Regarding the drop down menu, yes, I think it would be better to only have the "task cell" option visible for markdown cells.

The documentation asserts this too "If you select the “Manually graded task” option (available for markdown cells), the nbgrader extension will ..."

Sometimes the output percentages of flows in gradebook are not the actual percentages, when the grading rule set credit_percentage other than 100, for the purpose, e.g., to encourage in-time submission of assignment. However, there are cases we need to know the actual percentages (actual correctness if neglecting credit_percentage).

Unfortuately, the credit_percentage is only used [here](

The API is well-defined in the code, but it would be nice to have it in ReadTheDocs.
It would also allow us to add bits of useful information, such as the fact that #470 exists.

If a teacher goes through to give specific grades to a number of different students in a group, the "apply to all" button becomes destructive. The button is useful if a teacher accidentally applies a score incorrectly to everyone, but the destructive nature of the button should be conveyed if it is hit accidentally.

To avoid annoyance when teachers want to use to to apply to everyone, possibly

There are several new features of GatorGrader that are not documented in the README file:

• Use of Docker
• Plugin-based approach
• New sample repositories
• Description of the check for markdown components
• Description of the checks for different kind of counts
• References to other auto-grading tools (see my new internal list)
• Reference to article showing what happens when auto-grade

Create some sort debugging system where students can go to the menu and trigger an email that contains all the important debugging information.

Stuff to include:

• OS version
• Markbot version
• Java version
• The assignment & course
• Computer folder paths
• Markbot cheat status
• URLs for the different servers (same as Activity window)
• HTML & CSS va

We should take the time to consolidate these functions into one and add the 'code' key into the dictionary so that feedback is more easily analyzed. This should probably happen over the winter break when BlockPy currently isn't in use so we can fix all the 1064 Curriculum.

The parsec parser does allow to provide a "Feedback judgement" like this:

# Assignment 1: 10/10
## Feedback:
  * Good work!
## Task 1: 10/10

Then, we can extract the feedback using the CLI: