SOA is a big tent. On the one hand you have developers working with sketchy APIs, services of questionable granularity, metadata, and what the building trades call 'finished carpentry' - stuff like making some architect's REST scheme actually work at a reasonable speed.
On the other hand you have the strategists and architects who are sometimes not too very high above the daily fray. These people must reckon with ESBs, governance, dueling developers, business alignments and 15 types of middleware and data architecture.
Nowhere does the dichotomy appear more clearly than in the recent muddy river of Microsoft marketing around 'cloud' computing and SOA. At its PDC conference the company grandly showcased its Azure computing OS and associated updates to the .NET framework. But Microsoft's folks more than once positioned cloud as a replacement for SOA.
This is far from the truth. If you can't do SOA, you will not get ahead in the cloud. If you fail at creating efficient services down here on earth, the cloud will not turn you into a winner. In one group or another, people at Microsoft know this is true. We put the PDC 'the-cloud-is-better-than-SOA' gaffe to marketing people - the people that reliably bring you 'the next big thing.' Hey, people! Why didn't you have Azure evolving from Vista?
Microsoft is a big tent now too. The developers and the platform folks see most things differently, including SOA and the cloud. As in much of application
Notes: Read about SOA doings at PDC in Rich Seeley's reports from L.A. 's Staples Center.
In other news we have word on an Evans Data study (''Services reuse drives ROI for SOA''); a discussion with noted architect Leonard Feshkens on SOA skills (''Domain bridging, skills of abstraction define the software architect''); and ZapThink's Jason Bloomberg on how SOA can drive business change (''Business agility as an emergent property of SOA''). Come gather in the SearchSOA tent!