I think I have sorted out the Bottom and Middle Tier Architectures for the most part. I've chosen to go with my own quirky namespacing as opposed to the general practices. So what, I like it better my way.

I have tested none of this at the moment. Specifically, I'm unsure if the Respository is going to work at all because of the new STEs and my Repository being built on old POCO and generics. I don't see any need to use a custom logic since I've got complete control over the datasource and I intend of extrapolating data detail there not in the BL. I'm really hoping to capitalize on STEs more efficient db connection handling, disconnected tracking and especially not require custom contracts for each type.

I think this approach pays off in TCO because additions to the data schema later down the road will not require a lot of coding and because learning to take this approach is more valuable in the long run for me.

Last edited Mar 7, 2010 at 12:57 AM by omayhemo, version 2


No comments yet.