Last weekend I gave a pair of talks about Service Oriented Architectures at the Great Lakes Software Symposium. Creating the talks was difficult; in the end, I created an abstract overview followed by an in-depth look at Axis and the WS-I specs. I was anxious to see the feedback from the students, which ranged all over the map.
One student suggested I just change the name of both talks to "SOAP with Java and Axis". Another thought the topic was too advanced, while yet another found it boring and basic. Most thought it was the right mix of material. But almost all asked the same question at the end:
"Is this really useful? Why do I want to know this stuff?"
I was stunned. Almost too stunned to formulate an answer. First, I mumbled something about Gartner studies showing that 60-70% of IT shops were investing in web services and/or SOAs in 2005 and 2006. Then I stopped myself. Nobody here cared about Gartner studies. They didn't want to know why their bosses should care; they wanted to know why THEY should care. I think I answered it ok, but I'm putting it down here because I can do it more succinctly:
We should care because Java isn't alone in the universe. We should care because the definition of the Enterprise Application has changed radically. We should care because our heavy-weight J2EE frameworks and application servers aren't useful when one of the components is a remote .NET object, and another is a Ruby script, and another is PHP. We should care because the transparent services we've been working so hard to get through those frameworks need to extend now to other frameworks.
SOA is largely about crafting a standardized syntax for communicating intent across platforms and runtimes. Web Services are about localized, standardized endpoints on a specific protocol that meet the needs of the SOA. What amazed me is not that people were new to the technologies behind them, but that they were insulated from the climate of cross-platform interop that seems to be welling out of the cracks of the Earth right now. Which makes me wonder if all the talking-brains-in-jars (a category I'm conceding that I inhabit) are missing something important, namely that regardless of what we or the Gartner group thinks, this imperative isn't reaching the trenches.