behaviour of manually graded task (and answer) differs from original nbgrader

[szazs89]

(Maybe it is a documentation issue) EDIT: this report is false, the behaviour is quite the same as described in the last column -- see the comments in the follow up

The Teaching with nbgrader section refers to the nbgrader documentation about creating different question type. However, I found the following major dfferences:

Cell type	nbgrader	CoCalc
Manually graded answer	emptied, editable	emptied*, editable
Manually graded task	editable*	locked

where

• emptied* means, that CoCalc makes fully empty this cell, if there is no === BEGIN SOLUTION === / === END SOLUTION === snippet in that (in the latter case Cocalc removes only that snippet)
• editable* means, that nbgrader enables to edit this cell only if there is a === BEGIN MARK SCHEME === / === END MARK SCHEME === snippet, which is removed from the student version. (CoCalc totally discards any kind of markings end shows the original content but locked in the student version.)

It is quite confusing... (I hope I did not make any mistakes above.)

[williamstein]

Regarding "Manually graded tasks", the nbgrader docs say

"manually graded tasks cells are not edited by the student.",

so it seems confusing to me that nbgrader actually makes manually graded tasks editable in some cases. Is there documentation wrong or confusing?

[szazs89]

Hm... it seems I was wrong. Really, the manually graded task cells are not editable either. What confused me is the MARK SCHEME section which is removed by nbgrader but not by CoCalc. (I guess this is because of the CoCalc special Intructor only cells, which can be used for the same purpose, i.e. helping the grading.)

(I am trying to migrate collected assignments in CoCalc to submissions in nbgrader. Thus, I could use Formgrader and its generated html feedback. And the notebooks use octave kernel... :-) )

[williamstein]

Thus, I could use Formgrader and its generated html feedback.

In addition to this issue, feel free to create a precise issue related to adding an equivalent to "formgrader and its generated html feedback" to CoCalc. I would be very likely to implement that if it was thought through in a way that makes sense for CoCalc.

[szazs89]

In addition to this issue, feel free to create a precise issue related to adding an equivalent to "formgrader and its generated html feedback" to CoCalc. I would be very likely to implement that if it was thought through in a way that makes sense for CoCalc.

As far as I see, there is already a similar issue: #4723 (I continue there...)