[apps-discuss] draft-ietf-appsawg-json-pointer-00 feedback

[Julian Reschke <[e-mail]>]

Hi Paul,

below some feedback:

Section 3:

>    ABNF syntax:
>
>    json-pointer = *( "/" reference-token )
>    reference-token = *( unescaped / escaped )
>    unescaped = %x00-2E / %x30-5B / %x5D-10FFFF

Is 0 really allowed? (I see JSON allows this as well; maybe that's 
something that needs to be discussed in the security considerations).

>    escaped = "\" ( "/" / "\" )
>
>    It is an error condition if a JSON Pointer value does not conform to
>    this syntax.

I'm fine with this because it allows using \ to escape more values in 
the future.

Section 4:

>       If the currently referenced value is a JSON array, the reference
>       token MUST contain an unsigned base-10 integer value, and the new
>       referenced value is the array element with the zero-based index
>       identified by the token.

Do we want to say something about leading zeroes in index values? I 
assume they are allowed?

Section 5:

>    A JSON Pointer MAY be represented in a JSON string value.  Per
>    [RFC4627], section 2.5, all instances of quotation mark '"' (%x22),
>    reverse solidus '\' (%x5C) and control (%x00-1F) characters MUST be
>    escaped.

The MAY is a statement of fact, not a normative requirement. Just say "can".

Section 6:

>    A JSON Pointer MAY be represented in a URI fragment identifier.  The
>    JSON pointer MUST be UTF-8 [RFC3629] encoded as octets; octets in the
>    URI "unreserved" set SHOULD be percent-encoded, per [RFC3986],
>    section 2.5.

See above.

Appendix A:

The semantics of fragment identifiers depends on the media type. You 
need to be clear whether you're trying to define the fragid semantics 
for application/json (which as far as I recall doesn't define any), or 
something else.

That being said:

>    http://example.com/example.json#
>       Resolves to the object value at the root of the JSON text
>       document.

The same would be true for

   http://example.com/example.json#/

right? I think we should stay away from empty fragment identifiers if we 
can. (There MAY be dragons here).

Best regards, Julian