Re: Checkout API spec published

> On Feb 8, 2016, at 9:05 PM, Manu Sporny <msporny@digitalbazaar.com> wrote:
> 
>> On 02/08/2016 04:33 PM, Ian Jacobs wrote:
>> It is not appropriate to have a normative reference to material that 
>> is not required for conformance.
> 
> It's perfectly appropriate to have normative references to material that
> tells people how to interpret/extend messages using JSON-LD if they want
> to use JSON-LD. If you're not going to interpret the messages as
> JSON-LD, you can ignore the normative text. You include text like this
> in the API specs so people know that they have the option to interpret
> the data as JSON-LD.

The external spec itself could have normative language. That does not mean that the base spec needs to have a normative reference to it, especially if there is no dependency.

>> I do not believe we understand the ecosystem well enough to require 
>> JSON-LD.
> 
> Why do you think we're "requiring JSON-LD"? If you don't have a
> normative extensibility mechanism, then what's the extensibility
> mechanism you're proposing?

I am saying I think we need more time to understand the ecosystem. I don't have a proposal other than to get more information before making requirements.

>> I support the idea of a standalone specification as a way to build 
>> the conversation around the use of JSON-LD.
> 
> Is that specification REC-track?

Personally I prefer a Note until we have a better sense, but I can live with rec track.

> Does it contain normative language on
> how to interpret the messages as JSON-LD?

I do not think we will achieve Interop unless we have a better sense of the ecosystem and its stakeholders. Until then, I am more comfortable saying "here is one way to do things" rather than  "here is the only way." It may be that your proposal accomplishes that; I need to study it more.

Ian