W3C home > Mailing lists > Public > public-linked-json@w3.org > August 2011

Re: IANA Considerations

From: Alexandre Passant <alex@seevl.net>
Date: Tue, 23 Aug 2011 12:15:16 +0000
Message-ID: <CALF6uCnCCgiEYnr3RMfiVwEN7L4KMysu_73fLz0rcJF=QfS9vA@mail.gmail.com>
To: Manu Sporny <msporny@digitalbazaar.com>
Cc: Linked JSON <public-linked-json@w3.org>
On Tue, Aug 23, 2011 at 4:29 AM, Manu Sporny <msporny@digitalbazaar.com> wrote:
> On 08/19/2011 09:00 AM, Markus Lanthaler wrote:
>>
>> Gregg Kellogg wrote:
>>>
>>> Also, we discussed format=context to describe the format of an @context
>>> document.
>>
>> Yes, we should add that as well. I think it would also be interesting to
>> link from a context document to a data document as Glenn proposed:
>
> Who, on this mailing list, currently needs this feature for an application
> that they're building? I'm not saying that we shouldn't do it - I'm just
> wondering who can't accomplish what they want to accomplish without this
> feature.
>
> That is, why wouldn't fetching the JSON you want to interpret as JSON-LD and
> injecting a pre-known context into the returned object work?
>
>>> On 07/06/2011 11:46 AM, glenn mcdonald wrote:
>>>>
>>>> If it's a goal to be able to impose context on an "existing" JSON
>>>> feed, it might be even more effective, instead of allowing @context
>>>> to point to an external file, to invert this and have a context file
>>>> that can point to a data file. That way the data file itself can be
>>>> completely untouched.
>>>
>>> Would something like this work for you, Glenn?
>>>
>>> {
>>>   "@context": "http://example.org/microblog",
>>>   "@data": "http://foo.bar/posts/15"
>>> }
>>
>> We could define an optional "data" or "instance" MIME type parameter for
>> JSON-LD context documents. TO link from an data/instance document to a
>> JSON-LD context document we could use a link header (RFC5988), e.g.:
>>
>> Link:<http://www.example.com/context.jsonld>; rel="describedby";
>> type="application/ld+json";
>
> This is an interesting direction. I like it because we don't have to
> introduce another keyword to JSON-LD. The down-side is that JSON REST APIs
> would have to be modified to include the Link header. That may not be
> difficult in some cases, but may be impossible in other cases. We should
> discuss it on the call.
>

+1 for this design choice.

This is also what JSON-Schema proposes to link instance data and
schema data (i.e. data and context in our case)

http://tools.ietf.org/html/draft-zyp-json-schema-03#section-4

I don't think that's a big deal to add this Link in the header. We can
probably assume that if a service is able to add a context in the JSON
file (a step that would be often required to make the JSON => JSON-LD
transition), it can also add the link in the header. It also saves
bandwidth when the context becomes bigger and developers just need the
JSON (or only fetching the context once in a while), and may be
another good step for adoption.

Alex.

> -- manu
>
> --
> Manu Sporny (skype: msporny, twitter: manusporny)
> Founder/CEO - Digital Bazaar, Inc.
> blog: Uber Comparison of RDFa, Microformats and Microdata
> http://manu.sporny.org/2011/uber-comparison-rdfa-md-uf/
>
>



-- 
Dr. Alexandre Passant - @terraces
Founder, CEO - seevl.net - @seevl
Reinventing Music Discovery
Received on Tuesday, 23 August 2011 12:15:46 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:35 GMT