V2: What should the API look like for horizontally mapped data via tuples?

Historically, we have the multi-generic Query<T1,T2, ... ,TResult> API for this which is ugly as sin and clunky to use.

We have an opportunity to improve things using tuples, with the nice advantage that since tuples are only one T (for any reasonable number of partitions), it won't cause overload explosion. Long term, this might even be seen as a replacement to the Query<T1,T2, ..., TResult> API (in a breaking major, etc).

I've thrown some early ideas into a branch; I want to invite feedback, or other original ideas - it is entirely possible that I'm missing an obvious + better way of doing this!

Again, the target scenario here is when you're selecting from multiple tables in one horizontal piece, and want to access the pieces into separate DTOs - in SQL terms:

select c.*, a.*
from Customer c
inner join CustomerAddresses ca on ca.CustomerId = c.Id
inner join Address a on a.Id = ca.AddressId

which is going to return a bunch of columns that we want to map (for each row) into a Customer object and an Address object.

Note: the discussion of what to do for duplicated Address records or Customer records (same Id, multiple instances vs single instance) is separate - please don't get distracted by that aspect (although that's also something we'd like to improve at a later data)

Here are a lot of examples in that test branch