> I think having the query functions named data_is_ascii_* may be sufficient to alleviate these concerns.

In general, I understand that it's a bit odd to have something like
is_ascii_digit('0') be potentially false and run a purely numerical
test that ignores literal encoding, but every alternative seems worse
here. I don't feel like saying "data_is_ascii_digit('0')" meaningfully
improves this.

We also usually say "lerp" or "midpoint", not "numbers_lerp"; adding
"data_" to the name feels like it repeats information that is evident
from the parameter. Isn't a char a datum? Why would we need to say
"data_", as if that added any sort of information?