A couple weeks back, in a post entitled "The time may be ripe for ‘Guerrilla SOA’", Joe McKendrick wrote about a meme that Jim Webber of Thoughtworks initially introduced in an interview with Infoq.com. The idea is that if SOA adoption for a company requires a massive mobilization of resources and people, it's often a non-starter. It's impractical to justify the initial investment, and it's impractical to realize actual benefits from that approach.  A more reasonable approach may be to adopt a practical, pragmatic approach to employing SOA, with a very incremental, iterative approach.  Use SOA in an initial project, and evaluate your results.  Then apply the lessons learned and broaden the use.  And so on.  This is something that I had independently described, although not using the Guerilla metaphor, as a common practice among Microsoft customers in a separate interview on infoq.com.

Are we reaching a tipping point on the real world approach to employing SOA?