Re: "RDF node" and "node" in SHACL document

Very early on in the document there is

(In this document, the verbs specify or declare are sometimes used to express
the fact that a node has property values in a graph.)

This looks unproblematic, but would be problematic if there are any cases of
wording like

.... if n declares ...

where n might not be a node in the graph in question.

Another pass is needed to flush out any other of these cases.

peter



On 12/11/2016 10:38 PM, Holger Knublauch wrote:
> Hi Peter,
> 
> I have gone through the whole document, searching and updating all usages of
> "node":
> 
> https://github.com/w3c/data-shapes/commit/3ce4495f47c91c711e99968b6ce2ee5b763193e5
> 
> 
> The intent of the terminology is such that
> - "node" is only used if it's clear from the context that the node must exist
> in the graph
> - "RDF term" is used more generally, e.g. when talking about targets
> - "focus node" is used to refer to RDF terms in the context of a data graph
> 
> I have switched to "value node" in many cases to avoid the general term "node"
> (value nodes may be focus node, which may not exist in the graph...)
> 
> It may be beneficial for another pair of eyes going through these changes to
> make sure it's more precise now.
> 
> Theoretically we would need to rename "focus node" to "focus term", but I
> believe the term node is more widely understood than term and having focus
> nodes that are not also nodes in the graph is a very rare corner case.
> 
> Thanks,
> Holger
> 
> 
> On 12/12/2016 8:06, Peter F. Patel-Schneider wrote:
>> The SHACL document uses "RDF node" in several places.  However, RDF node is
>> not a term defined in RDF.  The replacement should probably be "RDF term".
>>
>> In other places, only "node" is used, in particular at the beginning of
>> section 3.  "Node" only has a definition in RDF in the context of a particular
>> graph.  This causes problems for targets that are not nodes in the data graph,
>> as in
>>
>> Data graph
>>
>> ex:i1 rdf:type ex:c .
>>
>>
>> Shapes graph
>>
>> se:rdf:type sh:Shape ;
>>    sh:targetNode ex:i2 ;
>>    sh:class ex:c .
>>
>> Some occurrences of "node" should probably be replaced with "RDF term".
>>
>>
>>
>> Peter F. Patel-Schneider
>> Nuance Communications
>>
> 
>

Received on Monday, 12 December 2016 17:37:16 UTC