RE: Naming a Web service resource

just a few thoughts,

 >-----Original Message----- 
 >From: Anne Thomas Manes [mailto:anne@manes.net] 
 >Sent: Fri 04/07/2003 15:10 
 >Subject: Fw: Naming a Web service resource
 
 >Thinking through this scenario, I've realized that one of the must unwebby
 >features of the Web services framework is that a Web service doesn't have a
 >URI. It has a description (WSDL). It has one or more endpoints. But there is
 >no one URI that represents the resource that *is* the Web service.
 
 yes, and I would add that probably most folks would argue that the concrete physical endpoint is the URI you are speaking of..though this is prob because we may think of web services in terms of HTTP get/post binding. I would imagine that GET folks consider the concrete physical URL to be their URI, or to use reverse logic the http URI that represents the physical endpoint is their URL [1].
  
 >Reading through the various documents that describe URIs and the Web
 >architecture, it seems obvious to me that a Web service is an "important"
 >Web resource, therefore it should have a URI.
 
 
 a uri pointing to metadata about the 'thing' is a urc, which is probably what the URI pointing to the wsdl file could be thought to be. Since WSDL confines itself to describing the interface to the web service ( ok, though not solely interface description...), ; perhaps you desire more meta data in the form of a http URI pointing to a RDDL document....perhaps you are stating that there is a need to have an http URI bounded to a 'web service',  or maybe I am completly misunderstanding.
  
 to cross reference with inline comments
  
 point A: would be satisfied by a RDDL document
  
 point B: to put it simply if the interface changes, then the definition should change resulting in a new versioned wsdl file ( for me at least )
  
 the analogy with the requirement for a name is very true, though I think the problem is more related to the issue of making a service orientated architecture jig with the document centric web;
  
 an approach I use is to define a standard service, with HTTP Get binding, which points to a RDDL document for statisfying that innate 'webby' need to get a document after typing in a http URI; the uri is still pointing to the "thing", and I get the indirection of pointing to a RDDL doc or whatever else is useful.
  
 cheers, jim fuller 
  
 
 [1]: http://www.w3.org/TR/2001/NOTE-uri-clarification-20010921/

Received on Monday, 7 July 2003 04:07:50 UTC