Re: Blank nodes semantics - existential variables?

On 7/5/20 12:43 AM, Aidan Hogan wrote:
>>     :Bob  :has  _:b1 .
>>     :Bob  :has  _:b2 .
> [ . . . ]
> Many agree with this idea that _:b1 and _:b2 might refer to two 
> different things in an RDF graph (and many do not). I think that RDF 
> semantics does not contradict this assumption (it states that Bob has at 
> least one thing, which does not contradict him having two or more 
> things). However the semantics is weaker than what some might want (Bob 
> has at least two things, or Bob has exactly two things). Like you say, 
> other semantics would give those stronger meanings ... but not without 
> leading to other undesirable cases, like if you parse the document 
> twice, using different parsers or from different locations, now you 
> might end up assuming that Bob has (at least) four things, six things, 
> eight things (unless you now lean blank nodes).

Again, if a higher-level RDF language required each object to have a 
(possibly composite) key, this problem would go away.

David Booth

Received on Monday, 6 July 2020 01:22:30 UTC