W3C home > Mailing lists > Public > public-rdf-wg@w3.org > August 2011

Re: Container Deprecation

From: Richard Cyganiak <richard@cyganiak.de>
Date: Tue, 30 Aug 2011 12:37:27 +0100
Cc: ivan@w3.org, public-rdf-wg@w3.org
Message-Id: <C720444A-D8F3-46F1-89AC-901D7340C46A@cyganiak.de>
To: Jesse Weaver <weavej3@rpi.edu>
Hi Jesse,

On 29 Aug 2011, at 20:42, Jesse Weaver wrote:
> It would be nice to be able to say in a single statement:
> 
> <group> <contains> <thing> .
> 
> You cannot do this with RDF collections.  You can do it with RDF
> containers using rdfs:member.

Not really. What you certainly can do is this:

  <group> <contains> [ a rdfs:Bag; rdf:_1 <thing>.].

and this implicitly uses rdfs:member because rdf:_1 is a subproperty.

The usage

  <group> rdfs:member <thing>.

is not documented anywhere in the RDF specifications. I don't think it is the intended usage. The specifications are too fuzzy to allow a definitive statement on whether it's correct or not. They don't assign any meaning to the property at all, except that the property is implied by membership in a Seq, Bag or Alt. (Which is a failure of the spec, and one of the reasons why containers are best avoided.)

So, the usage you propose assumes that rdfs:member has a meaning that is sort of implied by the property's name, but not given explicitly in the spec.

As Dan said, dcterms:hasPart is more appropriate for your use case.

Best,
Richard


>  It seems that we should have some way of
> doing this, regardless of whether RDF containers are deprecated.
> 
> Jesse Weaver
> Ph.D. Student, Patroon Fellow
> Tetherless World Constellation
> Rensselaer Polytechnic Institute
> http://www.cs.rpi.edu/~weavej3/
> ==============Original message text===============
> On Mon, 29 Aug 2011 8:09:22 EDT Ivan Herman wrote:
> 
> Jesse,
> 
> - There is no formal resolution on containers yet. There has been some
> discussions, and deprecating containers has indeed been raised as one of
> the candidates.
> 
> - However, your usage of rdfs:member v.a.v. lists (a.k.a. collections) is 
> a slightly different question. At them moment, there is no semantic
> relationships in RDFS between the terms used for lists (rdf:first,
> rdf:next, or rdf:List) and rdfs:member. Put it another way if I have 
> 
> <a> <b> (<c> <d> <e>) .
> 
> I cannot infer something like
> 
> <c> rdfs:member _:a . # _:a is the 'head' of the list above.
> 
> I may have misunderstood what you said, though.
> 
> Ivan
> 
> 
> 
> 
> On Aug 26, 2011, at 20:59 , Jesse Weaver wrote:
> 
>> Hello RDF working group.
>> 
>> Has there been consensus concerning deprecation of RDF containers in 1.1? 
>> Specifically, I am curious about the rdfs:member property.  It is very
>> useful for stating membership of something in a uniform way (unlike using 
>> rdf:_1, rdf:_2, ...) and in a single triple (unlike using rdf:first and
>> rdf:rest).  I am well aware of the distinction between containers and
>> collections, but it seems that RDF really needs something as simple as
>> rdfs:member.
>> 
>> Please let me know.
>> 
>> Thank you.
>> 
>> Jesse Weaver
>> Ph.D. Student, Patroon Fellow
>> Tetherless World Constellation
>> Rensselaer Polytechnic Institute
>> http://www.cs.rpi.edu/~weavej3/> 
>> 
>> 
>> 
>> 
> 
> 
> ----
> Ivan Herman, W3C Semantic Web Activity Lead
> Home: http://www.w3.org/People/Ivan/mobile: +31-641044153
> PGP Key: http://www.ivan-herman.net/pgpkey.htmlFOAF: http://www.ivan-herman.net/foaf.rdf
> 
> 
> 
> 
> 
> ===========End of original message text===========
> 
> 
> 
> 
> 
Received on Tuesday, 30 August 2011 11:37:57 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:45 GMT