Re: [Json] Proposal to move forward on acceptable roots

I think non-normative (but strong) advice to use an object at the
top-level and have must-ignore semantics (unless you have a really
solid reason for doing otherwise) is an excellent idea. In particular
advising the use of strings for 64-bit integers and pushing for
ISO8601 for dates/times could help library authors converge on the
encodings they use, which would definitely help interoperability. You
could also include advice on binary encoding (presumably base-64). Not
that JSON is a great format for binary data, but sometimes that's what
you've got.

On Fri, May 30, 2014 at 9:15 AM, Tim Bray <tbray@textuality.com> wrote:
> This doesn’t sound insane. I’ll try, over the next few days, to draft such a
> section.  If nothing good emerges, I think staying with the current language
> is still a good outcome.
>
>
>
> On Fri, May 30, 2014 at 9:13 AM, Joe Hildebrand (jhildebr)
> <jhildebr@cisco.com> wrote:
>>
>> One way we could go on the topic of what goes in the root that would work
>> for me would be for us to add a non-normative section to I-JSON giving
>> protocol design advice, and move the issue of potential incompatibility of
>> non-[object,array] roots to that section.
>>
>> Into that section, we can also put some suggestions about how to encode
>> integers, floats, and dates in strings, how to do forward-compatibility
>> with objects (MUST ignore what you don't understand), and whatever other
>> pearls of wisdom we come up with.
>>
>> Yes, I know that will make the document take longer, and open up other
>> areas for us to argue about.
>>
>> --
>> Joe Hildebrand
>>
>>
>>
>> _______________________________________________
>> json mailing list
>> json@ietf.org
>> https://www.ietf.org/mailman/listinfo/json
>
>
>
>
> --
> - Tim Bray (If you’d like to send me a private message, see
> https://keybase.io/timbray)
>
> _______________________________________________
> json mailing list
> json@ietf.org
> https://www.ietf.org/mailman/listinfo/json
>