W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > October to December 2009

Re: Comments on RDF Update Protocol for HTTP

From: Steve Harris <steve.harris@garlik.com>
Date: Tue, 6 Oct 2009 17:54:15 +0100
Message-Id: <BFDB6909-63C8-404C-9387-D720D8C25592@garlik.com>
To: SPARQL WG <public-rdf-dawg@w3.org>
On 6 Oct 2009, at 17:21, Chimezie Ogbuji wrote:
>
>> == 5.1 HTTP POST
>>
>> Can the POST create a graph that does not already exists?  HTTP  
>> POST is
>> neutral on this.  You can get back 201 (Created).
>
> It seems to me that this behavior is already covered by HTTP PUT.   
> I'm not
> sure, however, if that suggests that we should clearly indicate that  
> the use
> of HTTP POST on a non-existent graph is not allowed. I have added a  
> brief
> note to highlight this.

I've sent a separate mail on this topic, but I also don't feel that  
HTTP condones that behaviour. Further, it's quite inconvenient.

Imagine a service that hooks into Git pushes, and writes triples  
describing the blob pushed into a graph named after the resource URI.

With the behaviour above the required sequence would be:

    GET $resource
    IF status is 404
       PUT triples to $resource
    ELSE
       POST triples to $resource

with the potential for PUT/PUT races, which lose data, as opposed to  
just

    POST triples to $resource

Which the store has a greater chance of processing correctly, and is  
simpler.

- Steve

-- 
Steve Harris
Garlik Limited, 2 Sheen Road, Richmond, TW9 1AE, UK
+44(0)20 8973 2465  http://www.garlik.com/
Registered in England and Wales 535 7233 VAT # 849 0517 11
Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10  
9AD
Received on Tuesday, 6 October 2009 16:54:45 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:40 GMT