- From: Eric Prud'hommeaux <eric@w3.org>
- Date: Mon, 8 Jul 2013 10:47:22 -0400
- To: public-ldp-wg@w3.org
It's not great, but here's some example text I came up with during the WG call: For instance, Santa's list of children could include a disposition of whether they are naughty or nice. A representation requiring inferencing would prevent other LDP applications from consuming this mission-critical data: |Invalid LDP Container contents|: <Children> rdfs:membershipPredicate np:knownChild . np:naughtyChild rdfs:subPredicateOf np:knownChild . np:niceChild rdfs:subPredicateOf np:knownChild . <Children> np:naughtyChild <Alice> . <Children> np:niceChild <Bob> . Instead, Santa's container of children must serialize the ldp membership predicate explicitly, though the application may include the extra triples to indicate the naughtiness of the contained children.: |Valid LDP Container contents|: <Children> rdfs:membershipPredicate np:knownChild . np:naughtyChild rdfs:subPredicateOf np:knownChild . np:niceChild rdfs:subPredicateOf np:knownChild . <Children> np:naughtyChild <Alice> ; np:knownChild <Alice> . <Children> np:niceChild <Bob> ; np:knownChild <Bob> . The above serves as a counter-example to the earlie example which would require inferencing for use by generic LDP clients. It does not imply that assertions about resources should be confined to the container. The following example shows the behavior information stored instead in resources. |Valid LDP Container contents|: LDP Container: <Children> rdfs:membershipPredicate np:knownChild . <Children> np:knownChild <Alice> . <Children> np:knownChild <Bob> . LDP Resource <Alice> a np:NaughtyChild . LDP Resource <Bob> a np:NiceChild . -- -ericP
Received on Monday, 8 July 2013 14:47:52 UTC