2009/4/1 Gregg Wonderly <[email protected]>: > Rob Eamon wrote: >> Dave Cutler springs to mind as well. And in a completely different >> field, Clarence "Kelly" Johnson. > > One of the compelling feeling that I get about people like Kelly Johnson, > are > that when someone can, and will take up a task to be done, and has the > technical > knowledge to manage the task, including managing the development of work > items > and the execution of the work items, good stuff gets done. > > What is happening with SOA, from my perspective, is that the breadth of the > subject is so wide, that no one technology, person, or thought process is > actually sufficient, without reducing choices to be made. For example, if > you > remove the choice of application protocol, that "simplifies" a lot of things > by > reducing what any related software has to deal with. If you removed > programming > language as a choice for software, that removes additional complication. > > So, we see people who continue to focus on smaller parts as in the choice of > "application protocol" such as WS-* vs REST, so that they can deal with the > complications of just that choice alone and restrict the number of issues to > deal with. > > The SOA chiefs want to ignore technology whereas the SOA warriors have to > consider it at every step and just can't seem to understand why the chiefs > think > it doesn't matter. > > Sure, it doesn't matter when you are at the 40k ft view, you see the SR-71 > and > see that it can carry out its mission. The people who are making sure the > elevons move when the control stick goes back and forth don't care about the > 40k > ft view, because its not where "works or doesn't work" is meaningful; i.e > you > can't get to 40k ft without working controls...
I think this is a good analogy in that if you are the sort of person who can design an aircraft then you are someone who is liable to have an extensive background in various different parts of a plane, some places you won't be a detailed expert in (e.g. RADAR) others might be your historical sweet spot (e.g. wing aerodynamics). The point is that your greatest skills are in compartmentalising the problem and enabling all of the people with the job of the detail to focus on their detail within the context of the whole solution but without having to understand the whole solution. You can't get to 40,000ft without working controls, engines, avionics, aerodynamics, pilot controls, RADAR and lots of other pieces. That is the real skill, breaking a problem down so it can be managed as a series of small projects that are combined in one large scale programme. I've said before that I think complex defence and ATC programmes have historically been Service Oriented because its the only way to tackle such challenges. A good SOA chief applies those skills to their enterprise. The other type of SOA chief, the one who disregards technology, is in part a fake. IMO if you haven't been through the detail of delivery you will struggle to break the problem down in an effective manner. I blogged about the differences between the two types of architects who hate technology http://service-architecture.blogspot.com/2008/04/terry-pratchett-architects.html I think a good chief is a ptarchitect. Steve > > Gregg Wonderly >
