I've come across this problem while proofreading someone else's questions. The author has used the variable `x_0` in the answer to a mathematical expression part. I expect that some students will just type `x0` without an underscore if they're not told explicitly what to do.
Numbas accepts both forms, and they're both rendered exactly the same, but they're not considered to represent the 'same' variable.
I wonder if they should be considered identical. I think my preferred way of implementing this would be to normalise the version without an underscore to the version with, during parsing.
Does anyone have any objections to that, or suggestions for other solutions?