- From: Dan Brickley <danbri@w3.org>
- Date: Fri, 8 Jun 2001 15:19:10 -0400 (EDT)
- To: Brian McBride <bwm@hplb.hpl.hp.com>
- cc: <Ora.Lassila@nokia.com>, <phayes@ai.uwf.edu>, <w3c-rdfcore-wg@w3.org>
On Fri, 8 Jun 2001, Brian McBride wrote: > Ora.Lassila@nokia.com wrote: > [...] > > For "incomplete", unordered containers, some type of > > simple "membership" relation might also do... > > I've generally assumed the reason for using rdf:_n properties for bags was > that a graph was considered to be a set of statements, and if one wanted the > same object in a bag more than once a simple membership property wouldn't > do. Was this part of the original WG's thinking or no? I'm just curious. I don't remember, but another consideration was related to an artifact of the XML encoding: RDF had to something that could be serialized within HTML docs, which means using XML attributes for literal data, eg. <FOO rdf:_1="a" rdf:_2="b" />. Since XML doesn't allow for repeated attributes with the same name, so we needed the numbered container-membership properties. Dan
Received on Friday, 8 June 2001 15:20:18 UTC