Re: [Json] Media types, extensibility in draft-ietf-json-i-json-02

["Joe Hildebrand (jhildebr)" <[e-mail]>]

That's what I'd want you to do as a consumer of my service.  In a perfect
world, ECMAscript would have a third parameter on JSON.parse that was
either a boolean controlling IJSON mode (defaulted to the old behavior),
or the second parameter could either be the existing reviver function or
an object with reviver:function, ijson:bool, and whatever else needs to be
added later.  Other language/library mechanisms would then follow suit
eventually.

On 7/4/14, 8:04 PM, "Tim Bray" <[e-mail]> wrote:

>So, would an I-JSON media type help in producing the desired effect? Not
>obvious to me.  In practice, I’m supposing that if I’m writing a consumer
>for JIP (Joe’s Internet Protocol) I’d toggle my JSON parser
> into I-JSON mode because I know that’s what the RFC said.
>
>
>
>On Fri, Jul 4, 2014 at 6:30 PM, Joe Hildebrand (jhildebr)
><[e-mail]> wrote:
>
>There are times when, in the documentation for a service that is producing
>JSON, I would like to say:
>
>"If you receive a message from me that is not strictly I-JSON, I would
>prefer you treat it as an error on my part, and therefore treat my message
>as invalid, because if something went wrong enough that what I sent you
>wasn't valid I-JSON my server room is likely on fire."
>
>
>On 7/2/14, 1:42 PM, "Mark Baker" <[e-mail]> wrote:
>
>>On Wed, Jul 2, 2014 at 11:54 AM, Tim Bray <[e-mail]> wrote:
>>> I’ll say that just on Web-architectural grounds, I think distinct data
>>> formats should have distinct Internet Media Types, and so it bothers me
>>>that
>>> we don’t have one for i-json.  But the WG couldn’t perceive any real
>>>value
>>> in having one that’s distinct from JSON’s and I didn’t have  a
>>> forceful-enough argument to move the consensus on this.
>>
>>Were there driving use cases behind I-JSON? I could imagine a
>>(atypical) scenario where it's used as a "publishing profile",
>>defining a best practice for publishers (Postel-ian style), with no
>>regard for what happens once it's consumed. You wouldn't need to say
>>anything about media types then.
>>
>>But if you want to be able to exchange I-JSON over the Web or
>>Internet, so that recipients know, e.g., not to add duplicate keys,
>>then that can only practically be indicated with a new media type.
>>
>
>
>>_______________________________________________
>>json mailing list
>>[e-mail]
>>https://www.ietf.org/mailman/listinfo/json
>>
>
>
>--
>
>
>Joe Hildebrand
>
>
>
>
>
>
>
>
>
>
>-- 
>- Tim Bray (If you’d like to send me a private message, see
>https://keybase.io/timbray <https://keybase.io/timbray>)
>
>
>


-- 
Joe Hildebrand