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

Re: Questions about Update 1.1

From: Steve Harris <steve.harris@garlik.com>
Date: Fri, 16 Oct 2009 16:35:45 +0100
Message-Id: <6C6006EB-5165-4AD1-951C-7D63E2CC3388@garlik.com>
To: "public-rdf-dawg@w3.org Group" <public-rdf-dawg@w3.org>
On 16 Oct 2009, at 16:14, Birte Glimm wrote:
> [snip]
>>>> 2) Can you delete bNodes with DATA? DELETE DATA { <a> <b> _:z }
>>>> doesn't really make much sense. Some stores (including Jena and
>>>> 4store) allow you to quote bNodes like DELETE DATA { <a> <b>  
>>>> <_:z> }
>>>> but that's non-standard.
>>>
>>> Yes - this is more of a problem now we have update and not just  
>>> query.
>>>
>>> We could document the <_:...> usage.
>>
>> Garlik would be in favour of that. We've found it invaluable when  
>> dealing
>> with FOAF data in particular.
>> However, I'm not quite sure how it fits in with the semantics of  
>> RDF. It's a
>> form of skolemisation I suppose, in our case it just externalises  
>> what's
>> happening inside the store.
>
> Hm, it could be quite difficult for users to know what the actual
> bnode IDs are because you might have to rename them in a merge. If I
> think ahead to OWL where you have imports, we just name them all apart
> in any case, so if a user asks our OWL reasoner to load some ontology
> that contains <a> <b> _:z, it will end up as <a> <b> _:genID_1 in the
> reasoner. Now the user would first have to query and hope that the
> query reveals the internal bnode name and then it can be deleted.
> It does not really speak against the usage of <_:z> it would just be
> hard to use with systems that freely rename bnodes, which is perfectly
> ok for a system to do. So I am not totally against it, but it might
> cause confusion for users because the system behavior is hard to
> predict.

Yes, agreed. A similar thing happens in 4store, but for different  
reasons. The only promise we make is that if you get a bNode ID out in  
SPARQL results, then you can reuse that ID in <_:$id> until you do an  
update on the graph it came from, at which time it may no longer be  
valid.

I agree that this is tricky to define, and has potential for confusion  
though.

- 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 Friday, 16 October 2009 15:36:18 GMT

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