W3C home > Mailing lists > Public > public-prov-wg@w3.org > November 2011

Re: PROV-ISSUE-145 (Tlebo): qualified identifiers may not work well with named graphs [Data Model]

From: Graham Klyne <GK@ninebynine.org>
Date: Sun, 13 Nov 2011 11:50:13 +0000
Message-ID: <4EBFAEF5.6030901@ninebynine.org>
To: Luc Moreau <L.Moreau@ecs.soton.ac.uk>
CC: public-prov-wg@w3.org
On 10/11/2011 17:13, Luc Moreau wrote:
> Hi Graham,
>
> When you say 'documents', do you mean accounts?

No.  I was talking generically about "documents" with a formal semantic 
interpretation, and more specifically I meant resources containing provenance 
information.

Reflecting on this, and qualifying the comment: it applies to using RDF without 
named graphs.  If named graphs are used then whether or not it applies will 
depend on the formal semantics of named graphs.

Generally, though, I think it is not a good idea to allow different accounts to 
use the same URI for different entities.  While accounts may contain statemenbts 
that are specific to the account, they should also provide for inferences about 
things (specifically, Entities) that hold outside the context of an account; 
e.g. entity1 derivedfrom entity2, if true, should be true independently of any 
account considered.

#g
--

We are not saying
> that two different accounts have to be combined. On the contrary,
> they exist, but we may have to establish relations between the 'provenance records'
> we made in these accounts. (wasComplementOf is an example of this)
>
> Luc
>
> On 11/10/2011 03:45 PM, Graham Klyne wrote:
>> On 07/11/2011 08:57, Luc Moreau wrote:
>>> Hence, two different entities could be asserted by two different asserters in
>>> different accounts,
>>> and both may have chosen the same URI to identify them.
>>
>> That would be contradictory at the level of RDF.
>>
>> I think any attempt to combine documents with such usage would render the
>> result formally meaningless (or at least fail to formally convey the intended
>> meaning).
>>
>> #g
>> --
>>
>>
>>> If we want to avoid this, then it must be a requirement that new URIs are minted
>>> for all
>>> entities. But I thought the WG wanted to move away for that, since it prevents
>>> lightweight assertions
>>> of provenance.
>>>
>>> So, the minting burden, i think, is put at the level of accounts.
>>> BTW, prov-dm is currently not suggesting that, since it says that account
>>> identifiers are also scoped,
>>> but I think this is not good.
>>>
>>> Luc
>>>
>>> On 11/06/2011 11:49 PM, Tim Lebo wrote:
>>>> Hi, Luc!
>>>>
>>>> Sent from my iPhone
>>>>
>>>> On Nov 6, 2011, at 17:44, Luc Moreau<L.Moreau@ecs.soton.ac.uk> wrote:
>>>>
>>>>> Hi Tim,
>>>>>
>>>>> We need to explore this in detail.
>>>>> To me, it is crucial to be able to assert that entity e1 in account acc1 is
>>>>> complementOf entity e2 in acc2.
>>>> Absolutely.
>>>>
>>>>> How do you propose doing this?
>>>> Would the one triple
>>>>
>>>> :e1 prov:wasComplementOf :e2 .
>>>>
>>>> work? This would stand independent of what account the entities are in --
>>>> which gets to the point I was making that the asserter needs to make distinct
>>>> URIs and not depend on some account-based scoping mechanism to establish its
>>>> identity.
>>>>
>>>> Perhaps I am misinterpreting DM again from a "too RDF" perspective like the
>>>> literal versus URI discussions earlier today.
>>>>
>>>> Best,
>>>> Tim
>>>>
>>>>>
>>>>>
>>>>> Professor Luc Moreau
>>>>> Electronics and Computer Science
>>>>> University of Southampton
>>>>> Southampton SO17 1BJ
>>>>> United Kingdom
>>>>>
>>>>> On 6 Nov 2011, at 21:24, "Timothy Lebo"<lebot@rpi.edu> wrote:
>>>>>
>>>>>> By "appropriately scoped", I mean "predefined, consciously selected; URIs".
>>>>>>
>>>>>> the note "refer to an identifier in the scope of a given account" sounds
>>>>>> like we are going to permit lazy naming that can be computed in the future,
>>>>>> which current named graph implementations do not support.
>>>>>>
>>>>>> -Tim
>>>>>>
>>>>>>
>>>>>>
>>>>>> On Nov 6, 2011, at 2:32 AM, Luc Moreau wrote:
>>>>>>
>>>>>>> Hi Tim,
>>>>>>>
>>>>>>> What do you mean by appropriately scoped?
>>>>>>>
>>>>>>> Luc
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> I propose that we require the asserters to define appropriately-scoped URIs
>>>>>>> for their identifiers. Letting them be lazy up front will cause headaches
>>>>>>> when actually trying to use it.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>
>>
>
Received on Sunday, 13 November 2011 12:25:26 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:51:04 UTC