Re: PROV-ISSUE-142 (Tlebo): Can roles only be Literals? [Data Model]

Hi Stian,
Answer/comments appear below.

On 12/01/2011 10:52 AM, Stian Soiland-Reyes wrote:
> On Wed, Nov 30, 2011 at 13:06, Luc Moreau<L.Moreau@ecs.soton.ac.uk>  wrote:
>    
>> We believe that we have addressed this issue. The latest WD indicates that
>> URI can be expressed
>> using *prov-dm literal* but their interpretation is outside the scope of
>> prov-dm.
>>      
> I've checked http://dvcs.w3.org/hg/prov/raw-file/tip/model/ProvenanceModel.html#record-literal
>
> In general this looks good.
>
> Grammar: "@ (U+40) by followed a nonempty sequence" ->  ".. followed by a .."
>
>
> I like "In particular, a PROV-DM Literal may be a IRI-typed string
> (with datatype xsd:anyURI); such IRI has no specific interpretation in
> the context of PROV-DM."
>
>
> However there is also in the example:
>
>    
>> The following example shows a literal of type xsd:QName (see QName [XMLSCHEMA-2]). The prefix ex must be bound to a namespace declared in the record container.
>>    "ex:value" %% xsd:QName
>>      
> Here there seems to be a particular interpretation for QNames (which
> seems to match typical XML usage) - should this not be stated outside
> the example?
>    

> Do such QName literals have any further interpretation in PROV-DM,
> such as when used with prov:type or prov:role - or when the namespace
> name resolves to<http://www.w3.org/ns/prov-dm/>?
>
>    

good point, the text is not specific here. I think it should be.

Above this text, there is an example, with a QName:

    agent(ag, [prov:type="prov:SoftwareAgent" %% xsd:QName])

We want the prefix prov to be recognized as a prefix declared in the 
container.

> (I assume that in PROV-DM to replace such QName literals with
> xsd:anyURI literals would give a different interpretation)
>
>
>
> In the PROV-ASN syntax:
>
>    
>> typedLiteral ::= quotedString %% datatype
>> datatype ::= IRI
>> IRI ::= an IRI compatible with production IRI in [IRI]
>>      
>    

> However most examples here and elsewhere in the document use QNames or
> CURIEs for the datatype - which are not covered by [IRI]:
>
>    
>>   "http://example.org/foo" %% xsd:anyURI
>>      
>    

I don't follow you, what do you mean not covered by IRI?

> We will have to choose one or introduce extra syntax for one of them,
> because some (but not many) IRIs can also be interpreted as QNames,
> for instance:
>
> skype:soiland
> h323:example.com
> news:example.group.this
>
>
> (To be fair all QNames can be parsed as anyURIs - but semantically
> they are not anyURIs)
>
>    

We are not saying that QNames are URIs, far from it. We allow for asserters
to use QNames and URIs, according to their preference.

In Prov-dm, identifiers and attributes are defined as qualified names, 
and for these,
we say: "A qualified name can be mapped into an IRI by concatenating the 
IRI associated with the prefix and the local part".


Luc


-- 
Professor Luc Moreau
Electronics and Computer Science   tel:   +44 23 8059 4487
University of Southampton          fax:   +44 23 8059 2865
Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
United Kingdom                     http://www.ecs.soton.ac.uk/~lavm

Received on Thursday, 1 December 2011 11:39:54 UTC