- From: Champion, Mike <Mike.Champion@SoftwareAG-USA.com>
- Date: Sun, 15 Jun 2003 15:00:40 -0400
- To: www-ws-desc@w3.org
> -----Original Message----- > From: Mark Baker [mailto:distobj@acm.org] > Sent: Sunday, June 15, 2003 2:05 PM > To: www-ws-desc@w3.org > Subject: Re: targetResource wording > > > The resource(s) that a service uses are part of the > implementation of a > service, not its interface. Please don't conflate the two. OK, how about an ERP system whose functionality one wants to expose as Web services to other applications in an enterprise; its "implementation resources" are RDBMS tables for the most part. Surely you're not suggesting that the other applications directly read/write/delete records in those separate tables are you Mark? Doesn't it make a *lot* more sense for the "service" to be an interface to some component that performs business-level operations ("add new employee" or whatever) on those tables but hides the implementation details from the service consumer? This is the essence of the difference between REST and [whatever the alternative is called], IMHO. REST seems to imply that the "real" or "implementation" resources should be exposed via their URI and some minimal set of operations directly on them; the [alternative] approach allows the identified resource to be an interface that encapsulates and may well hide the implementation details of the back-end resources. So I agree that one should not conflate the interface and implementation resources, and see REST as a special case of [the alternative] where the interface and implementation are identical, but strongly disagree that WSDL should expose the resources that a service uses as part of its implementation rather than those that define the interface.
Received on Sunday, 15 June 2003 15:00:56 UTC