The actual code 'complexifies' real or 'integer' numbers instead of directly computing.
Example:

real(x) * complex(y) =>

x = number_to_complex(x);
y = number_to_complex(y);
z11 = number_times(x->cmp.cmp_real, y->cmp.cmp_real);
z12 = number_times(x->cmp.cmp_imag, y->cmp.cmp_imag);
z21 = number_times(x->cmp.cmp_imag, y->cmp.cmp_real);
z22 = number_times(x->cmp.cmp_real, y->cmp.cmp_imag);
z1 =  number_minus(z11, z12);
z2 =  number_plus(z21, z22);
vz = make_complex(z1, z2);
return(z);

Is it possible to use ? =>

z1 = number_times(x, y->cmp.cmp_real);
z2 = number_times(x, y->cmp.cmp_imag);
vz = make_complex(z1, );
return(vz);