Re: ldp-ISSUE-79 (ldp:contains): ldp:contains

On 31 May 2013, at 18:48, Henry Story <henry.story@bblfish.net> wrote:

> 
> On 31 May 2013, at 18:11, Arnaud Le Hors <lehors@us.ibm.com> wrote:
> 
>> Hi Henry, 
>> 
>> I think the proposed text below has several issues: 
>> 
>> "relates an LDP Container to the elements it contains, ie LDPRs that were created through this LDPC or that act as if they had been" 
>> 
>> 1. LDPCs aren't limited to containing LDPRs. They can contain any types of resources, including binary ones. 
>> 2. LDPCs aren't limited to containing resources that are created from the LDPC. Although the end of the sentence opens up to that possibility I think the text unnecessarily implies a tie that just doesn't exist. 
> 
> agree. 

Ok but after the recent discussion with John Arwe I'd like to find a way to tie the relation
to the ldp HTTP actions. Because that is exactly what I want to focus on.

> 
>> 
>> I would suggest something like this instead: 
>> 
>> "relates an LDP Container to the resources it contains". 
> 
> Yes, I was hesitant about that way of expressing things, because it seemed nearly circular.

And I think doing that is what creates this incredbile confusion we have in the
group since the beginning relations and containment. 

> 
>> 
>> I think you're right that having an LDP specific predicate would prevent any ambiguity. Sadly, this is a good example of why it is so difficult to reuse existing vocabularies.
> 
> Reuse is often done by inference. So one could add to the definition
> 
> ldp:contains rdf:subPropertyOf rdf:member .
> 
> Then one gets the benefits of aligning intuitions of those who have understood rdf:member relation.
> But yes, in this case we use the concept ldp:contains in a very specific way, and rdf:member covers
> much more ground.
> 
>> 
>> --
>> Arnaud  Le Hors - Software Standards Architect - IBM Software Group
>> 
>> 
>> 
>> 
>> From:        "Linked Data Platform (LDP) Working Group Issue Tracker" <sysbot+tracker@w3.org> 
>> To:        public-ldp-wg@w3.org, 
>> Date:        05/31/2013 02:43 AM 
>> Subject:        ldp-ISSUE-79 (ldp:contains): ldp:contains 
>> 
>> 
>> 
>> ldp-ISSUE-79 (ldp:contains): ldp:contains
>> 
>> http://www.w3.org/2012/ldp/track/issues/79
>> 
>> Raised by: Henry Story
>> On product: 
>> 
>> replace all (most) references of rdf:member in the spec to ldp:contains . 
>> 
>> ldp:contains a rdf:Property;
>>    :comment "relates an LDP Container to the elements it contains, ie LDPRs that were created through this LDPC or that act as if they had been";
>>    :domain ldp:Container;
>>    :range ldp:Resource .
>> 
>> The advantage of using this relation is that:
>>  - it is more specific than rdfs:member which can be applied much more widely than LDPCs
>>  - it does not require the client to know that { <> a ldp:Container }, and so does not need to
>>     parse through all the triples before it can start interpreting the meaning of an rdf:member .
>>  - LDPRs that wish to refer to their LDPCs can do this in one relation with 
>>     { <.> ldp:contains <> . } this otherwise requires two relations 
>>     { <.> a ldp:Container; rdf:member <> }
>>  - ( very minor: it may reduce the need to import the rdf namespace ) 
>> 
>> 
>> 
>> 
>> 
> 
> Social Web Architect
> http://bblfish.net/
> 

Social Web Architect
http://bblfish.net/