--- In service-orientated-architecture@yahoogroups.com, "Steve Jones" <[EMAIL PROTECTED]> wrote: > > 2008/6/23 Rob Eamon <[EMAIL PROTECTED]>: > > --- In service-orientated-architecture@yahoogroups.com, "Steve Jones" > > > > As you point out below, the identification of events without an > > obvious home aren't the problem. It tells you where your service > > analysis is lacking. > > Or to be more accurate it told us that there were people we should > be speaking to who were not engaged. This was mainly because IT had > never come into contact with them before.
Excellent. > The Events (and other interaction models) are the lines between the > services, this means they are very much part of SOBA and fit nicely > within the overall context. So event analysis certainly helped and > it helped because it was within a consistent context. A consistent context seems very much like a Good Thing. Being event driven is another principle to apply to BA. Clearly, events fall in nicely as the lines between services, but I still think EDA is distinct from SOA, both styles to be applied to a BA. SOA says "there will be service providers and consumers which interact." The interaction is left undefined. EDA says "there will be events passed between producers and consumers." The exact nature of the producers and the consumers is left undefined. Applying SO and ED principles to a BA says "there will be service providers and consumers which interact, and that interaction will, at least in part, be the exchange of events." Neither style subsumes the other. Knowing that events and services are major components aids in context and analysis. > I think its important to address _all_ interactions explicitly and > that some of these will be events, the process I do has only three > parts > > "What" - The services > "Who" - external actors (including external services) > "Why" - reasons that they interact > > The Why is all about the interaction part and its that part of the > analysis that EDA comes into, often highlighting more "Who" > and "What" examples (especially "Who" IME). I agree that all interactions, event or otherwise, should be addressed (avoiding analysis paralysis of course!). I think there is value in examining events as part of the "what" discovery. -Rob
