Hi all,

we were talking about designing an adaptive question with multiple choice, with the feature that, in case the student answers wrongly, in the subsequent adaptive section a text dependent on the given answer appears.

For example, the question may be: "which of the following football teams is not an English one?". If both Chelsea and Liverpool are among the alternatives, then if Chelsea is answered we would like to show a consistent text as it would be "wrong: Chelsea is a district in London", while if Liverpool is answered the text would be like "wrong: Liverpool is a city in England".

Unfortunately, response-specific feedback does not fit for this case, since it is shown while the question is evaluated, that is, at the very end, not while the question is verified, i.e. when the correctness of the answers in the current adaptive section is checked.

I guess it would require to catch the answer given by the student, after having assigned a text to every (wrong) alternative, but I don't know how to perform it; I worked with chained questions in the past, but there a specific question made use of the answers given in a previous question, not in a previous adaptive section belonging to the same question.

Thanks in advance, best regards,

Fabio

We solved: it is sufficient to remove the lines relative to "privacy", "automodule" and "exportFrom" from the XML file, because they misguide the TA 10 interpreter, while being not necessary for the correct interpretation of the contents. Thanks anyway!

Hi all, again. While trying to export a question, as a CourseModule (doing it as a Question Bank makes substantially no difference), from the newer version to the older one, I get the following error:

Error loading the course module:
privacy : privacy : privacy : privacy
---- Debugging information ----
message : privacy : privacy
cause-exception : com.thoughtworks.xstream.mapper.CannotResolveClassException
cause-message : privacy : privacy
class : gateway.question.xml.CM
required-type : gateway.question.xml.module.Mod
line number : 7
-------------------------------
Error reading XML file:Illegal character in path at index 0: 221ea3ce-e2e9-4974-a26c-a966bff48628

Of course, I tried to export a question which type is present in Maple TA 10, i.e. a simple multiple choice one. Is there a way to circumvent this error, and export backwardly?

Fabio

Hi all, is there a way to create an assignment which makes use of questions external to the course (they could be also from a larger shared database, or even the Maple Cloud), such that if a question is updated independently (people working on the former one will not necessarily work on the latter ones, and vice versa), then also the assignment would present it in its updated form?

Thanks in advance.

Hi again; so, after solving the doubts discussed in the previous topics I opened, we tried with some students to simulate an exam, in order to check the state-of-the-art of our knowledge of the platform. We had basically two problems:

- In homework/quiz mode, it seems that two people at the same time are able to login with the same credentials, and furthermore if the second student does the login while the first one is already logged, the second one is also able to see all the answers given by the first one before he/she entered the system. This is a problem if a student gives his/her access data to a mate (which is of course illicit, but it would need means to prosecute it).

- In proctor mode, the previous problem does not hold, but proctor authorization is a bit unstable, sometimes the proctor login goes down without apparent reason (no timeout or specific actions) and this is a problem because evaluation remains locked until the proctor is logged again (which requires to pass through Moodle since we do not accept direct Maple TA login, so taking precious time while doing an exam).

Is there a way to overcome them?

Thanks in advance.

Ok, thank you. The possibility of allowing the student to navigate through other sites or programs while doing the exam is in discussion since it can be considered not compliant with standards concerning safe exams procedures, but this option can be taken into account.

If you are playing with JavaScript, I would thank you very much in the case I will be able to hear something interesting from you in a more or less far future.

Two weeks ago we discussed ( https://mapletacommunity.com/t... ) about the possibility of giving a student the means for drawing an image as part of his/her answer.

Unfortunately, we saw that the scanned document question requires more assistance from the teacher (e.g. he/she has to upload personally all the figures by associating them to the codes automatically generated for the students, without the possibility to proceed for these ones) than we thought. On the other hand, the sketching question seems to aim at certain types of plots (elementary functions like parabola, exponential, logarithm), rather than at a generic graph that could be, for instance, the one relative to f(x) = exp(sin(x)+1), a function which is a composition of elementary ones, but not an elementary one itself. Last but not least, the writing of a custom JS code seems that would take a consistent amount of time.

So we were asking ourselves if there was a possibility for the students to upload, for example in an essay-type question, an image made by them with another software on the same PC on which they are doing the online test. Although for designing a question it is immediate to find how to upload pictures, the student seems to have less options in his/her WYSIWYG editor, and the one relative to uploading is apparently missing. It is peculiar that he/she has the possibility to modify directly the source (and for example to insert a <img> tag, able to hotlink potentially every reachable image on the www), but neither upload an image from local nor from remote via wizard. Is there a possibility to do it?

Thanks in advance.

Yes, I'm using the 2016 version. It worked, thank you! I used $x*$response.provachain1.1.1 in the response area, where $x=4; was defined as the only line of the algorithm. Good to know also that it is wrong to consider the (addable by the user) field "id" from "Information Fields" as equivalent to the "id" appearing in the source, which is not the case.

Thank you, I just sent you the texts.

Ok, thanks. We explored both ways since HTML is more versatile but chaining seems to be more immediate, not requiring a particular code knowledge.