Re: "property values"

Hi Karen and thanks for this

some comments inline

On Tue, Oct 4, 2016 at 12:56 AM, Karen Coyle <> wrote:

> There are 6 instances of the use of "property values" in the SHACL spec.
> From my research, I do not believe that RDF supports the concept that
> properties have values (unlike, for example, key/value pair models). The
> object of a triple may have a literal value, but that is not what is
> intended here.

Actually there are a few hundrends, as a convention the spec uses a lot the
term "value" and links to the property value definition.

> Here are my suggestions for changes, in the order in which they appear in
> the document:
> 1 Now reads in the terminology section:
> "Property Values and Paths
> The values of (or for) a property p for a node n in an RDF graph are the
> objects of the triples in the graph that have n as subject and p as
> predicate."
> This could become "Object nodes" but in fact I think that "node n" here is
> referring to the subject node? Not sure. I would define "object nodes" and
> "property paths" separately, and the latter definition could be copied from
> SPARQL. In fact, it might be good to reference the SPARQL documentation at
> this point.

For simple properties object values is correct but for paths the node might
also be a subject node since we allow inverse relations.
The spec already reuses part of the SPARQL definition for property paths
and splitting the terms sounds good to me

> 2 Now reads:
> "Some of the property constraints specify multiple constraint components
> in order to restrict multiple aspects of the property values."
> I have no idea what "multiple aspects of the property values" means here.
> I think this sentence and the two that follow it should be changed to
> something like:
> "There can be multiple constraints directed at a single node."
> 3&4 drop
> 5 now reads:
> "It is a common scenario that certain property values are derived from
> other values. "
> In the RDF documentation, "value" is used exclusively with literals. If
> this is the case here, then it may be appropriate to refer to "literals" or
> "literal nodes" (which by definition includes only object nodes) rather
> than property values.

Looking at R2RML the term value is also used for non-literals but the
wording there is a little more verbose without needing a definition


> kc
> --
> Karen Coyle
> m: 1-510-435-8234
> skype: kcoylenet/+1-510-984-3600

Dimitris Kontokostas
Department of Computer Science, University of Leipzig & DBpedia Association
Research Group: AKSW/KILT

Received on Tuesday, 4 October 2016 07:44:08 UTC