Re: Question on closed shapes

On 5/8/2015 5:06, Arthur Ryman wrote:
> Holger,
> The text is ok but we are missing a precise spec for the triples that
> are matched by a shape.

I was hoping that my text clarifies this - if it is not precise enough 
then we need to rework it. Currently it states:

An|sh:Error|must be reported for each triple that has thefocus nodeas 
itssubjectand apredicatethat is not explicitly enumerated as 
a|sh:predicate|of the|sh:property|constraints at the surrounding shape. 
The properties|rdf:type|and|sh:nodeShape|are excluded from this 
constraint. The produced|sh:Error |must have thefocus nodeas 
its|sh:root|, and the corresponding values of the triple 

In other words, it will look at all triples that have the focus node as 
subject, except the rdf:type and sh:nodeShape triples. If any of those 
triples contains a predicate that has not been explicitly enumerated via 
sh:property/sh:predicate, then report an error for that triple.

In Example 16: A closed shape:

 a sh:Shape ;
 sh:constraint sh:ClosedShape ;
 sh:property [
  sh:predicate ex:exampleProperty1 ;
 ] ;
 sh:property [
  sh:predicate ex:exampleProperty2 ;
 ] .
 ex:exampleProperty1 ex:someValue .
 ex:exampleProperty2 ex:someValue ;
 ex:someOtherProperty 42 .

The last resource is invalid because it has a value for 
ex:someOtherProperty, which is not declared as a property.

I would like to get clarification from the ShEx people if this was the 
intention, or if this should be any more complicating.


> -- Arthur
> On Wed, May 6, 2015 at 2:12 PM, Karen Coyle <> wrote:
>> Holger, the text looks fine, but I think we need to come up with a term
>> other than "closed shape" -- it seems to me that is not going to be how most
>> users express this concept. That said, I'm struggling to come up with a
>> usable suggestion -- but I'll continue to think on it.
>> kc
>> On 4/30/15 4:46 PM, Holger Knublauch wrote:
>>> FWIW I have added some support for closed shapes to my draft
>>> To those who suggested this feature: does this look about right?
>>> The design currently excludes rdf:type and sh:nodeShape. Does that make
>>> sense or must even those properties be explicitly enumerated via
>>> sh:property?
>>> (Other interpretations of "closed" shapes could be expressed via
>>> Thanks
>>> Holger
>> --
>> Karen Coyle
>> m: 1-510-435-8234
>> skype: kcoylenet/+1-510-984-3600

Received on Thursday, 7 May 2015 22:42:48 UTC