[eeps] Fourth draft of JSON proposal

[Richard A. O'Keefe]

On 12 Aug 2008, at 4:05 am, David-Sarah Hopwood wrote:
> UTF-EBCDIC seems to be mainly intended as an internal format for
> systems based on EBCDIC that are migrating to Unicode, rather than
> as a format for interoperable data. Applications on those systems
> are likely to have support for producing UTF-8, so I don't think
> that support for UTF-EBCDIC should be required.

The very paragraph that mentions UTF-EBCDIC says

     Until the day
     someone ports Erlang to a z/Series machine, this is mainly of
     academic interest, but we don't want to paint ourselves into
     any corners.

That is to say, the EEP does *not* require any level of support
whatever for UTF-EBCDIC.  The point of mentioning it was to
emphasise that the specification "JSON text SHALL be encoded in
Unicode" is really hopelessly vague, and to argue for representing
JSON strings internally in UTF-8 and UTF-8 *only*.

The argument that "those systems are likely to have support for
producing UTF-8" in fact *strengthens* the claim I was making,
which was "Erlang JSON strings should be binaries using UTF-8,
not whatever the external encoding was."