There seems to be a multiplicity of "official" JSON standards.
The ECMA official JSON standard:
http://www.ecma-international.org/publications/files/ECMA-ST/ECMA-404.pdf
This is a subset of what's in the RFC, and shows signs of having been derived from it, though editorially modified.
The IETF official JSON standard:
http://tools.ietf.org/html/rfc4627
Technically, this is a MIME media-type registration rather than a definition of JSON. The new version being worked on in the IETF JSON WG will expand it into a definition, but will only warn about interoperability problems with numbers, unpaired surrogates in strings, and multiple occurrences of names in JSON objects rather than attempting to lay down the law. The current draft (in unofficial HTML form) is at <
https://www.tbray.org/tmp/draft-ietf-json-rfc4627bis-06.html>.
The WG may have further work product which will lay down the law on these points; it hasn't been decided yet. Anyone interested in the issues should join the JSON WG mailing list (
json-subscribe@ietf.org).
The JSON web site official JSON standard:
http://www.json.org/
Nothing official about that; it's one person's point of view, admittedly the inventor. That gives what he says a peculiar interest, but not a peculiar authority.
Elsewhere on this thread, David Lee writes:
For example I dont happen to know of any "JSON" processors that can *actually* handle integers > 53 bits.
JSON processors written in Scheme or Common Lisp do so as a matter of course, since those languages support big integers (not all Scheme implementations do, but it's common).
