Re: Properties v classes in validation

Holger Knublauch <holger@topquadrant.com> wrote on 08/31/2015 04:12:51 PM:

> From: Holger Knublauch <holger@topquadrant.com>
> To: public-data-shapes-wg@w3.org
> Date: 08/31/2015 04:14 PM
> Subject: Re: Properties v classes in validation
> 
> 
> 
> On 9/1/15 1:24 AM, Karen Coyle wrote:
> > I'm not sure that this is an actual issue, so I thought I'd ping the 
> > group before making it official...
> >
> > Many of the validation requirements coming out of the cultural 
> > heritage community are conceived as requirements on "things" (defined 
> > as rdf:type classes) not on properties. As an example, a rule might 
be:
> >
> > for every subject of type ex:CulturalObject
> > -- there can be one or more subjects in the same graph of type 
ex:Person
> 
> As written above, I see no constraint that could be evaluated. Probably 
> you wanted to say "there MUST be one or more subjects..."? If yes, is 
> there a triple/relationship between them, but the problem is that you 
> don't know which property?

I have to say that I'm not sure I completely understand either. Karen, are 
you saying that for every X rdf:type ex:CulturalObject there can be one or 
more X rdf:type ex:Person ? Does that mean you really want to enforce that 
for every X rdf:type Person there must be one X rdf:type ex:CulturalObject 
?

> 
> >
> > There are also rules regarding what rdf:type(s) are allowed in general 

> > in a graph. (This would be a closed shape.)
> 
> I believe this could be expressed (albeit a bit geeky) with an inverse 
> property constraint on rdf:type:
> 
> ex:MyShape
>      a sh:Shape ;
>      sh:scope [
>          a sh:InversePropertyScope ;
>          sh:predicate rdf:type ;
>      ] ;
>      sh:inverseProperty [
>          sh:predicate rdf:type ;
>          sh:allowedValues ( ex:Class1 ex:Class2 ex:Class3 )
>      ] .

And this has to be set as a Graph-level/global constraint, right?

> 
> Holger
> 
> >
> > SHACL, however, has property validation rules, but no class validation 

> > rules. In some cases in CH data, there is a single property that 
> > connects the subjects, but for example in the case of cultural 
> > resources there are literally hundreds of different properties that 
> > can link a person to an object.

It sounds like this is something a Graph-level/global constraint, should 
be able to address but I don't understand enough. An example would help.

> >
> > It's possible that we just need to adjust our thinking, but I'd like 
> > to hear if others have similar situations with their data.
> >
> > Thanks,
> > kc
> 
> 


--
Arnaud  Le Hors - Senior Technical Staff Member, Open Web Technologies - 
IBM Software Group