On Monday, I blogged about Dave Linthicum's comments regarding the tenure of the term SOA, along with some fellow panel members who were also in the room. Today Dave answered and rated those comments,. Dana Gardner got an A+. Dave flunked me. But while I may have been distracted, it certainly wasn't due to Dave's alleged "brut good looks."
Now, I might not have gotten the exact quote right (I didn't put it in quotation marks) but I certainly understood what Dave meant. I just don't happen to agree with him about the term going away in 5 years. I do agree that SOA is good enterprise architecture. As a matter of fact I have been speaking and writing about it for the past 15 years (I KNOW Web services aren't that old - so please don't flunk me in math). In fact, I agree with just about everything Dave was saying (don't sweat your looks Dave, we love you for your mind). Dave's final comment in his post was "I don't think the value or the notion of SOA will go away, it will just be discussed in the context of architecture in general."
My initial take was "I don't think that time frame is remotely possible given the current state of the maturity of SOA adoption.". SOA is as much about application architecture as it is about enterprise architecture. The people in the room who Dave was speaking to were enterprise architects. They are the first to get SOA. But it's going to take developers and those responsible for application architecture a very long time to transition skill sets, because the architecture itself does not say HOW to create reusable services. Loosely coupled does not sufficiently describe how to determine the right level of granularity to maximize reuse. While the architecture components and infrastructure technologies are becoming mature, application developers are still lost at sea. They will need the SOA beacon for a long time to come.
That is, unless the failure rate becomes so high that the term itself becomes discredited - like what happened with Client/Server. And 5 years is certainly long enough for that to happen. So i think that IF SOA as a term falls out of use it will be because it was implemented so poorly - not because it was done well and became an integrated practice. But regardless of what we call it, it is likely to become THE way we do IT. Maybe in 10-15.
Leave a comment