This post is part of my collaborative research with Shinsei Bank on highly-evolvable enterprise software.  It is licensed under the Creative Commons Attribution-ShareAlike 3.0 license.  I am indebted to Jay Dvivedi and his team at Shinsei Bank for supporting this research.  All errors are my own.

Judging from comments on earlier posts, there seems to be some confusion whether these posts articulate design principles that Shinsei seeks to follow, versus descriptions of how Shinsei’s systems actually work in practice.  Jay has mentioned to me on several occasions that he believes current implementations capture only “15%” (I think the number should be understood as an impressionistic estimate rather than a precise value) of the design principles. From studies of specific systems at Shinsei, I’ve observed implementations that adhere to these design principles to greater or lesser degrees, but none that could be characterized as complete, reference implementations.

Thus, the ideas developed here should be understood as a sketch of systems as they might be and are, at least in some cases, in the process of becoming, rather than as descriptive research about existing systems.  That said, audited cost estimates suggest that Shinsei may be capable of building systems at a small fraction (perhaps less than one tenth) of the cost of traditional vendors, and I think it’s likely that these architectural principles–even if only partially implemented–contribute to this performance differential.

In conclusion, then, this research falls squarely in the domain of theory building rather than theory testing.  Ultimately, the value of such research is perhaps a metaphysical question; I refer skeptics to my defense here.