Re: Additional data vs metadata

Hi Laura

Sorry, the example I gave probably makes it look like I'm talking about the
annotation syntax {| |}, but the problem I'm describing exists even with
just the quoted triple syntax << >>.

I don't know if I understand the question clearly. But I think the answer
> to your problem is simply to use different properties names, more
> vocabularies, or nested {||}. Not a more complex syntax.
>

That's where the problem lies though, I think. If I take the example from
the CG report, which appropriately is metadata in my opinion:

<< :Employee38 :jobTitle "Assistant Designer" >> :accordingTo :Employee22 .

And I decide to say something that's additional data rather than metadata,
so something that extends the description rather than being about the
description:

<< :Employee38 :jobTitle "Assistant Designer" >> :desk 8 .

Doesn't that mean the quoted triple is now being used to denote two very
different things, an abstract description in the first statement, and a
relationship/event in the second statement? Maybe someone with more
experience than me can tell me if that's a problem or not, I was always
under the impression that identifiers and blank nodes should only ever
describe one thing, not two. Doesn't reasoning break if people use quoted
triples to denote two different things?

To me, allowing usage like that, seems similar to allowing something like:

:x schema:familyName "Smith"
:x schema:organizer :W3C

Is "x" a Person or an Event? Can it be both? I don't have enough experience
to know if it's a problem or not.

Regards
Anthony

On Tue, Feb 15, 2022 at 1:22 AM Laura Morales <lauretas@mail.com> wrote:

> I don't know if I understand the question clearly. But I think the answer
> to your problem is simply to use different properties names, more
> vocabularies, or nested {||}. Not a more complex syntax. The {||} extension
> exists for solving a practical problem, due to the extreme verbosity of
> reification. It's not just a place to put metadata to. Or at least this is
> my understanding.
>
>
>
>
> Sent: Monday, February 14, 2022 at 2:48 PM
> From: "Anthony Moretti" <anthony.moretti@gmail.com>
> To: public-rdf-star@w3.org
> Subject: Additional data vs metadata
>
> I'm putting this in a separate thread because I think it's important
> enough. I'm of the view that there should be separation between additional
> data and metadata. RDF-star is a great model for metadata, but because
> there's no corresponding way to add additional data, people are using
> RDF-star for both. It's like a step was missed. The solution I propose for
> it is in another thread, "Three ideas[
> https://lists.w3.org/Archives/Public/public-rdf-star/2022Jan/0024.html]",
> specifically what I call a "complex statement". I'll quickly try to explain
> what I think the problem is, I'm rushing this email a bit, so if
> something's wrong let me know.
>
>
> There are three different things that in my view are constantly getting
> muddled in RDF-star examples. If a triple is in the subject position it's
> common to see annotations that, in my view, actually have different
> subjects, but you wouldn't know it from the way they're all grouped
> together. The subject could be:
>
> The relationship/event.The abstract description of the
> relationship/event.Specific occurrences of the abstract description of the
> relationship/event.
> The terminology I'm using:
>
> Triples where (1) is the subject are what I describe as "additional
> data".Triples where (2) is the subject are what I describe as
> "metadata".Triples where (3) is the subject are also what I describe as
> "metadata", but the proposed "occurrenceOf" vocabulary already comes to the
> rescue here.
> The intention of RDF-star, if I understand correctly, is to solve (2). But
> because there's no equally simple way to express (1) people are also using
> RDF-star for that. There's a difference between a description of a
> relationship/event, and a description of a description of that
> relationship/event. Very quick example where (1) and (2) are separated:
>
> :LizT :starredIn :JaneEyre
>     {
>         :role :HelenBurns,
>         :pay-USD 10000,
>     }
>     {|
>         :statedBy :Bob,
>         :statedIn :Wikipedia,
>     |}
>
>  So, very quick summary because I'm hurrying this, my suggestion is that
> it be made clear to users of RDF-star that it's only for metadata, or in
> parallel we come up with an equally simple syntax for n-ary relationships
> so that there are simple ways to model both things without getting them
> muddled.
>
> If I'm missing something here and it's not so black and white, or if you
> don't think it's a problem, please suggest some examples that we can talk
> about. Sorry for the hasty way this was written.
>
> Regards
> Anthony
>