Hmm.  I do like the OASIS SOA-RM draft; but that seems to take a strong stand against the definition "Formerly called a "distributed objects" architecture, the SOA term was coined at the turn of the century as Web services were evolving. CORBA and DCOM are examples of earlier SOAs.".  That seems the most wrong of the three in Len's original post, partly because it conflates implementation technologies and architectural principles, and partly because both the W3C WS-Arch document and the OASIS SOA-RM draft take pains to distinguish service architectures from distributed object architectures.  
 
That's why W3C WS-Arch failed, in a nutshell:  "web services" technologies such as SOAP (and, ahem, plain ol HTTP) can be used to *implement* RPC/distributed object architectures that expose objects and methods over the web, and can implement service architectures that hide these and only expose the concept of a service contract.  So, "web services" are a set of architecturally-neutral technologies, whereas "service architectures" are a technologically neutral design principles.
 
All my personal opinion, of course.  



> Date: Tue, 10 Jan 2006 10:51:35 -0500
> From: chiusano_joseph@b...
> To: len.bullard@i...; xml-dev@l...
> Subject: RE: [xml-dev] Will The Real SOA Please Sit Down?
> 
> I would say that they are all correct, and all equally useful. The best
> situation would be to have a single definition that comprises all of
> these, as they each point out valuable and important aspects of SOA.
> 
> Now would be a good time for me to mention that the OASIS SOA Reference
> Model (SOA-RM) TC has just approved as a Committee Draft the following
> spec:
> http://www.oasis-open.org/committees/download.php/15966/wd-soa-rm-11.pdf
> (distribution is unlimited, document is public). We have not yet updated
Express yourself instantly with MSN Messenger! MSN Messenger