Graham > For example, if we have a suitably flexible core model, expressed in RDF, then I > would expect a key extension point to be refinement (subclasses and > subproperties) of the core elements. That needs no process. agreed, but as you say: > I accept that not all desirable extension points may be as obliging, but for now > I'm happy to leave the details unresolved. what may be needed is /principles/ for doing extensions right. I can make up any extended class structure which is a formally correct refinement but makes no sense to anyone else :-) -PaoloReceived on Friday, 3 June 2011 09:29:14 UTC
This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:58:05 UTC