Eliminate implicit mapping

[Beliavsky]

@certik

@mfurquan

It offers best of both worlds. Straightforwardly interpreted strong typing, sans explicit declarations.

I think more useful “implicit typing” would be to infer the type from the right hand side, so that you can do things like:

a = [1., 2., 3.]
sin(a)

This would not be possible with the Fortran’s implicit typing because you can’t declare arrays that way (I think), only scalar variables.

I have suggested using ASSOCIATE for this, dropping the current requirement that each ASSOCIATE statement be paired with an END ASSOCIATE statement. There would be an implicit END ASSOCIATE when the end of the program unit is reached.