(j3.2006) Question about DIM

[Van Snyder]

Does anybody remember a compelling reason why the arguments of DIM are
required to be of the same type and kind, whereas, e.g., the arguments
of DOT_PRODUCT can be any numeric type and kind with the type and kind
of the result determined as for the product VECTOR_A*VECTOR_B?

Why not allow X and Y to be either integer or real, of any kind,
independently, with the type and kind of the result as if for X-Y as
specified in 7.1.9.3?