- From: Holger Knublauch <holger@topquadrant.com>
- Date: Mon, 14 Sep 2015 15:55:53 +1000
- To: public-data-shapes-wg@w3.org
On 9/14/2015 15:52, Karen Coyle wrote: > > > On 9/14/15 2:43 AM, Holger Knublauch wrote: >> In my current work they look like >> >> ex:MyShape >> a sh:Shape ; >> sh:constraint [ >> a sh:LessThanConstraint ; >> sh:predicate1 ex:property1 ; >> sh:predicate2 ex:property2 ; >> ] >> . >> >> i.e. the ordering is defined by being predicate1 or predicate2. > > OK. So "sh:predicate1" is an actual predicate name. But of course > "property1" is not probably an actual property name. It would be more > readable to do an example where the instance properties are something > that make sense, like: > > ex:MyShape > a sh:Shape ; > sh:constraint [ > a sh:LessThanConstraint ; > sh:predicate1 ex:birthDate ; > sh:predicate2 ex:deathDate ; > ] > . Agreed, and I did just that today (among other things): http://w3c.github.io/data-shapes/shacl/#LessThanConstraint Holger > > or > > sh:predicate1 ex:startTime ; > sh:predicate2 ex:endTime ; > > I often have trouble telling which names are "real" and which are made > up in the examples because of how the examples read, so I would > suggest making examples that are more real, which would help > readability. I know that "sh:" should always be a real property name, > but this is where naming becomes important for clarity. "property1/2" > unfortunately looks a lot like "predicate1/2" in terms of "fictiveness." > > Perhaps this is something that Arthur could look at in his review - > and since it's my idea I volunteer to help wherever I can. > > kc
Received on Monday, 14 September 2015 05:56:26 UTC