Re: RDF Shapes has published shacl (Shapes Constraint Language) draft

On 2015-03-22 19:00, Dietrich Schulten wrote:
>>
>> Ok, I see you answer my above question. This may be a problem indeed.
>
> Or maybe not, after you solved my @type cast problem so elegantly :)
>
>> Now, upgrading to JSON-LD for me kind of
>> implies that the underlying data model is in fact RDF and you can't
>> escape that fact.
>
> I do not want to escape that fact. What I do want is: keep the ability
> of json-ld to *interpret* communication in plain json in terms of RDF as
> much as possible.

Sure. And sorry, I replied to part of your email I only later realized 
wasn't a message you wrote directly to me anyway.

>
> If the server understood a query
>
> /orders?status=ORDER_PROCESSING
>
> until yesterday, and now it adds a Link header to its application/json
> responses which interprets them as linked data, then it does not sound
> reasonable that all client requests suddenly MUST be linked data, too,
> and that the status query value must be a URL from now on.
>
> In fact, I was assuming the opposite all the time and thought that the
> expanded form in IriTemplates is only required if the server
> specifically applies ExplicitRepresentation. Otherwise, if the server
> says "status":"ORDER_PROCESSING" in a response, then the client also may
> say ?status=ORDER_PROCESSING in a query which filters by that value.

That would be very neat. I'm not sure how the standards stand on this 
matter. For one thing I don't think that adding the JSON-LD context 
header implies that the requested URI parameters should be interpreted 
in any special way. I don't think we've stated anything like it in 
Hydra. However this is a problem you usually would face from inside of 
an API - as developer slash publisher. I think it should be safe to make 
such non-standard assumption. You won't break existing non-LD clients 
and there is no effect on LD-aware clients, which likely won't make such 
a request in the first place.

>
> I can see of course that the above URL query is not linked data and that
> a pure RDF service wouldn't know what to make of it. My understanding
> was that this is why such a service may ask clients to use
> ExplicitRepresentation.
>
> What do you think?
>
> Best regards,
> Dietrich
>
>

Received on Sunday, 22 March 2015 19:10:28 UTC