So the project is split into 3 parts:
* Stack
* Client SDK
* Server SDK
Stack is all the built-in datatypes, structures, channels, serialization, transports, etc... along with the scaffolding to build a client or server. Specified mostly in part 4 and 6 of the specs.
The Client and Server SDKs then build on the stack and make it easier to build an actual application (but still just an SDK... you have to write code).
This split is common in the OPC world. Stacks can be open, easily maintained, smaller in scope, and basically a proof of the basics of the spec. It also makes interoperability easier when you can build on the same stack as others. SDKs are usually provided by various toolkit vendors, built on top of the stack for whatever language.
One of the things that makes Milo a little special is that it provides both stack and SDK under a free, open source, license. Typically you'll find the stacks have a restrictive or dual license and the SDKs from toolkit vendors are commercial.
