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

Re: [ACTION-33] Trying to sort the SPARQL/Update issues.

From: Steve Harris <steve.harris@garlik.com>
Date: Wed, 27 May 2009 21:32:52 +0100
Cc: Kjetil Kjernsmo <Kjetil.Kjernsmo@computas.com>, "public-rdf-dawg@w3.org" <public-rdf-dawg@w3.org>
Message-Id: <9BD3225E-F54F-4063-B810-7E29374FBC15@garlik.com>
To: "Seaborne, Andy" <andy.seaborne@hp.com>
On 27 May 2009, at 18:06, Seaborne, Andy wrote:
>> -----Original Message-----
>> From: Steve Harris [mailto:steve.harris@garlik.com]
>> Sent: 27 May 2009 17:31
>> To: Seaborne, Andy
>> Cc: Kjetil Kjernsmo; public-rdf-dawg@w3.org
>> Subject: Re: [ACTION-33] Trying to sort the SPARQL/Update issues.
>>> I don't see the HTTP protocol use as adding operations that can't be
>>> done by the language.  They should be aligned.  The language will
>>> probably be able to do more.
>> I disagree. The language can't take "local" (to the client) data in
>> RDF/XML syntax and write it to a remote store. You can do
>>   $ sparql-update http://store.example/ 'LOAD <file:///tmp/data.rdf>
>> INTO <http://example.com/data.rdf>'
> I think pushing rather hard on a design that does not fully exist  
> yet in saying "can't".  You have a suggested requirement.

Given the above situation, I don't think "can't" is too strong. The  
endpoint which is going to receive the SPARQL/Update request has a  
different context for file:///.

My requirement is to be able to PUT data to an endpoint as per RFC  
2616. It doesn't require explicit standardisation in SPARQL (I already  
do it, and it's the only way to write data into one of our stores),  
but it really wouldn't hurt.

> Not RDF/XML per se, but that's it in the area of INSERT DATA so we  
> might have an operation like LOAD INLINE (and multi part MIME? Bit  
> messy).

That would require a dedicated SPARQL/Update client, I can't trivially  
do that with the curl command line tool for example.

>> but that's not the same as
>>   $ curl -T /tmp/data.rdf
>> 'http://store.example/?graph=http%3A%2F%2Fexample.com%2Fdata.rdf'
> This requires a service to manage the store service endpoint.
> I see that as putting language into the URI query string
> 'http://store.example/?loadInto=http%3A%2F%2Fexample.com%2Fdata.rdf'
>> due to the whole client/server thing.
>> I suspect that the PUT/POST/DELETE type tuff is a more natural fit
>> into the SPARQL Protocol doc, but no strong feelings on that.
> But again, what if there is no protocol engine?

Well, that's why I think it's a better fit in the protocol doc.  
Doesn't feel like a language feature to me at all.

> A language enables "perform this script (ref file) on that  
> endpoint".  The ability to record, pass around, reply is valuable.

Sure, but the webarch fundamentals are important too. I see these as  
complimentary myself. Neither on their own address all cases.

- 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  
Received on Wednesday, 27 May 2009 20:33:28 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:00:55 UTC