Re: UC&R ready for review

Alberto Reggiori wrote:
> 
> (replying as the original author of the "identity management" use case)
> 
> On Oct 4, 2004, at 10:05 PM, Pat Hayes wrote:
> 
> 
>>>  RDF Data Access Use Cases and Requirements
>>>  Revision: 1.138, Date: 2004/10/01 19:56:52
>>>  http://www.w3.org/2001/sw/DataAccess/UseCases
>>>
>>>is ready for review for publication.
>>>
>>>Kendall
>>
>>2.15 Mr X has two personae, why does he need three PPDs?
> 
> 
> in the scenario I had in mind (the Web?),  he might have several 
> separated PPDs handed around (one sitting on his home directory on the 
> laptop /Users/MisterX/foaf.rdf, one on a Web site at work, one on his 
> palm/iPod, one on a 3rd part server annotated by somebody else and so 
> on) - perhaps not all synchronized (merged) together and having 
> different trust levels - and overlaying to describe his two persona(s) 
> (characters).
> 
> my whole point with the "identity management" UC was to try motivate 
> the need of named-graphs (named containers) as bNodes

I would rather the UC&R document did not try to motivate technical solutions but 
  explained use cases and requirements.  I think it does that.

I don't think we can address any trust issues - the impact on query is just a 
small part of an overall approach.  The approach needs to be in place first 
before we can design to it.

As Matthias is revealing the source of triples, to Johanna and others, he's 
going to have to use URIs if source information is to be returned in query results.



A simple example might have been newsfeeds - needing to track which newsfeed 
says what is important.  Especially near elections.

	Andy


> and not simply 
> URIs i.e. here is my profile persona1.rdf, here is my profile 
> persona2.rdf  VS. here is my FOAF profile source about my persona(s). 
> In a certain sense have the possibility to refer to the whole 
> 'collection of named containers' about MisterX FOAF information 
> (instead of each separated conatiner) and distinguish each of them at 
> query time, as outlined in Andy's proposal
> 
> http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0581.html
> 
> is there any reason why a named-graph can not be a bNode and 
> referenced-by-description?
> 
> Yours
> 
> Alberto
> 
>

Received on Tuesday, 5 October 2004 09:08:21 UTC