I think I can see the logic of having source-specific satellites in the Raw Data Vault, and I've concluded that it is probably something we will want/need to do.
However, there are SOME cases where there is enough commonality between source system field names and the source systems are stable and such... making it not really that important in those cases.
I'd hate to lose too much consistency by doing some tables one way and some another, but two patterns isn't bad. Has anyone implemented with some source-specific satellites and some shared across-source-system satellites in the RDV? If so, are there things you learned later that would make you choose differently?