I'm sure you get complaints on all sides, but thought I would advocate for tightening/returning the restrictions around calculating with numbers from nominal variables. It is confusing to create means, show histograms, allow them as regression DVs, etc.
I'm guessing the disagreement comes over the treatment of binary variables, since many consider them nominal. But, if one is going to take advantage of their unique mathematical properties, they are not. Or, maybe it is because lots of variables are [mistakenly] imported as categorical and it is better not to have restrictions immediately.
However, there are so many types of "in-between" values, including continuous with labeled missing values, discrete (or structurally discrete) continuous values, likert, skewed, and more. The ordinal category seems sufficient for such variables that can be treated multiple ways or need to be treated specially. Perhaps the label could indicate this.
Setting and enforcing restrictions on nominal variables would make jamovi so much better for teaching and new researchers, and experienced researchers should understand enough to set their variable types accordingly.
Thanks for creating jamovi and listening.