Screaming Penguin

While I am writing "what Cooper is thinking about" stuff, I want to draw your attention to this presentation by Major Joe Paiva, Enterprise Architect for the US Army. I find it really compelling on a number of levels. First, it is one of the most pragmatic perspectives on the role of SOA in a large enterprise I have ever heard. There is no hype, but rather it is the perspective of SOA as a design pattern that solves a particular problem set, rather than some kind of magic bullet. It is to be applied where sensible and not where it is an imposition. While he does use some language that if I heard a manager use would be a "red flag" to me, it is in fairly minor points. Otherwise, spot on. Secondly, the discussion of the failings of WS-SE and our current SOA models in dealing with access control and data granularity are something that, which Major Paiva sees them as special-case to the DoD, strike me as something that we are going to need to address in the larger world of Web 2.0 and loosely coupled integrated services. I have been thinking about a new project of late -- a applet base provider with a browser plugin enhanced provider to allow user-by-user registry of HTTP data services that are proxied at the client. So that, for a limited time I could allow access to my GMail Atom feed to, say, the JPenguin aggregator, without ever giving it my login information, so that data I want to share is channeled through my browser and authentication is restricted. As there are more and more pay-to-play podcasts (kiss my ass Air America, btw) and things like private mail feeds or private filesystem feeds in the operating system, this kind of cross-application access control on limited basis becomes important. While I am looking at it at the client-server end, the major talks about the problems on enterprise to enterprise level, and they sound very similar. Lastly, one thing that stood out was the assertion that the Army has one job: to fight and kill the enemies of the US. I have spoken to this in this space before, but I want to go back to it. While I know this mission statement is coming down from the leadership and not the responsibility of the Major, I find it to be one of the problems with the modern military. The reason we have Halliburton scandals, delapidated levies in NOLA and closures locally at Lake Lanier, and chaos in post invasion Iraq is because the non-warfighting business lines of the Army have been allowed to atrophy. The founding fathers of this country were opposed to having a standing army because they felt that having an Army designed to fight wars in idle would provide an impetus to put it to use. (See also H.G. Wells on democracies and "adventurous foreign policies.) This is a very real issue, and in effect, what doomed the empire of Alexander. I feel that a large part of the problem we have now with the Neocon set stems from this line of thinking. However, I don't think it is the only one. Both FDR and Julius Caesar understood that you could maintain a large standing peaceable army by enhancing the roles it performs. Caesar's built roads. FDR's built dams and levvies. Moreover, losing aspects of the quartermaster role in the military means that in response to things like a hurricane, the massive logistical support needed to, say, feed and water a large number of people in an isolated/decimated area is lost. The one exception to this multiroling, however, is posse comitatus. To quote a line from (forgive me) Battlestar Galactica, "The purpose of the police is the keep the peace. The purpose of the army is to engage enemies of the state. If the army becomes the police, the people start to look a lot like enemies of the state." In this case, the separation of roles is very important.