> On Jan 11, 2017, at 09:00, Martynas Jusevičius <marty...@graphity.org> wrote:
>
> Paul,
>
> isn't that what SPARQL CONSTRUCT is doing?
That just gives you another graph, not a hierarchical structure as the OP
wanted. Other than named graphs, RDF has no notion of containment, which is
essential for creating meaningful displays of information. Nor is it good for
sequencing, another essential feature for good information display. A general
graph transformation language should make it easy to specify an output
information structure built around containment and sequence (e.g., XML).
Regards,
--Paul
>
>> On Wed, Jan 11, 2017 at 3:18 PM, Paul Tyson <phty...@sbcglobal.net> wrote:
>>
>>
>> On Jan 10, 2017, at 18:06, Grahame Grieve
>> <grah...@healthintersections.com.au> wrote:
>>
>>>>
>>>> statements to describe the graph itself, e.g. here's a graph which
>>>> identifies a particular person as the foaf:primaryTopic of the graph:
>>>>
>>>> {
>>>> "@context": {
>>>> "Person": "http://xmlns.com/foaf/0.1/Person";,
>>>> "name": "http://xmlns.com/foaf/0.1/name";,
>>>> "primaryTopic": "http://xmlns.com/foaf/0.1/primaryTopic";
>>>> },
>>>>
>>>>
>>>
>>> thanks. Is it expected that every domain application will come up with
>>> their own (different) way to do this?
>>
>> Wrong list for this suggestion, but what is missing from the w3c semantic
>> web stack is a general graph transformation language, analogous to XSLT,
>> that would allow one to write domain-specific transformations from graph to
>> hierarchy. Perhaps it was imagined that RDF/XML plus XSLT would fill this
>> role, but that is painful. I've written srx-to-*ml transforms, but that is
>> awkward also. Of course you can always roll your own rules in JavaScript,
>> but that hides business logic in imperative code.
>>
>> Regards,
>> --Paul
>>>
>>> Grahame
>>
