Re: [Json] Proposal: the minimal edit

[Paul Hoffman <[e-mail]>]

On Jun 24, 2013, at 8:20 PM, John Cowan <[e-mail]> wrote:

> Paul Hoffman scripsit:
> 
>> Add Section 1.2, "Changes from RFC 4627"
>> 
>>   This section lists all changes between this document and the text in
>>   RFC 4627.
>> 
>>   - Applied errata #607 from RFC 4627 to correctly align the artwork
>>     for the definition of "object".
>> 
>>   - Applied errata #3607 from RFC 4627 by removing the security
>>     consideration that begins "A JSON text can be safely passed"
>>     and the JavaScript code that went with that consideration.
>> 
>>   - Added Section 1.3, "Differences from the JSON Definition in ECMAScript". 
>> 
>>   - Changed the [ECMA] and [UNICODE] references to be non-version-specific.
> 
> Given these changes, and *only* these changes, then I do not believe that
> 
>>   - ECMAScript implementations can generate and consume code points
>>     in JSON strings that are not Unicode characters.
> 
> belongs in Section 1.3, for it is equally true of ECMA-JSON and RFC 4627
> JSON.  Both of them allow code points that are not Unicode characters,
> e.g. unassigned codepoints.  So this point is true but not relevant.

Hrm. Either you are focused only on the grammar section and not the whole document, or you missed the Introduction to RFC 4627, which say unequivocally:

   A string is a sequence of zero or more Unicode characters [UNICODE].

That statement disagrees with ABNF in the grammar, so the RFC is ambiguous. Some people in the WG believe that this sentence is right and the ABNF is wrong; others believe the other.

Does that help make the sentence about the differences more believable?

--Paul Hoffman