Re: Blank Nodes Re: Toward easier RDF: a proposal


Mornin'! Alex here, playing the part of the Devil's advocate today ...

On Sun, Nov 25, 2018 at 6:13 AM Pat Hayes <> wrote:
> The generic pronoun is actually 'something'. The triple
> ex:PatHayes ex:owns _:x17 .
> says 'Pat Hayes owns something', without saying what it is that I
> own. One can conclude things from this: I am not destitute, for
> example.

Hmm. Since the range _:x17 is from "nothing" to "everything", I'm not
sure that's true. Inference is a bitch.

If you know more about what I own:
> _:x17 rdf:type dbpedia:Real_estate .
> then you can infer more: that I am actually in reasonable
> financial circumstances.

Again, the data is fine, but your inference is not. I'm sure "sales
poster" is part of the type system. One can also infer that you've
gone broke, the house is forfeited, and you're underselling out of
manic desperation before driving off a bridge ...


> Also, it takes work to create a URI, and a quite
> unreasonable amount of work to create a 'cool' one.

URIs are pesky at times. I'm almost tempted to say they're
counter-productive and pointless if the reason for their existence is
not what we use them for ...

> OK, bnodes do make RDF more complicated than it would be without
> them. But RDF without blank nodes is just data graphs.

So, what are they now, *with* them? :)


 Information Alchemist / UX consultant / GUI developer for hire     |

Received on Saturday, 24 November 2018 20:20:23 UTC