Re: [foaf-dev] Re: privacy and open data

Phil Archer wrote:
> Henry, all,
> I've jumped back to the start of this thread to raise an additional 
> concern/use case about the whole issue of data portability and privacy 
> protection.
> I've been told many times that once asserted, an RDF triple persists 
> until the entropic heath death of the universe. Therefore any privacy 
> protection system that depends on triples, at least in theory, cannot 
> support the rather human practices of growing up, changing one's mind, 
> changing jobs etc.
> If I allow my e-mail address to be seen by anyone who is no more than 
> 2 hops from my FOAF file, I would want that check be carried out in 
> real time so that if I deleted someone from my FOAF file, the 
> permission would not be granted. But does this break the Semantic model?

Simple answer is: No.
> In the POWDER WG, with help from Jeremy Carroll, we've defined a 
> semantic extension to make it possible to put an expiry date on 
> triples [1]. Can something like this be worked into data portability?
> I could back this up with a nice gentle "I've changed job so no more 
> e-mail about its staff social functions please" but here's a more 
> scary one:
> Imagine a teenage girl who is being abused at home. She uses her 
> social network to call for help. Luckily, she finds it and manages to 
> escape the dangerous home life. Now she wants to keep in touch with 
> her new support network but become invisible to her former abuser.
She assumes a new Identity via a new URI in here new Social Space.
> In short, any privacy control needs to support changing circumstances.
One URI dies and another is born :-)

> Phil.
> [1]



Kingsley Idehen	      Weblog:
President & CEO 
OpenLink Software     Web:

Received on Wednesday, 26 March 2008 13:54:25 UTC