Jonathan Rosenne wrote:
>
> I feel that when identifiers are case sensitive, such as in C, there
> may be a case for respecting points,
Indeed, C++ and C99 legislate that points are acceptable part of
identifiers' names (but OTOH cantillation marks are not portable).
> although this would cause a problem with cross-system portability,
Why? C++ Standard, followed by C99, which introduced the possibility to
use character outside ASCII in identifiers, also introduced a way to
"type in" them in a portable way, using the \uxxxx notation also used
in Java. So even if users are not able to type in the characters at
keyboard, they have an escape way...
Of course, following what I wrote above, transferring sources from
one machine to another need to avoid "converters" that drop the
points, in the same way that a converter that convert all unrecognized
Devanagari codepoints to U+FFFD or similar is also to be avoided.
Antoine
