- From: Stasinos Konstantopoulos <konstant@iit.demokritos.gr>
- Date: Sun, 5 Oct 2008 17:49:49 +0300
- To: Phil Archer <parcher@fosi.org>
- Cc: Public POWDER <public-powderwg@w3.org>
On Thu Oct 2 14:56:04 2008 Phil Archer said: > > Andrea, Stasinos, > > Thank you for pursuing this, I can see that it is not simple - but I'm > hoping that between you it's possible to decide on a way forward. For me > the important things are: > > 1. At the moment, FOAF is the vocab that a lot of people use and > therefore we should support it. foaf:Agent must be a possible object of > wdrs:issuedby. > > 2. dcterms:Agent has greater formal long term stability and we should > therefore make it possible to use that as the object of wdrs:issuedby as > well. > > 3. We should not make any statement that in any way affects the > semantics of either of those external vocabularies. Under the strict interpretation of "the formal specification of wdr:issuedby must be such that, as it stands, does not affect the semantics of either external vocabulary". The *potential* to affect said semantics by adding triples that make assertions about the domain or range of wdr:issuedby, or subordinate wdr:issuedby under another property is not covered by this statement. This leaves both A1 and A2 open. The stronger interpretation (covering potential damages) closes up option A2. > 4. IMO, we should /only/ support those two, but, this may not be best > practice. If we make it so that other hypothetical Agent classes can be > used, I don't think anyone will object. In general, the fewer > restrictions that apply to a vocab term the better. If *only* is a requirement, A1 is the only way forward, as far as I can tell. But I might be wrong. > 5. We always used to use rdf:Description to describe the POWDER-S doc. > We changed to owl:Ontology in response to a comment from Ivan. If the > present discussion means that we should go back and revert to > rdf:Description, OK, we can do that. As I recall, it was a suggestion > from Ivan, no more. Like I said, rolling back owl:Ontology only allows us to syntactically pretend we are OK, although in essence no instance can ever be an rdf:Description and an owl:Thing. s
Received on Sunday, 5 October 2008 14:50:34 UTC