- From: Jacek Kopecky <jacek.kopecky@deri.org>
- Date: Tue, 04 Jul 2006 17:08:51 +0200
- To: verma@cs.uga.edu
- Cc: SAWSDL public list <public-ws-semann@w3.org>
Hi Kunal and John, I'm not sure that this approach would in fact be correct. We can have a situation where a small printing interface is extended by a big aggregated service interface, which may break for example categorization. While the small printing interface may fall into a "printing services" category, the big aggregated interface that isn't limited to printing functionality could fall into a "general services" category. In some taxonomy the "printing services" category can be defined as services that only do printing. I'd say that all the model references do apply (i.e. none are overridden or precluded) in their scopes, in other words they do not propagate anywhere. 8-) Just my opinion, Jacek On Fri, 2006-06-30 at 17:08 -0400, Kunal Verma wrote: > Hi all, > > Following the general principle of all modelReferences applying to the > other components, we could compute a transitive closure of all the > modelReferences of all the interfaces extended by an interface and let > all of them apply to that interface. Also given our former discussion, > it might be wise to not imply any relationship between the > modelReferences. > > Thanks, > Kunal and John >
Received on Tuesday, 4 July 2006 15:09:04 UTC