Tom Jordahl wrote: > > How about we just prevent two different services with the same name from getting >deployed? > > Humm.. But how can we detect this error case as different from a re-deploy of the >same service.
-1 The names we choose to key off of may not have any real meaning, and in fact be computer generated. If you remember, the IBM Smalltalk guy had to pretty much sit out the entire event because of a similar consideration. This doesn't lead to happy customers. For that matter it is quite possible (one could even say likely) that someday somebody will define a bunch of services which only differ based on SOAPAction (see http://radio.weblogs.com/0101679/stories/2002/01/25/whatObjectDoesSoapAccess.html if you are interested in why). Some form of generalized mapping (with reasonable defaults) will ultimately be required. - Sam Ruby
