Anyway, I thought about the reason to require all MPI characters be
stored in a certain representation but could not come up with a reason.
Setting such a requirement would means any machine that uses a different
set of characters representation would have to do conversion between its
native character sets and the ISO-8859-1. But what good does it do?
Can someone explain the benefits of requiring such a standard for the
MPI_CHAR and MPI_CHARACTER data types?
If a standard should be used, is ISO-8859-1 the one to use?