- From: Jonathan Marsh <jmarsh@microsoft.com>
- Date: Wed, 21 May 2003 10:20:59 -0700
- To: <public-ws-desc-meps@w3.org>
Session Start: Wed May 21 09:06:54 2003 Session Ident: #ws-desc *** Now talking in #ws-desc *** dbooth has joined #ws-desc *** alewis sets mode: +o Marsh *** alewis sets mode: -o alewis * alewis likes that better ... *** umit has joined #ws-desc <dbooth> http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/wsdl12/meps-vs-iops/mep s-vs-iops_clean.htm *** Marsh is now known as Scribe <Scribe> David: Previously uncovered differences in assumptions. <Scribe> ... For example, how messages are viewed. <Scribe> ... Go through assumptions, see if we all agree. <Scribe> 1) is agreeable to all <Scribe> 2) ? <Scribe> 4) WSD has a single "meaning" <Scribe> Jonathan: what's the meaning of "meaning" - different users will view it through different lenses. <Scribe> Umit: Tells client how to connect to a service. Contract for the client as presented by the service. <Scribe> Umit: How it came into existence has nothing to do with it. <Scribe> David: Anyone who reads the contract should infer the same meaning. <Scribe> 6) <Scribe> Jonathan: "complete"? <Scribe> David: includes endpoint uris, bindings, interfaces. <Scribe> Amy: "interface" is a generic term here - that might confuse someone. <Scribe> Umit: "A complete WSD" -> "A full WSDL document including imports" <Scribe> "A full definition of the service, which can be defined by multiple WSDL documents and Schemas". <Scribe> Umit: Transitive closure of documents that define a service. <Scribe> Amy: WSDL can contain multiple services. <Scribe> Amy: Don't have a concept of a "service" (other than the one being developed by the WG) <Scribe> Amy: Need a term for the chain of WSDL constructs defining a particular service? <Scribe> David: My definition is a service that contains endpoints. <Scribe> Jonathan: Should we look at it starting with the particular Web service and all the description components associated with it? <Scribe> David: Trying to look at it from the point of view of the description. <Scribe> Amy: Multiple clients may make use of the same contract. *** dbooth has quit IRC (Ping timeout) <Scribe> Amy: Don't need "complete WSD" concept to get this across. *** dbooth has joined #ws-desc <Scribe> Umit: Need an endpoint to interact. <Scribe> Amy: You can do interesting things with parts of WSDL (e.g. abstract) without actually interacting. <Scribe> David: Suppose you have a WSDL that includes interfaces, bindings, endpoints - what is it useful for? <Scribe> Amy: Depends who's looking at it. <Scribe> ... Registry use is different than client. <Scribe> ... Item 2 and item 4 amount to the same thing? <Scribe> Umit: 1) seems to cover multiple clients. <Scribe> Jonathan: ... "it's client or clients" <Scribe> Jonathan: 2) "primary" purpose (not all purposes are mechanics of interaction) <Scribe> Amy: Need to distill these assumptions down. <Scribe> Amy: 1) = from point of view of service, WSD is intended for use by service _and_ it's client(s). <Scribe> ... 2) WSD represents a contract. <Scribe> ... 4) Is a complete (only?) contract. <Scribe> David: no <Scribe> Umit: Two different people can infer the same meaning from the contract. <Scribe> Amy: Exclusionary statements will never be complete (phase of moon, etc.) <Scribe> David: Called these out because they were questioned <Scribe> Jonathan: Seems to be a vague assumption that a WSD should be adequate to capture the mechanics of interacting with the service. <Scribe> David: No, I'm saying that given a WSDL document, what is the intended purpose. <Scribe> Jonathan: Description can be used for many purposes. <Scribe> 1) A WSD is intended for use by a WS and its client or clients. <Scribe> 2) A primary purpose of the WSD is to ensure that the interacting parties agree on the syntax, datatypes and protocols they will use to interact. <Scribe> Amy: (drop 2a) <Scribe> 2) A primary purpose of the WSD is to establish a contract on the syntax, datatypes, and protocols which the interacting parties will use to interact. * alewis gives Scribe a trout * Scribe is not sure that is a good thing <Scribe> David: OK, but is it exclusive of those things? * alewis suggests that it's a good thing if it's fresh (or even alive), and that maybe the goodness is inversely correlated with freshness .... <Scribe> Jonathan: not exclusive. <Scribe> Umit: This should be a sufficient condition. *** umit has quit IRC (Quit: umit) *** alewis has left #ws-desc (Bye!) *** Disconnected Session Close: Wed May 21 10:19:23 2003
Received on Wednesday, 21 May 2003 13:43:26 UTC