My ZDNet blogging colleague David Berlind talks about the "uncomputer," based on the fact that the computer as we know it is evolving into a set of APIs.
Perhaps we can extend this logic and call service-oriented architecture the "untechnology."
Untechnology because applications may evolve into sets of services that are one step removed from technology, and do not depend on one specific technology over another. While we're at it, SOA has nothing to do with Web services, either. It's entirely possible to build an SOA on your own or somebody else's specifications, without the accepted Web services standards.
However, some say we're taking this concept too far, and selling "architecture," versus implementable technology, is bound to be a flop. I recently caught up with PolarLake's Ronan Bradley, who say that we need both the vision and the tangible goods to execute on the SOA promise. (The full Q&A is posted here at Webservices.Org.)
"It’s amazing; SOA seems to be the first technology that doesn’t require technology," he said. "The big vendors say it doesn’t really matter you do in technology terms, so long as you follow the virtuous road of SOA-based architectures. If you then look at the measurements coming back about SOA failures, not surprisingly, if you’re taking an approach that’s just about virtue, and not about action, you end up with failure. And it’s only going to get worse."
Bradley argues that we need tangible products that can deliver the system-to-system integration that SOA promises.
Bradley is not down on SOA by any stretch. Ten years out, he sees us looking back up SOA as laying "the first layer of the true integration of enterprises. It was the first layer that actually allowed people to start getting benefit from the integration between applications. And it was the first true switch from all the business value being locked in applications, to the business value actually extending across not only the applications, but also the whole network."
Can untechnology make technology work this well?