- From: Holger Knublauch <holger@topquadrant.com>
- Date: Wed, 16 Sep 2015 09:50:05 +1000
- To: public-data-shapes-wg@w3.org
On 9/15/2015 19:42, Karen Coyle wrote: > > > On 9/14/15 2:43 AM, Holger Knublauch wrote: >> The unique language case is different because it's about a single >> property only. Its syntax could simply be >> >> ex:MyShape >> sh:property [ >> sh:predicate skos:prefLabel ; >> sh:uniqueLang true ; >> ] . >> >> We have ISSUE-82 open for this topic. > > Great. Thanks. Then there is the one-to-many question. A case in point > comes again from SKOS, but without the requirement to look at language > codes. Essentially the rule is that you cannot have the same string as > both a prefLabel and an altLabel (there's more, the hiddenLabel, but > let's not go there right now). An instance could be: > > <term1> skos:preflabel "dogs" . > <term1> skos:altlabel "canines" . > <term1> skos:altlabel "pups" . > > a violation would be: > > <term1> skos:preflabel "dogs" . > <term1> skos:altlabel "canines" . > <term1> skos:altlabel "pups" . > <term1> skos:altlabel "dogs" . > > I presume that this would not make use of the pair-wise constraint > that you started this thread with. Is that the case? Making prefLabel and altLabel pairwise disjoint was exactly the scenario motivating sh:NotEqualConstraint. Holger
Received on Tuesday, 15 September 2015 23:50:39 UTC