Re: regrets & skolem predicates

Nathan wrote:
> Hi Pat,
> 
> Pat Hayes wrote:
>> a simple literal scheme for skolemization, along the lines 
>> "<text-of-bnode-id-label>"^^yourBaseUri:genid
> 
> I've noticed you mention this a couple of times now (iirc), and have to 
> say that my gut instinct says "Pat's got something here". Care to share 
> any more details or ways it could work?
> 
> What I'm reading here initially is that "_:b1" in syntax could be sugar 
> for a typed literal of the form "b1"^^rdf:BlankNode. Which seems a 
> rather nice little solution, however you have ^^yourBaseUri:genid, which 
> I can see has the URI in there and some obvious benefits from that, but 
> at this point things go gray and fuzzy for me!
> 
> Care to expand a little?

O! just noticed, a few weeks(months) ago, I was banging on about names 
being a pair and doing things like ( http://example.org/foo , ∃(bar) ) 
-> http://example.org/foo#[bar].

Anyway, just noticed that what you say above appears to be pretty 
similar, but using typed literals instead:

abstract ->
   ( http://example.org/foo , ∃(bar) )
full lexical ->
   "bar"^^<genid:http://example.org/foo>
syntax sugar ->
   _:bar

where genid: is the scheme used for all blank nodes, and indicates that 
anything with a datatype belonging to the genid: scheme is a blank node, 
bearing the identifier "bar" within the namespace/g-box 
http://example.org/foo.

Almost crazy enough to work?

Received on Wednesday, 20 April 2011 21:43:43 UTC