Re: Peter's friends

On 3/15/13 9:20 AM, Roger Menday wrote:
> One thing in the discussions last night we want a want of referring to "Peter's friends" as well as "Peter".
>
> "Peter" is a resource, and so is "Peter's friends".

Not really clear. I think you mean:

The entity <#Peter> (an entity of type: foaf:Person) in with a label 
property "Peter" that is described by a  Document (an entity of type: 
Web Resource).

For "Peter" and friends you can have an entity of type: foaf:Group which 
has a label "Peter's Friends" and all of the relevant entities in a 
foaf:member relation expressed in the Document that describes this 
particular entity.

>
> "Peter's friends" can be paginated when it becomes too large.

The description documents can be viewed using pagination. Of course.

>
> "Peter's friends" can be appended to.
> e.g.
> simply POSTing "Mary" to "Peter's friends" will assert this new friendship.

You can add and remove members to and from the foaf:Group.

>
> We called "Peter's friends" a value set, and it is a useful construct it seems.
>
> What this could mean, is that a set of such atomic operations, i.e. "Add Mary to  Peter's friends", "Add Andrew to Peter's friends", "Remove Susan from Peter's friends", ... is a PATCH format (?)
>
> Roger
>
This already works using SPARQL.

The issue here is just about a SPARQL alternative that matches what 
SPARQL already delivers. Basically, this is just about not mandating 
SPARQL even though (IMHO) productivity realities will lead all 
implementers to adopt SPARQL, rather than painfully reinventing the 
wheel :-)


-- 

Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen