Re: [Json] Wording on encoding; removing the table

[Larry Masinter <[e-mail]>]

" The transfer encoding
    used to represent the characters on-the-wire is beyond the scope
    of this document."

On the contrary, it seems to me that this document is (or should be)
MAINLY ABOUT normatively establishing how JSON text is encoded
in an application/json message body "on the wire" -- something
ECMA 404 doesn't do.


JSON in ECMA-404 is a sequence of Unicode codepoints.
application/json is a sequence of octets. Those octets are used
to represent ECMA 404 JSON text, in the manner prescribed
in this document.

Insofar as ECMA 404 is updated to include the substance from this
draft (guidelines for encoding JSON text and registration of
the media type), then anything that remains (implementation
advice, non-normative ABNF) could remain in IETF, and
obsolete 4627 but normatively point to 404.

This might help resolve some of the liaison  issues as well
as clarifying the relationship of the two documents.

>From a "first principles" point of view:

* MIME types for a format are best defined in the same document 
  which defines the format. That sometimes means enhancing
  the format definition to talk about encodings in bytes in a way
  that it wouldn't otherwise be defined.

*  Having two overlapping normative documents with different
   change control is bad for multiple reasons. Even if non-optimal
   for one side, it's better to avoid overlap; one way to do that
   is by making one document    normatively reference the other,
   even if it (for information) it restates some of the information.

Larry
--
http://larry.masinter.net