[LRUG] Linked Data & Ruby

[Eleanor McHugh]

On 30 Jun 2009, at 14:38, Matthew Willson wrote:
> Yeah it could be done inside the model if every fact is given its  
> own URI. (is this typically the case with RDF? forgive my memory).  
> Although still need a good standard schema for how to go about that,  
> how to do signatures etc, and for this to be widely adopted. I see  
> one proposal here: http://rdfa.info/wiki/Security-and-trust#Subject-signing 
>  although it seems they haven't quite nailed it yet.
>
> Reason I think it could be better as a special metadata property of  
> a fact though, is the parallels with modal epistemic logics, which  
> have a special modal operator for things like  'agent x believes',  
> and (in some cases IIRC) are able to do some pretty clever  
> inferencing about trust and distributed knowledge as a result. And  
> just the general feeling that, on the web, a fact isn't much use  
> unless you know who's asserting it, and making it a part of the  
> model would encourage this.

Also trust is dependent on identity and managing that is definitely  
outside the scope of RDF. The last thing the internet needs is further  
proliferation of ad hoc identity management formats...

> Timestamp is another very key piece of metadata about a fact or  
> assertion, as the lifespan of facts is often not infinite, and quite  
> relevant to inferences being made from them... in this case  
> parallels with temporal logics.
>
> Anyway getting a bit OT now :)

Quite like old times then ;)


Ellie

Eleanor McHugh
Games With Brains
http://slides.games-with-brains.net
----
raise ArgumentError unless @reality.responds_to? :reason