Fractional formatting sometimes is expected to be restricted to a sequence of preferred denominators. The most frequent such sequence should be 2,4, 8, 16, 32 - or more generally the powers of 2. If it can be done without breaking the concept, a code for fractional formatting like # ?/[2^5] (5 denoting the maximum exponent; if omitted 14 e.g.) or (to give a funny example) ?/[3;7;11] should be established. Implementation should choose the smallest denominator usable without loss of accuracy. For export to an alien file format the sequence can be replaced then by the highest (realistic) deniminator matching it, or to the respective number of decimal digits in the denominator.
I'm not sure to get it, what is the advantage of having "# ?/[2^5]" instead of "# ?/32"? So far to me, extra complication in number format notations. Same for ?/[3;7;11] which is not clear the exact meaning. Can you elaborate on it?