A short introduction with my take on the subject as part of a multi-presenter panel.
I used my now much used slide of Jean-Luc Picard (see “Just make it SOA, Number One“) to get over the simple message about what SOA should be: just give me the service, and spare me the details, please…
In the eGovernment space, this is not as easy to do as it is to state however: in the European space, an absence of a central authority or legal basis for action in public administration; strongly nationally focussed work and desires for high-levels of accountability.
SOA is an enabler, not an end in itself, as Colin Wallis stressed also in his remarks: get your mission and high-level “enterprise architecture” sorted out first, and worry about SOA later.
SOA is not about one or more technology stacks: it’s also a frame of mind. The experience of working with the Reference Model for SOA taught me that we had taken a lot for granted and just sort of assumed hat basic terms and concepts in SOA are shared by everyone – wrong. We discovered that even some of the most “obvious” tenets of SOA really needed unpacking and nailing down (someone here today talked about some things being so close to your nose that you can’t focus on them – nicely put). None of us claimed in that work that all the definitions and concepts were correct, but the importance was planting a flag for a common terminology that crosses all SOA implementations and stacks.
If we want to roll out eGovernment services that can be used in cross-agency and cross-border situations, then SOA would seem to be the best underlying paradigm with which to start (particularly if you start with the Reference Model) – my question was what else is needed as an enabler? XML as the underlying data structure is certainly a very good idea, but my concern is also to provide some “semantic glue” to help agencies to work together – particularly if you work in complex environments such as the European Union, in which services are to be composed between 27 Member States working with more than 30 official languages. SO one key question is: would some sort of “eGovernment upper ontology” or “reference information model” be a step in that direction? That is certainly a point that could be raised at this week’s eGovernment technical committee meeting.