Re: Max Expressivity of SPARQL (was... Re: Introductions: Lee Feigenbaum and Ben Szekely)

On Wed, Mar 3, 2010 at 11:51 AM, Marcelo Arenas
<marcelo.arenas1@gmail.com>wrote:

> Hi Juan,
>
> On Wed, Mar 3, 2010 at 12:26 PM, Juan Sequeda <juanfederico@gmail.com>
> wrote:
> >>
> >> """
> >> Ahmed: ericP, can you please ask Ivan also about SPARQL
> >> ... about the max expressivity of SPARQL because we don't want to have a
> >> mapping language that supports more expressivity than SPARQL
> >> """
> >
> > The paper by Angles and Gutierrez [1] states that SPARQL and
> non-recursive
> > safe Datalog with negation have equivalent expressive power, and
> therefore
> > the expressiveness of SPARQL is equivalent to relational algebra.
> > We have represented the translation of sql schema to rdf schema/ owl in
> FOL
> > and Datalog [2].
> > Marcelo, correct me if I'm wrong... if we represent our mapping language
> in
> > Datalog, then we don't have to worry if it is more expressive than
> SPARQL...
>
> Yes, if you use non-recursive Datalog with safe negation then you
> don't have to worry if it is more expressive than SPARQL.
>
>
+1 :)

> > AND.. we don't have to worry about the semantics of the language. The
> only
> > thing we would have to work on is the syntax.
>
> This is a nice feature of the Datalog approach, you don't have to
> worry about the semantics of the language as the semantics of Datalog
> is well defined.
>
>
+1 :)


> Does the mapping language include some built-in predicates (like
> order)? If this is the case, then some care has to be taken when
> defining the semantics of Datalog and comparing the expressiveness of
> this language with that of SPARQL.
>

+1 :)

When we start defining the requirements of the mapping language, we should
just make sure that it can be represented in datalog.

For the case of order, if I understand correctly, it is just post-processing
and I quote: "The ORDER BY clause establishes the order of a solution
sequence." [1] which means that we need to have the solution already in
hand, therefore order does not need to be in the semantics of datalog.


[1] http://www.w3.org/TR/rdf-sparql-query/#modOrderBy


> Best regards,
>
> Marcelo
>