Experimental assertion on menubar using NVDA

Hi everyone,

As an experiment I made an assertion using the template Matt shared to us 2 weeks ago.
It took me about 3-4 hours, and some of the thoughts/challenges I came up with were as follows.

1.Difficulty having comprehensive key commands. (column H)
-I wasn't quite sure the commands that I tested were comprehensive for each assertion, but this could be solved after some learning phase.

2.Accuracy of test results. (column I)
-There were some assertions which I wasn't quite sure how I should've marked.
For example, when users open a list of submenus in the menubar with enter key, should NVDA announce both menu/menuitem role?
It could be avoided if a tester was experienced, but these kinds of confusion delayed testing time in my case.
To keep the accuracy and quality, perhaps it'll be useful to create double-checking system using JF's prototype since it seems more intuitive for less SR experienced testers.

3.Unclear which output correspond to different key commands? (column J)
-This is a discussion made several times in the past few weeks.
Even in a same assertion, different types of announcements were made depending on different key commands and could be confusing.
I put numbers in the column for clarification.

Best,
Yohta