Re: Bnodes or all-URIs? Straw poll will do. [was: Re: Help needed: *brief* online poll about blank-nodes]

On 19 June 2011 02:59, David Booth <david@dbooth.org> wrote:
> And if anyone wishes to add thoughts or ideas to the Bnode Slokemization
> wiki page that we started after a previous discussion, please do:
> http://www.w3.org/wiki/BnodeSkolemization

While it's a tasty-sounding phrase, I don't think it's accurate.

Plonking a name on something isn't the same as transforming it into a
constant with a function. Constant, sure, and HTTP likes that. But
there's a bit more expected no?

I'm not sure how much the RDF model as it stands requires, but as a
model of the real world we are looking at a seriously partial function
between the abstract syntax and the reality.

No technical problem in the approach of giving names wherever
existentials occur, but I'd suggest naming unknowns is likely to cause
problems further down the line.

Grr, sorry, no strong arguments yet. It's mostly just an intuitive thing.

As a teenager I was stopped by the police on a motorcycle ride, I was
not known, not wanted. My cousin (on the front seat with the helmet)
was known, not wanted. The wanted on one side maybe as HTTP GETs. But
known was without anything like he needed to be locked up, just he
existed in the police universes. I was there on the road, but not in
the police universe. Damn, I'm in the wrong universe to make a point.

Grr, still intuitive, having trouble pinning things down.

Cheers,
Danny.


http://danny.ayers.name

Received on Sunday, 19 June 2011 01:49:26 UTC