Simon Fell: I believe that the version model Sam is pushing for makes a lot of sense, and feels more like the XML way, it'd be an interesting exercise to see if this is possible in any of the current toolkits.  ASP.NET does pretty well.  Let's take a concrete example. 

Suppose there were an API named blogger.newPost, which took six parameters.  Imagine a client and server were generated from this definition.  Suppose somebody were to want to add seventh argument, say, a title perhaps?  Imagine a new client and new server generated from this new definition.  Clearly old client will continue to talk to the old server, and the new client will talk to the new server, but what will happen when you mix and match?  And old client talking to a new server won't send a title.. therefore this parameter will be defaulted to null.  A new client will send a title to an old server which will ignore it. 

Not bad.  No configuration options required.  No new "meta" interfaces need to be defined.  There is a clear mechanism, via WSDL, to discover what a given server supports.  Things just work.

In my opinion, this is a key reason why HTML has been so successful - the contract between the producer and consumer isn't brittle.  Browsers are expected to ignore markup that they don't understand and to provide reasonable defaults for missing elements.

  9:47:02 AM    

James Snell: It's not a debate, Sam and I are just seeing the issue from two different points of view, both of which are equally valid. Personally, I think that it is time for James to take a vacation.  The actual sentence in the original article that jumped out at me:

WSDL version indication is done via the targetNamespace attribute of the definitions element. This namespace gives the SOAP messages meaning by relating the messages to a specific bit of code implemented somewhere on the server.

The net of the article is to make a specific recommendation about how one should change values in the abstract definition of a service based on how one particular SOAP implementation choses to dispatch in order to address an important issue such as versioning.

By the way - thanks to Dare, I've been in contact with the author of this paper, and he seems like a reasonable guy who is willing to listen and work through this, once he returns from TechEd.  A place I would like to be, and probably would have gone to had I been given enough notice, not because I am interested in the kool-aid (I'm up to date on my shots and am fully immunized), but because of all the kool people who are there: Mike, Keith, Jonathan, Don, Brad, and doubtless countless others, including Scott.

  9:11:41 AM