worthwhile goal.
So, I think I would make a few comments:
1. If IMS want their specifications implemented, then they need to start writing in language that developers can understand - and understand not by struggling for hours, by just by reading it.
2. They should fix their web site. Last time I wanted to ask some questions about one of their specs, I went to their forums (they were empty) and tried to register an account. The email sign-up system just gave an error. I email some support address, and never got a response. Compare that with the Moodle community. So if they want their specs implemented, they should try to get a supportive community started.
Taking notes here... I think your point "0" of interoperability vs innovation is key... and wondering how to say all I have to say and avoid getting kicked out of the meeting!
