Interesting point and the reuse discussion has been, well, 'reused' by SOA proponents and opponents alike. The reality is that reuse became a bit of an amorphous term and more of an ideal once IT realized that the concept was stronger than the execution in a lot of instances.

However, I take issue with the way that Woods continues, 'what will save SOA is to split the process of creating services into two stages. In the first, you just use or build whatever services you need to support a new application. You don't worry about governance or reusability. You just get the users what they need.'

To me, this approach could be one of the keys to a stalled or failed SOA initiative. Think about it. If you subscribe to 'get er done' tactics, this approach may seem reasonable. However, it may not take into account the longer-term impact of how that application or parts of it will inevitably extend beyond the original team it was intended for. 

Actually, not worrying about governance or reusability strikes me as a bit counter-intuitive to the whole SOA value proposition. After all, reusability was one of the most often uttered phrases in the SOA pitch. Yet the lack of reusability is often due to inconsistent and erroneous development approaches. Approaches that may have never been put in place, nor would they have reached the deployment phase, had governance been in place at the get go to alert the architects and developers that something may be awry in the code.

In fairness, the opinion article takes into account the second stage where 'you look at the portfolio of applications and services you created and attempt to identify patterns and opportunities to collapse several of the stage-one applications and services into reusable service and common applications.' 

This is a sound strategy though perhaps cancels out phase one? Further, it brings up the somewhat circular conversation that's happening these days about services ownership, creation and cost. Who, in fact, does own the original service and can they charge a colleague to use parts of it as part of a reuse strategy? While that is a discussion for another day, what the industry can agree on by and large is the incremental approach to SOA. 

What I believe, however, is that agreement on design and architecture in accordance to business goals before services are created is a critical first step. And, of course, I'm going to tell you that this is where governance has the most impact on the longer-term benefits of the SOA project.

Agree or disagree with the Forbes's piece on 'SOA Revival?' Drop me a line and let me know as I don't think this one is truly black or white.

-Jeff