I also don't think there is a good way to do this, as
c(1, NULL, 2) reduces automatically to c(1,2).
Only if data gets to the function as a list, then it could have the NULL
value. But it is unlikely that would happen.
Should all the property functions return NA in the situation the sequence has invalid characters, or at least don't consider the character in the normalization by length (NULL and NLL should return the same)? It seems some functions that look into the different scales, return NA, but others don't.