Re: Radical change proposal

On 14/03/2024 13:01, James Anderson wrote:
> good afternoon;
> 
>> On 13. Mar 2024, at 12:37, Franconi Enrico <franconi@inf.unibz.it> wrote:
>>
>> All the problems and discussions about well-formedness, forbidden simple entailments, unsupported RDF 1.2 graphs, etc, will vanish if we give up option 3 and we go back to have only triple reification terms in RDF 1.2:
>>
>> graph             ::= (triple)*
>> triple            ::= subject predicate object
>> subject           ::= iri | BlankNode | tripleReification
>> predicate         ::= iri
>> object            ::= iri | BlankNode | literal | tripleReification
>> tripleReification ::= identifier triple
>> identifier        ::= iri | blanknode
>>
>> Too radical?
> 
> no it is not too radical.
> except, that, from the perspective of a sparql implementation, if graphs are to be used, their abstract form needs to be
> 
>      graph ::= identifier triple*
> 
> in which case, the difference to a tripleReification is just the size of the triple set.

This is the RDF data model for a graph, not a dataset.

Using (named) graphs is one implementation choice - there are others.

Any implementation choice should not require an additional change to the 
RDF data model because in doing so all complete implementations of RDF 
1.2 would then be expected to provide that data model feature.

     Andy

> 
> what of your proposal does this perspective misunderstand?
> 
> best regards, from berlin,
> ---
> james anderson | james@dydra.com | https://dydra.com
> 
> 
>

Received on Thursday, 14 March 2024 13:59:20 UTC