- From: Peter F. Patel-Schneider <pfpschneider@gmail.com>
- Date: Thu, 13 Feb 2014 14:52:53 -0800
- To: Gregg Kellogg <gregg@greggkellogg.net>, "Jason Johnson (BING)" <jasjoh@microsoft.com>
- CC: "public-vocabs@w3.org" <public-vocabs@w3.org>
On 02/13/2014 01:20 PM, Gregg Kellogg wrote: > [...] IMO schema.org has an implicit entailment rule that if a literal is provided as the value of a predicate having on object range, it's a stand in for an anonymous object with a schema:name of that literal. For example: > > <> schema:role "Coach" . implies <> schema:role [schema:name "Coach"] . > > The last one could also work if schema:Coach is an instance of RoleType with the name "Coach". > > It looks like we may be converging! > > Gregg > > It would be very nice to get some buy-in for this handling of strings where objects are expected. I proposed a variation of this before. I'm not particularly keen on using name as the generated role, partly because some objects don't have names, but it sure is better than the current unspecified behaviour. peter PS: It's not really an entailment rule, but instead a special treatment of a particular construct.
Received on Thursday, 13 February 2014 22:53:23 UTC