non-IRs, was Re: Editors' proposal for membership predicate names

On 02/24/2014 09:26 AM, Kingsley Idehen wrote:
> On 2/21/14 12:55 PM, Roger Menday wrote:
>>
>> hello Kingsley,
>>
>> Thanks for your feedback.
>> I would like to think about your diagramming suggestions over the 
>> weekend.
>> But, I have a comment now on your point #2.
>>
>> My picture was the DirectContainer case.
>> Here, all the boxes are Documents.
>> So, there is nothing to change.
>> (for the IndirectContainer case, we *will* have other shapes in the 
>> picture I beleive)
>>
>> However ...
>>
>> MembershipTriple's link non-informational resources and 
>> ContainerTriple's link information resources. But, in 
>> a DirectContainer, the Document which is created is the object 
>> position of the membershipTriple. Therefore the membershipTriple is 
>> linking to an Information resource, which is a contradiction.
>>
>> My conclusion is that we shouldn't really have DirectContainers ...
>>
>> ?
>> Roger
>
> Roger,
>
> I don't believe a document about "net worth" and what the term "net 
> worth" denotes are one an the same, hence my suggestion about entity 
> depictions. Naturally, I need a medium (e.g., a Web Document, Paper 
> Document etc.) through which I perceive your "net worth" but that 
> doesn't make the aforementioned perception medium (i.e., Document) the 
> signifier (i.e., identifier) of your net worth.
>
> The eternal issue is that the perception medium (so called 
> "information resource") and the signification mechanism (identifier) 
> continue to be conflated.
>

This is another case where I think it clarifies things to split 
Containers and Selections [1].    "Containers" manage information 
resources, identified by URLs.    They're an interface to a web hosting 
service, basically.    Selections involve collections of resources in 
general, including non-information resources.

In practice, in linked data, behind the scenes, each non-information 
resource has an associated RDF Source (an information source, with a 
URL), which is found by chopping off the fragment if necessary, and 
following certain redirects, if necessary.   The current LDP specs get 
things wrong by saying these are always 1-1, and that the 
non-information resource is necessarily the Primary Subject of the 
Information Resource.   (Even worse, it says that every resource in a 
container MUST use the same primary subject predicate, which is also not 
a great plan.)

          -- Sandro



[1] https://www.w3.org/2012/ldp/wiki/Collection_Types

>
> -- 
>
> Regards,
>
> Kingsley Idehen	
> Founder & CEO
> OpenLink Software
> Company Web:http://www.openlinksw.com
> Personal Weblog:http://www.openlinksw.com/blog/~kidehen
> Twitter Profile:https://twitter.com/kidehen
> Google+ Profile:https://plus.google.com/+KingsleyIdehen/about
> LinkedIn Profile:http://www.linkedin.com/in/kidehen
>
>
>
>