- From: Graham Klyne <GK@Dial.pipex.com>
- Date: Thu, 07 Sep 2000 22:22:44 +0100
- To: Ray Fergerson <fergerson@SMI.Stanford.EDU>
- Cc: rdf interest <www-rdf-interest@w3.org>, Mor Peleg <peleg@SMI.Stanford.EDU>
At 12:01 PM 9/7/00 -0700, Ray Fergerson wrote: >Graham Klyne wrote: > > (b) the lack of clarify about what rdf:Alt means, and the lack of a > > compelling prototypical application for it suggests to me that it might be > > struck from the core specification. > >I'd actually prefer to keep it but just remove the implication that >the items are "the same" in some unspecified way. It would then map >cleanly to the common programming language "enumerated type" concept >which has lots of uses. Enumerated types can also modeled as >subclasses or instances but sometimes it is just more natural to put >some things in a bag and say "pick one". Something like the colors on >a stop light can reasonably be modeled in this way rather than >introducing a totally artificial "stop_light_color" class. Hmmm... I think this is in danger of conflating union types with alternative values. (An enumeration type can be viewed as a union type of its individual enumerated values) If this is being used in the context of a type definition schema, then your comments make sense. But it seems to me that the primary use of rdf:Alt would in the _value_ of some _instance_ of a type, so the issue of enumeration does not arise -- it's just a value. #g ------------ Graham Klyne (GK@ACM.ORG)
Received on Thursday, 7 September 2000 17:25:02 UTC