Confirmation output from the previous attempt

[repagh]

I tried to slightly re-purpose the coderunner question type to applications for control theory, and possibly more general to applications where students supply the answer in Python variables.

The question template and support code is on a separate project:

https://github.com/repagh/practicon.git

I worked with the suggestion as given in https://coderunner.org.nz/mod/forum/discuss.php?d=309
but was not very happy with the result. In the end I forked the repository, and worked initially on using the showoutputonly option, but since that only works in the question feedback, and thus was tied to showing the same feedback during attempts and after completion, I changed course and added a new variable to the template output, named attemptconfirmhtml. If available, it is added to the question presentation. In the practicon template I use it to give confirmation on the answer variables the user has supplied.

I hope you consider this a worthwile and sufficiently general extension to coderunner.

[trampgeek]

Hi Rene

Sorry about the long delay in replying to this. It's been a busy time here.

Certainly the issue you raised in the original forum posting ("pre-check without hints on question score") is important. In programming questions it's not so important (I believe programmers should be told if their answer is right or wrong) but as CodeRunner is increasingly used for non-programming questions it's getting more important to find a good solution. Indeed, in the last week, two of our courses here at Canterbury have been dealing with the very question you raise.

The solution you've provided in this pull request is very ingenious. However I'm a little uncomfortable with using it because it changes the underlying Moodle question behaviour model. The standard assumption is that a question has associated feedback (immediate or deferred) and that the quiz author can control that feedback via review options.

The behaviour of the question over multiple submissions is managed by the question behaviour and I have a special subclass of the base adaptive behaviour called adaptive_adapted_for_coderunner. In principle (perhaps) I would have such behaviour subclasses for all the built-in behaviours but that becames exceedingly cumbersome. So I settled on locking CodeRunner questions into Adaptive mode which I think is right for programming questions (the original intended domain of application). But now we're moving into a space where we want a mix of adaptive mode (with a button providing limited feedback) and deferred feedback (where actual marking isn't done until the end).

I would prefer to accommodate your needs without changing the standard Moodle model, if possible. What we finished up doing at Canterbury for our exams was to hide the Check button in selected questions, as hinted at in this forum posting, but on a question-by-question basis. The Precheck button then provides as much in-exam feedback as we want to while actual marking is deferred until the quiz is finally submitted and closed.

This is a hack and needs to be properly integrated into both the question and question behaviour plugins but it is more in line with the Moodle design. So my question is: would having an option in the author editing form to disable the Check button solve your problems?

Thanks for all your work on CodeRunner. It sounds like you're doing great stuff.

Regards

Richard