W3C home > Mailing lists > Public > public-ldp@w3.org > March 2012

Re: defining RDF and RESTful

From: Dave Reynolds <dave.e.reynolds@gmail.com>
Date: Thu, 29 Mar 2012 22:11:20 +0100
Message-ID: <4F74CFF8.2060703@gmail.com>
To: Sandro Hawke <sandro@w3.org>
CC: public-ldp@w3.org
On 29/03/12 21:51, Sandro Hawke wrote:
> On Thu, 2012-03-29 at 21:35 +0100, Dave Reynolds wrote:
>> On 29/03/12 21:25, Sandro Hawke wrote:
>>> On Thu, 2012-03-29 at 10:16 -0700, ashok malhotra wrote:
>>>> On 3/29/2012 6:15 AM, David Wood wrote:
>>>>> I suggest that the Callimachus REST API be an input document to the LDP WG.
>>>> That would be great!
>>> Okay, since that's seconded, I've added a references section.  I also
>>> took out the explicit mention of GET and PUT.
>> The phrasing in the introduction:
>> [[[
>> This approach has been proposed and explored for some time [1][2][3][4],
>> but its use has only recently been reported in industry [5] [6].
>> ]]]
>> makes it sound like 1-4 have not been used in industry. In fact a number
>> of commercial organizations are using the Linked Data API [4], including
>> in media and financial sectors as well as for-profit organizations
>> serving the public sector.
> I did it that way because I was using the LDAPI as evidence that the
> approach had been "explored" (and generally to mention it).  I didn't
> think it was fair to claim that LDAPI demonstrated reading and writing
> RDF through a RESTful API, since (as I recall), clients see JSON not
> RDF.    But please correct me, if I've got that wrong.

No, clients see RDF (in both RDF/XML and Turtle formats).

It is true that they can *also* ask for the RDF to be translated to, and 
returned as, simple JSON/simple XML/CSV/etc.

The RESTful API machinery for paging, filters, views, text search etc 
all apply independent of whether you ask for RDF or JSON. The modelling 
of collections is all in RDF and returned in the RDF. All of the 
metadata which allows you to discover alternative views and API 
capabilities are expressed and returned in RDF.

It *is* true that the current spec only defines read usage, not writing. 
[Though it is quite straightforward to extend the list endpoints to 
support the standard REST approach to collections.]

Received on Thursday, 29 March 2012 21:11:51 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:16:33 UTC