- From: Holger Knublauch <holger@topquadrant.com>
- Date: Mon, 30 Mar 2015 08:50:12 +1000
- To: Richard Cyganiak <richard@cyganiak.de>
- CC: "public-data-shapes-wg@w3.org" <public-data-shapes-wg@w3.org>
On 3/30/2015 8:46, Richard Cyganiak wrote: > So maybe sh:class is the least bad after all? I'd be OK with sh:class. Holger > > Richard > > > >> On 29 Mar 2015, at 22:49, Holger Knublauch <holger@topquadrant.com> wrote: >> >> I am not happy with sh:type either, and welcome better suggestions. >> >> - sh:valueType is not clear enough to me, and would only make sense if it's about both datatype and object values. >> >> - sh:objectType, while aligning with owl:ObjectProperty, is misleading because the term "object" is also used for the triple position, and then it wouldn't exclude literals either. We'd have to use sh:subjectType for inverse properties. >> >> - sh:resourceType may be misleading to those people who understand Resource to include literals. >> >> - sh:rdfType is too geeky for people who don't even know/care that RDF is the foundation of this language. >> >> What other options do we have? Once again it seems to come back to the issue that there is no accepted term for "IRI or Blank Node". >> >> Holger >> >> >>> On 3/30/15 2:10 AM, Irene Polikoff wrote: >>> +1 >>> >>>> On 3/29/15, 11:41 AM, "Karen Coyle" <kcoyle@kcoyle.net> wrote: >>>> >>>> I really would advise against sh:type for exactly the reason you give. >>> >>
Received on Sunday, 29 March 2015 22:51:28 UTC