Re: PROV-ISSUE-230 (Name-scoping): Name scoping in DM is wrong concept [prov-dm]

Hi Graham,

On 01/30/2012 11:01 AM, Provenance Working Group Issue Tracker wrote:
> PROV-ISSUE-230 (Name-scoping): Name scoping in DM is wrong concept [prov-dm]
> Raised by: Graham Klyne
> On product: prov-dm
> The PROV-DM draft introduces name scoping, particularly with respect to Accounts.
> I think this is the wrong concept, as it tries to use name-scoping to capture different provenance accounts about the same entity.  I think that an entity id should refer to the same Entity wherever it occurs.  What may vary between accounts is the claims that are made about that entity.  Without this, I see no basis for comparing accounts.
> For PROV-DM, I imagine one one might say that the account+local id together form the common identifier (ala compound key), but then I think some additional mechanism would then be needed to link names from different accounts.
> When the names used are URIs, then I think that the notion of scoping is entirely wrong.  URIs are, by design, a *global* namespace, and it creates confusion (or worse) of one allows a URI to denote different things.  Personally, I would not prescribe the form of names used by the DM; the use of URIs is a syntactic matter, and as such it could be introduced for ASN.

What is scoped is what we say about entities: e.g.

- what is said about entity e1 in acc1 differs/is simlar/is 
specialisation of /... of what is said about entity e2 in acc2.
- the entity e1 as described in acc1 was generated by the activity a2 as 
described in acc2

We may not like the notion of scope but recognizing that account+local 
id  forms a key, in effect, constitutes a notion of scope.
The crucial thing is to be clear about what is scoped (and I agree that 
the document is not clear, given all the contradictions
about identifiers I have raised and that we are trying to address now, 
cfr ISSUE-183).

> I see the DM as an "abstract syntax" in the sense proposed by John McCarthy, where the terms and productions have the form of logical predicates, and in particular a "name" is distinguished simply as a predicate "Name(id)" which is True iff "id" is a name.  This avoids any need to prescribe the actual form of referenced by the DM.

BTW, the DM states that names are 'qualified names': I think this is 
important because, namespacing is part of the abstract model I believe.


Professor Luc Moreau
Electronics and Computer Science   tel:   +44 23 8059 4487
University of Southampton          fax:   +44 23 8059 2865
Southampton SO17 1BJ               email:
United Kingdom           

Received on Tuesday, 31 January 2012 09:29:39 UTC