- From: Saied Tazari <Saied.Tazari@zgdv.de>
- Date: Fri, 27 Jul 2001 21:51:05 +0200
- To: www-rdf-interest@w3.org
Aaron Swartz schrieb: > > On Friday, July 27, 2001, at 05:46 AM, Andrei S. Lopatenko wrote: > > > But how to say that it is a Bag of person-project? > > This is a known issue: > > http://www.w3.org/2000/03/rdf-tracking/#rdfs-constraining-containers Why can't we handle containers transparently? I mean, we can say the range of a property must be of some type and leave it free to specify arbitrary number of objects of that type for the property in a concrete case. If no or only one object is specified, it is unambiguous. If more than one objects are specified, from the type of the used container we can understand if they are alternative objects, a set of objects, or a list of objects. To have a control on the number of objects there could be some properties such as rdfs:cardinality, rdfs:maxCardinality, and rdfs:minCardinality all three with rdfs:domain equal to rdf:Property and rdfs:range equal to xsd:nonNegativeInteger. To control the type of the possible container, we can have another property such as rdfs:containerType with domain equal to rdf:Property and range equal to rdfs:Container. To my mind, none of these 4 proposed properties are needed in the case of alternative values, because a single value is logically equivalent with only one alternative. On the other hand if some user finds an rdf:Alt as value she only needs some policy to choose the most appropriate one but finally she has to use only one of the alternative values. Do I miss anything? Regards, -- Saied
Received on Friday, 27 July 2001 15:51:48 UTC