[Eulersharp-users] Fwd: EYE vs use_module(library('semweb/rdf_db'))

[Jos De R. <jos...@ag...>]

Hi Douglas,

Nice to see you back and hope all is fine!

> You wrote euler.yap initially for YAP, and then made it also
> work in SWI-Prolog. Since you already created it for YAP, you
> didn't need to leverage the library('semweb/rdf_db') system.
> Also you support .n3 such as :
> <#Socrates> <#beliefs> { <#Socrates> a <#Mortal> }.
>
> and rules such as:
>
> {?P rdfs:range ?C. ?X ?P ?Y} => {?Y a ?C}.

True.


> So Question:
> Did you find it would have been too much work to have made
>   SWI-Prolog existing
> SemWeb rdf_assert/3-4  a form such as?
> <#Socrates> <#beliefs> { <#Socrates> a <#Mortal> }.
>
> Would you needed to have made a bnode for that graph?

Already since 1999 when our initial prototype was made in Java,
we decided to use the p(s, o) representation for triples instead
of the triple(s, p, o) representation. We have kept it since
then as it is compact and performant (*): the memory load is
around 4 million triples per gigabyte and there are like 2
indexes, the p index and either the s or o index on JITI basis.
So this is basically a totally different design than the rdf/3
of library('semweb/rdf_db').


(*) https://github.com/josd/eyebmb

Kind regards,
Jos

Jos De Roo | Agfa HealthCare
Senior Researcher | HE/Advanced Clinical Applications Research
http://www.agfa.com/w3c/jdroo
http://twitter.com/josderoo

Agfa HealthCare NV, Moutstraat 100, B-9000 Gent, Belgium
http://www.agfa.com/healthcare