Re: ODP: ODP: Re: Replace hydra:Error with application/problem+json

Hi, 

Comments inline. 
Am 28.09.2015 13:46 schrieb Asbjørn Ulsberg <asbjornu@gmail.com>:
>
> Yes, that’s fine. I just don’t see the need to explain to anyone that application/problem+json is based on JSON-LD if the spec is rewritten to be based on JSON-LD. In that case, it will be a given just due to its content type (and the RFC explaining the fact). 

A json-ld file is json with a linked-data context. Without a context, the problem+json attributes are invisible to a json-ld processor. That is by specification,  "JSON keys that do not expand to an IRI, (...) are not Linked Data and thus ignored when processed." 

The context can be linked in via a Link header or it can be embedded with @context, in which case the media type *must* be application/ld+json, otherwise the semantics of @context is undefined. In problem+json there is no @context attribute, and there is no RDF vocab yet which defines its attributes.

> So, yes, we can always bridge the gap between the current spec and JSON-LD, but is that necessary? Is anything lost if application/problem+json is rewritten so it is JSON-LD compatible out of the box?

What do you have in mind when you say "rewrite so it is compatible"? 

Trouble is that the problem+json attribute semantics has not yet been published in a way that can be mapped to RDF by a machine. There is a relax-ng schema, but unlike xml schema with id attributes, relax-ng cannot be re-used for RDF (afaik). Just stating that the rfc is also an RDF vocabulary is not sufficient, it must be published and the json-ld response must point to it somehow.

The media types ld+json and problem+json are mutually exclusive, I am afraid there is no third way. 

>
> -- 
> Asbjørn Ulsberg          -=|=-         asbjorn@ulsberg.no
> «He's a loathsome offensive brute, yet I can't look away»
>
>
> On Mon, Sep 28, 2015 at 1:43 PM, John Walker <john.walker@semaku.com> wrote:
>>
>> Hi Asbjørn,
>>  
>>>
>>> On September 27, 2015 at 11:37 PM Asbjørn Ulsberg <asbjornu@gmail.com> wrote:
>>>
>>> Why do we need both? Can’t application/problem+json be JSON-LD compatible without announcing it in its Content-Type? Or is it expected that everything that is JSON-LD compatible will use application/ld+json as its content type?
>>
>> You can just add a Link header to the response to reference to the context that can be used to interpret an ordinary JSON documents as JSON-LD.
>>
>> You can find more info abotu this in the JSON-LD rec:
>>
>> http://www.w3.org/TR/json-ld/#interpreting-json-as-json-ld
>>
>> Regards,
>>
>> John
>>>
>>> -- 
>>> Asbjørn Ulsberg          -=|=-         asbjorn@ulsberg.no
>>> «He's a loathsome offensive brute, yet I can't look away»
>>>
>>>
>>> On Sun, Sep 27, 2015 at 9:38 PM, Karol Szczepański <karol.szczepanski@gmail.com> wrote:
>>>>
>>>> OK, I see your point now.
>>>>
>>>> Server may take into account accepted media types client sent and choose between RDF (if applicable) or not. It still may ignore it and consider a response as a subject not for content negotation and come with bare ‘problem’ media type.
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> Regards
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> Karol Szczepański
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>
>>>> Od: Dietrich Schulten
>>>> Wysłano: niedziela, 27 września 2015 15:38
>>>> Do: public-hydra@w3.org
>>>> Temat: Re: ODP: Re: Replace hydra:Error with application/problem+json
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> Hi Karol,
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> that was a misunderstanding.
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> Am 27.09.2015 um 13:56 schrieb Karol Szczepański:
>>>>
>>>> > Hi
>>>>
>>>> >
>>>>
>>>> > Your approach would put non-RDF clients to problem as you suggest to
>>>>
>>>> > respond with application/ld+json instead of application/problem+json.
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> That was not my intention at all.
>>>>
>>>> The idea is: *If* a client says in its POST that it accepts
>>>>
>>>> application/ld+json, then the service might respond with
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> Content-Type: application/ld+json
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> If another client comes to the same service and accepts
>>>>
>>>> application/json, then the server might respond with
>>>>
>>>> application/problem+json, similarly for XML.
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> The rules of conneg apply, of course. The server may respond whatever it
>>>>
>>>> wants to.
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>> Best,
>>>>
>>>> Dietrich
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>>>
>>>>  
>>
>>
>>  
>
>