- From: Guus Ramackers <Guus.Ramackers@oracle.com>
- Date: Thu, 03 Jul 2003 17:52:48 +0100
- To: David Booth <dbooth@w3.org>
- CC: www-ws-desc@w3.org
David, This is very helpful for me as an observer (responsible for WSDL IDE tool support at Oracle). Two suggestions intended to further sharpen: - the phrasings "The x defined by the <y> element" don't clarify the semantics. Perhaps move to the end of the definition instead of the start (i.e. more a side note than the central definition - which one generally expects as the first sentence)? - The word "agent" is a little confusing to me - it is not defined as a wsdl term, and in the real world (and in the IT world) may mean many things. Would it be possible to phrase the definitions without this term? Even if it would mean using "implementation" or such - if wsdl is a component model then the interface/contract terminology is typically contrasted with a realization/implementation behind it (e.g. in UML 2.0). Hope this helps, Thanks, guus David Booth wrote: > > I apologize for the lateness of this. I tried to send it last week, > but evidently it was lost when W2K crashed. > > ---- > > At today's call, I took an action to propose definitions for terms > related to the diagram[1]. My main concern is: > > There should be an obvious naming correspondence between WSDL > syntax > and the real world or conceptual things that WSDL documents > describe. > > I think this is important to avoid seriously confusing our readers. > So for our current syntax, I suggest the following terms for the > corresponding "real world" things. They are somewhat pedantic, but > perhaps they'll help. > > wsdl:interface > The (abstract) interface described by the <wsdl:interface> > element. > It is a collection of wsdl:operations. The wsdl:interface > defines how a client can interact with a wsdl:service through an > agent that is accessed at a location identified by a > wsdl:endpoint. > > wsdl:service (a/k/a "Web service" or "service") > The (abstract) service described by the <wsdl:service> element. > It is accessed through a collection of endpoints, and (if we keep > the wsdl:@targetResource attribute) it may manipulate some other > resource that is identified by the wsdl:@targetResource. > > wsdl:endpoint > The (abstract) thing that is described by the <wsdl:endpoint> > element. > It provides a location for accessing a particular agent that is > accessible via particular protocols described by its associated > binding. This agent is NOT assumed to be the same as the > wsdl:service, > since there may be several agents at different wsdl:endpoints > corresponding to the same wsdl:service. (The wsdl:service is > more abstract than the agent.) However, interacting with such > an agent (through a wsdl:endpoint) implies that you are > interacing > with the associated wsdl:service. > > wsdl:operation > The (abstract) operation template described by the > <wsdl:operation> > element. It is a template for a single interaction with the > wsdl:service. > > wsdl:binding > The (abstract) association described by the <wsdl:binding> > element. > It associates an (abstract) interface with transport and > serialization > details. > > wsdl:targetResource > The (physical or conceptual) resource[2] identified by the > wsdl:@targetResource attribute. Multiple wsdl:services that > refer to the same wsdl:targetResource are assumed to "manipulate" > the same resource, though what "manipulate" means is application- > defined. The wsdl:targetResource is NOT assumed to be the same > entity as the wsdl:service, though in some cases it could be the > same entity. > > References > > 1. Diagram: > http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/wsdl12.html#serviceresfig1 > > 2. RFC2396 def of "resource": http://www.ietf.org/rfc/rfc2396.txt > > -- _____________________________________________________________ Guus Ramackers Product Manager UML and Web Services Tools Oracle JDeveloper Tools group 520 Oracle Parkway, TVP Reading RG6 1RA, UK work: +44-(0)1189-245101 e-mail: guus.ramackers@oracle.com fax: +44-(0)1189-245148
Received on Thursday, 3 July 2003 12:51:14 UTC