device protocol interprete holding register as DINT/REAL

In the "New holding register" dialog of the device protocols it is possible to configure the transformation of the passed bits. The number of bits can be configured. However, only integer numbers are interpreted. However, the holding registers can also contain floating-point numbers of type REAL and only as DINT. It would be useful to provide an option in the dialog whether the bits are interpreted as integer or floating-point numbers.