Discovering the full schema.org vocabulary (Was: Re: Schema.org site updates: Added RDFa and JSON-LD incoming properties)

Hi Kingsley,

On Sat, Apr 5, 2014 at 7:27 PM, Kingsley Idehen <kidehen@openlinksw.com>wrote:

>  On 4/5/14 5:04 AM, Niklas Lindström wrote:
>
> For me, I would think that each term might reference the vocabulary
>> document somehow (perhaps rel=describedBy?). It would also be useful if
>> going to http://schema.org/ would somehow allow discovery of the
>> vocabulary.
>
>
>  I agree, that would be nice. I'm sure rdfs:isDefinedBy [1] is the right
> property for that. :)
>
>
> An rdfs:isDefinedBy relation associates Properties and Classes with the
> vocabulary in which they are defined.
>
> An wdrs:describedby (equivalentPropertyOf xhv:describedBy and
> inversePropertyOf xhv:describes) relation associates Properties and Classes
> with a Vocabulary document (e.g.,
> <http://schema.org/docs/schema_org_rdfa.html><http://schema.org/docs/schema_org_rdfa.html>)
> that is comprised of statements describing the aforementioned Properties
> and Classes.
>

Yes, that is true. I was rather vague in my statement about the "right"
property here. I was thinking more of linking the classes and properties
(in the dedicated pages describing them) to the vocabulary itself. I am not
sure whether the URI for that is <http://schema.org/>, or another resource
represented by the RDFa serialization. Given the status of the latter as a
"mere" representation, I do agree that describedby seems more correct. But
since schema.org in general avoids the distinction, what I had in mind was
more along the lines of this (expressed in TriG):

    graph <http://schema.org/Thing> {
        <http://schema.org/Thing> rdfs:isDefinedBy <http://schema.org/> .
    }
    graph <http://schema.org/> {
        <http://schema.org/> rdfs:seeAlso <
http://schema.org/docs/schema_org_rdfa.html> .
    }

(In general when I write scripts processing multiple vocabularies together,
I look for rdfs:isDefinedBy. Rather often though, I need to revert that to
string inspection of the term identifiers. Even for OWL. So I might be
wishing for too much here.)



> Ideally, you want to use both relations due to the fact that an Ontology
> is one entity, a Document comprised of what constitutes an Ontology is
> another entity [1].
>

Ideally, I would agree. :) But even this distinction was under debate
elsewhere some years ago [2] (ah, pedantic-web..). And since
schema.orgdoes "shadow" documents with what their URIs formally denote
(which I
accept), I was thinking of avoiding that debate by aiming for "what is
meant" rather than "what describes what". But I surely know this is hard to
do, and this question does raise the distinction. Hopefully not to the
point of debating httpRange-14 (let's not go there here), but to determine
a useful means for describing and discovering schema.org the vocabulary.

(Given how this goes I might pull out of the debate, since I get really
nervous when I say httpRange-14 in public like this. It's like saying
"Candyman".)

Cheers,
Niklas

[2]:
https://groups.google.com/forum/#!msg/pedantic-web/RZ6kxlAVIy8/8r_JE4gVXFAJ




> We use the pattern outlined above, across the board.
>
> [1] http://bit.ly/1dYlGqz -- Illustrating use rdfs:DefinedBy and
> wdrs:describedby relations
>
> --
>
> Regards,
>
> Kingsley Idehen	
> Founder & CEO
> OpenLink Software
> Company Web: http://www.openlinksw.com
> Personal Weblog: http://www.openlinksw.com/blog/~kidehen
> Twitter Profile: https://twitter.com/kidehen
> Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
> LinkedIn Profile: http://www.linkedin.com/in/kidehen
>
>
>
>