Re: Text about SPARQL-star built-in functions ready

On Wed, 3 Mar 2021 at 08:57, James Anderson <anderson.james.1955@gmail.com>
wrote:

>
> > On 2021-03-02, at 21:34:20, Gregg Kellogg <gregg@greggkellogg.net>
> wrote:
> >
> >> On Mar 2, 2021, at 12:09 PM, James Anderson <
> anderson.james.1955@gmail.com> wrote:
> >>
> >> good evening;
> >>
> >>> On 2021-03-02, at 18:56:40, Olaf Hartig <olaf.hartig@liu.se> wrote:
> >>>
> >>> ...
> >>>
> >>> The changes are in Section 4.2 (SPARQL-star Grammar) and in the new
> Section
> >>> 4.4 (Function Definitions).
> >>>
> >>> As agreed during our telco last Friday, I will leave this PR open for
> three
> >>> days so that your can take a look at it and raise concerns (if any).
> Unless
> >>> anyone brings forward reasons that this PR should not be merged, I
> will merge
> >>> it on Friday afternoon (CET).
> >>
> >> i raise the concern, that the approach needs to comprehend quads before
> it will be usable.
> >
> > I think this comment is more general than the specific functions added.
> It would be useful to have some examples that use various combinations of
> embedded triples, annotations, and these functions within named graphs.
>
> it would also be useful to know the consequences for paragraph 15.1 and
> section 17 of the sparql recommendation.
>

In the interest of saving everybody's time trying to puzzle out what these
sections are about and how it relates to the topic under discussion, I have
done some homework. I'll ignore the earlier remark about quads because I
have no clue what is intended there or how it fits in with the topic under
discussion (the new functions).

Section 15.1 is about establishing a partial ordering of RDF values for the
purpose of executing an ORDER BY clause. See
 https://www.w3.org/TR/sparql11-query/#solutionModifiers
<https://www.w3.org/TR/sparql11-query/#solutionModifiers> . Although it has
no direct bearing on the introduction of these new functions, there is a
valid point here that we'll need to establish how RDF-star triples fit into
that partial ordering.

For what it's worth, RDF4J has extended the partial ordering by adding
RDF-star triples at the end (behind RDF literals), and by comparing two
RDF-star triples by first comparing subject, then predicate, then object to
establish order.

Section 17 is about expressions and value testing. See
https://www.w3.org/TR/sparql11-query/#expressions. What I'm *guessing *mr.
Anderson is referring to as needing examination is the notion of
RDFterm-equality (see
https://www.w3.org/TR/sparql11-query/#func-RDFterm-equal). In particular,
it will need extending to explicitly state that two terms are considered
RDFterm-equal if they are both RDF-star triples and those triples are
equivalent under our predefined notion of equivalence (which I assume would
be most simply expressed in terms of each of its constuent parts being
RDFterm-equal, cf.
https://w3c.github.io/rdf-star/cg-spec/2021-02-18.html#dfn-rdf-star-terms).

Regards,

Jeen
-- 
Dr Jeen Broekstra (he, him)
*principal software engineer*

jb@metaphacts.com
www.metaphacts.com

[image: htps://www.metaphacts.com/] <https://www.metaphacts.com/>