RE: Action Item 2004-07-01 Solution to 168/R114

Wow, way too much to follow up on here!

Gudge said:
> I'm with Jim. I think operations are an illusion. All they 
> are is a *WSDL level* construct for grouping messages in an 
> exchange. The fact that when using RPC style the messages 
> exchanges are typically In-Out with different input and 
> output messages for each operation does not mean that all 
> message exchanges look like that. 

Why do we bother to even define such a thing in WSDL, then?  It seems to
me that the operation, regardless of how it is determined, represents
the *intent* of the sender of the message.  Here, I'm sending you a
stock ticker symbol - if I'm invoking "getStockQuote", I want one thing
in response, and if I'm invoking "findCompanyZipCode" I want something
else.  Note that I'm not talking about RPC here, I'm just talking about
the contract between the sender and the receiver.

If we agree (and perhaps we don't...) that WSDL describes Web Service
*operations* and not just random collections of messages, then I believe
it's important to ensure that there are interoperable ways in which
senders and receivers can ensure that they understand how to invoke
particular operations.  Again, not talking about code-level dispatch
here, but rather "WSDL-level" operation identification.

Even if we DO say that WSDL is in fact about operations, and that we
should enable you to know which operation is being invoked, you're free
to design a "doIt" operation and then let anyone send you anything on
that single operation.  But if we DON'T say that operations are
important, then we seriously reduce the usability of the language for
those who DO care about them as "buckets of functionality".

I'm not entirely sure, but I think we may STILL be having the "what is a
Web Service" argument! :)

--Glen

Received on Wednesday, 14 July 2004 14:09:32 UTC