Re: Blank Nodes and SPARQL

On Mon, Jul 11, 2005 at 01:18:54PM -0500, Dan Connolly wrote:

> > Thinking about it a little more, it does seem to change the semantics of
> > every graph that gets rewritten in that way.
> 
> The _!:... mechanism has the same effect, no? In the context in
> which it's used, it works logically like a URI; i.e. it matches
> the way URIs match.

I don't believe so. As Ron says in another message, OWL DL says some things
(i.e. restrictions) have to be anonymous. _!:foo is still anonymous, while a
URI isn't.

So they may be, from some point of view, equivalent, but they have different
effects, one of which is unacceptable IMO.

> > 1. It make sense (which is a good thing)
> 
> I accept that as your position; I don't share it.

Care to say why you claim that it doesn't make sense? (I assume you don't
mean that making sense isn't a good thing! :>)

> This is more than syntactic sugar for something that can already
> be expressed, isn't it? 

Hmm, yeah, maybe "syntactic sugar" is the wrong way to put it. I guess I
meant to say that it's a syntactic solution, rather than sugar for something
else that can already be expressed.

> What is this short for?
> 
>   SELECT ?MBOX
>    WHERE { _!:l55c33 foaf:mbox ?MBOX. }.

Okay, so it's not sugar per se.

This is the first I've heard you say that one of Andy's solutions to the
bnode problem isn't a solution. I guess we disagree about that.

> > 4. It feels right -- i.e., it's sufficient to the problem, but not overly
> > powerful, such that other, hard-to-see issues might be raised.
> 
> Adding a new type of term (in addition to URI, literal, bnode, and
> universal variable) makes me quite nervous about hard-to-see issues.

Fair enough. Nervousness, like mileage, obviously varies.

Kendall Clark