Re: [Json] Minimal edit proposal, second round

["Matt Miller (mamille2)" <[e-mail]>]

On Jun 26, 2013, at 2:07 PM, Nico Williams <[e-mail]> wrote:

> On Jun 26, 2013 2:15 PM, "Tim Bray" <[e-mail]> wrote:
>> 
>> Yeah, but our charter clearly doesn’t let us fix JSON.  What we need is
> to do something like this, then re-charter to produce a doc that provides
> real guidance, so that, as someone in this thread pointed out, other
> Working Groups will have something to point to and won’t, as is currently
> happening in Jose, have to write explicit language forbidding dupes, etc,
> into their drafts.  So I’m in favor of doing this minimal
> point-out-the-problems work and moving on. -T
> 
> We could also re-charter now (or conclude w/o producing any docs, as
> someone else pointed out).  There's no dishonor in acknowledging that the
> current charter is unsatisfactory, if we so conclude.
> 
> I'm for a minimal edit that describes the different interpretations of the
> old text that we have seen.  We could also note that there is a subset of
> JSON that will (MUST) interop (no dup names, smallish integers, only
> Unicode character data and unassigned code points), though we can't rewrite
> that encoders restrict themselves to that subset.  Then we can look at
> implementation guidance (e.g., parser options) and best practices.
> 
> Nico


/me doffs hat

I think the minimal edit useful as a way to move forward: it notes the original had warts, that "properly" (for varying definitions of "properly") correcting those warts could cause worse conditions, but that there are a plethora of implementations that coexist with very few reported problems.  I think the edits do bring to the surface information that has been unspoken for a long time.  I could easily live with it, and leave the definitive "this is what JSON software ought to do from this point forward" for a separate document.


- m&m

Matt Miller < [e-mail] >
Cisco Systems, Inc.



- m&m

Matt Miller < [e-mail] >
Cisco Systems, Inc.