Re: When N-Ary and When RDF-star

Hi David

It sure would be nice if the expressiveness of rdf-star and labeled
> property graphs were treated uniformly as mere sub-cases of a *standard*
> n-ary relationship syntax (which we do not currently have).
>

In the first email of the "Three ideas" thread, you'll see that the ideas
of neutral assertion, "<<" ">>", negative assertion, "<!" "!>", and
metadata annotations, "{|" "|}", are only applied after the rest of the
system for binary and n-ary relationships is defined. So to be picky about
what you're saying, in my view the expressiveness of RDF-star (metadata)
should be layered on top of a complete syntax describing binary and n-ary
relationships, rather than being a subcase, as you put it, of n-ary
relationships. It's a subtle but important distinction for various reasons
I've described throughout the thread.

It's good to see people falling back to arguments based in natural language
though, that's where the idea for simple, compound, and complex statements
comes from. We already have well-understood sentence structures to describe
n-ary relationships, and we're all already proficient at using them, these
statement types just mirror those and that's why I think they have the
potential to be a good solution.

Regards
Anthony

On Sat, Jan 22, 2022 at 2:37 AM David Booth <david@dbooth.org> wrote:

> On 1/21/22 9:40 AM, Jerven Tjalling Bolleman wrote:
> > When to use RDF-star and when to use N-Ary a personal perspective.
> >
> > RDF-Star is to be used when the relationship itself is enough for the
> data user in the vast majority of their use cases.
> >  . . .
> > When this evidence supporting such a relationship becomes the key
> concept for the users then it is time to go N-Ary.
>
> Excellent explanation!  And this brings up a major concern that I have
> about rdf-star: it means that we would have at least 3 *different*
> syntaxes for representing relationships of different arities.
>
> It sure would be nice if the expressiveness of rdf-star and labeled
> property graphs were treated uniformly as mere sub-cases of a *standard*
> n-ary relationship syntax (which we do not currently have).
>
> FWIW, some ideas for standardizing n-ary relations have been collected
> here, though the problem is far from solved:
> https://github.com/w3c/EasierRDF/issues/20
>
> David Booth
>
>