Define and use parameter for precision of floating point variables

[Beliavsky]

The codes use default REALs. It is generally recommended that modern Fortran programs define a working precision parameter for floating point variables and then declare variables using that, as described in the first answer to a question here. If this were done, it would be possible to run all of your codes in either single or double precision by just changing one line of code.

[MPenaR]

What about using iso_fortran_env parameters real32, real64 and real128 instead of using selected_real_kind?

In any case, the main purpose of this repository is for teaching numerics to non programmers.