> The reference implementation defines ascii-numeric? to take two
> arguments, x and radix, while the srfi text defines it takes one.

Good catch. Sorry about the mistake and thanks for reporting it.

`ascii-numeric?` is meant to be a direct counterpart to RnRS
`char-numeric?`, i.e. take only one argument (the character to be
tested) and check whether it is a digit 0..9. Therefore the SRFI text is
correct and the sample implementation is incorrect.

I just sent a PR to fix the implementation:
<https://github.com/scheme-requests-for-implementation/srfi-175/pull/13>.