Do we need scopes?

Commit https://github.com/corvus-dotnet/Corvus.Storage/pull/2/commits/1d5b049f09d18dad8ec2ee31762db9b0a4a0937e removed a bunch of half-baked code supporting the notion of "scopes".

A scope was some sort of context within which a logical container was resolved to an actual container. The two candidate scope-like things we knew about were:

* tenants
* the current process

However, it ended up looking more complex than it was worth. Instead, we've gone with a configuration-driven approach, with tenancy then layered on top of this. (And tenancy is just tenancy, rather than one particular example of this abstract thing called scope.)

As for per-process scopes, it actually looked more complex than just having process wide configuration settings and passing those in.

However, we might want to make a decision at some point in the future as to whether this was the right decision in hindsight, or whether some more reified notion of scope might in fact have been useful.