For SOA to soar like an eagle, not waddle like a turkey I think there are 5 changes needed:
- SOA sponsors need to find a Line of Business executive with a specific project where they can use the technology/concept. Then another, then another. If there are not enough volunteers give up - at least for a couple of years. It is the classic solution looking for a problem.
- Think constraints. Project team of no more than 3. Project time no more than 3 months. Total burdened cost less than $ 100K. And that's for the pilot. Even less for on-going ones. You are competing with web 2.0 team sizes and budgets.
- If you are a tools vendor, package everything in to one "wrapper". Do not show customers your sausage factory of 10-15-20 products
- If you are an application vendor, for every time you mention SOA plans, make sure you have 5 mentions of application functionality improvements.
- Everyone - it's about web services, not people services. Send the consultants home.
For those who call me a SOA cynic, they forget. I love architectures. Check out my blog name. Just talk to me in dollars and cents not acronyms like WSDL and UDDI. Give those - and the hype about SOA - a REST.
Update: Phil Wainewright says SOA may have hit rock bottom and should re-emerge with more realistic expectations.