Re: [Json] Consensus call: establishing name equality

[Jacob Davies <[e-mail]>]

Well, even given those givens, what about this:

http://www.unicode.org/versions/Unicode6.2.0/ch03.pdf
>From Section 3.2:

C6 A process shall not assume that the interpretations of two
canonical-equivalent character sequences are distinct.

• The implications of this conformance clause are twofold. First, a
process is never required to give different interpretations to two
different, but canonical-equivalent character sequences. Second, no
process can assume that another process will make a distinction
between two different, but canonical-equivalent character sequences.

• Ideally, an implementation would always interpret two
canonical-equivalent character sequences identically. There are
practical circumstances under which implementations may reasonably
distinguish them.

---

That argues pretty strongly against specifying name normalization
behavior for implementations, I would think.

Back on the point of what this is for: yes, JSON names in objects are
likely used as hash keys. But the specification itself does not
*require* that they be unique (it advises, yes) and therefore no
absolutely unambiguous definition of what "unique" means is necessary
within the specification.

The specification is already clear that a name is a string, and that
backslashed characters in JSON-encoded strings are to be interpreted
in certain ways to produce a native string; how to hash and compare
strings is really a platform function, not a decoding function. It
seems unnecessary to say that the four given strings for the slash
character are equivalent: if you have followed the instructions on
interpreting strings, they already will be by the time you get around
to comparing them.



On Fri, Jun 21, 2013 at 4:21 PM, Tim Bray <[e-mail]> wrote:
>
> On Fri, Jun 21, 2013 at 4:15 PM, Jacob Davies <[e-mail]> wrote:
>>
>>
>> Is this definition of equality needed elsewhere in the document though? Standing alone it seems a little odd - there's no discussion how to compare two numbers, two objects, or two lists for equality, so why are strings-used-for-names special?
>
>
> Because
> - We talk about duplicates so it's important to define what duplicate means
> - they are, in effect, hash-table keys, so comparison really matters
>