Re: IPTC Sport Schema

Yeah at this stage I wouldn’t advocate for the Role pattern in any new
work. W3C is busy updating rdf, sparql, json-ld etc to support arbitrary
properties of properties. The rdf-star thing. It would be good to  look at
whether that captures the usecase

On Sat, Nov 30, 2024 at 19:54 Ryan Levering <rrlevering@google.com> wrote:

> There's nothing complex about http://schema.org/athlete.  It's just
> SportsTeam -> athlete -> Person:
> { "@type": "SportsTeam", "name": "Cool team",
>   "athlete": { "@type": "Person", "name": "Cool Player" }}
>
> There's not even an inverse property for it currently.  I mean, you can
> use Person -> memberOf -> SportsTeam but that's an even more ambiguous role.
>
> Things like Role allow SportsTeam -> athlete -> Role -> athlete -> Person
> so you can add metadata that further explain the relationship.  So:
> { "@type": "SportsTeam", "name": "Cool team",
>   "athlete": { "@type": "Role", "startDate": "2023-12-1", "athlete": {
> "@type": "Person", "name": "Cool Player" }}}
>
> Google systems parse that, but it's very esoteric and I don't think anyone
> actively uses it.  I'd personally rather have a system that is a) more
> general/standardized like rdf-star or b) more specific but fully fleshed
> out like the IPTC sports schema.  Role was an attempt to do this purely in
> the scope of schema.org.
>
> On Tue, Nov 26, 2024 at 9:05 PM Hugh Paterson III <sil.linguist@gmail.com>
> wrote:
>
>> Hi Ryan,
>>
>> Can you drop a link to the description and example of the use of “so#athlete”?
>> I think understand what is going on but I’d like to read up on it again.
>> Thanks.
>>
>>
>> All the best,
>> -Hugh
>>
>> Sent from my iPhone
>>
>>
>> On Tue, Nov 26, 2024 at 9:33 AM Ryan Levering <rrlevering@google.com>
>> wrote:
>>
>>> schema.org usually has the unfortunate goal of having a simpler/less
>>> valid ontology coincident with a better one.  So in schema.org we can
>>> use Role to specify time-limited memberships like you encode in this schema
>>> directly.  But we also allow so#athlete as a time snapshot for membership
>>> (which is frankly more used).  We also have so#homeTeam and so#awayTeam
>>> which are again shortcuts for your participation alignments.  But the IPTC
>>> schema seems like a great model to push in the direction of and I think
>>> many of the types there could be directly co-typed with schema.org
>>> types without problems.
>>>
>>> On Tue, Nov 19, 2024 at 7:52 PM Paul Kelly <paul@polvo.ca> wrote:
>>>
>>>> Hi Public Schema/Schema.org <http://schema.org/> people, it’s been a
>>>> while since I’ve participated here. I’m Paul Kelly, lead of the IPTC’s
>>>> sports content group.
>>>>
>>>> I’m writing to tell you about the IPTC’s ontology for representing
>>>> sports competitions and results in RDF. It’s called IPTC Sport Schema and
>>>> you can check it out here:
>>>>
>>>> https://sportschema.org/
>>>>
>>>> Here’s a diagram:
>>>>
>>>> https://sportschema.org/schema-diagram/
>>>>
>>>> The site has samples, a SPARQL endpoint and copious documentation.
>>>> There have been initiatives in the past for representing sports data on the
>>>> web. Any interest still?
>>>>
>>>> Apart from structuring complex competitions and athlete histories, we
>>>> incorporated a slew of sports vocabs that arose from the IPTC’s SportsML
>>>> standard:
>>>>
>>>> https://cv.iptc.org/newscodes#sportcvs
>>>>
>>>> I’d like to hear any thoughts people on this list might have about what
>>>> we’ve done. This thing will evolve and your contributions are welcome and
>>>> appreciated.
>>>>
>>>> Regards,
>>>>
>>>> Paul Kelly
>>>>
>>>> -----
>>>> Paul Kelly
>>>> Lead, IPTC Sports Content Working Group
>>>>
>>>