Use of set vs. list in Aggregation.java. In general do we create a best practice document for modelling?

From a pure modelling perspective, if your collection is of type Predicate then it makes sense to use a Set, though the items really have implement hashCode() and equals() to properly match the RDF equality being used for this to make correct sense.

The collection types RdfBag and RdfAlt are best modelled by Collection, with the concrete instantiations using a List subclass (since they can contain duplicates).

For the other collection types (RdfSeq and RdfList) the field must be an array or a List.

In addition to the above mentioned constraints on items imposed by Set's, Set's also have the advantage or disadvantage (depending on your use cases) of effectively causing an eager load of all the items in the Set - this is because when put an item into a Set the Set must at the very least invoke equals() on the item to ensure uniqueness, and that invocation will cause OTM to load the item.

Milestone 0.9 deleted

Really need to get rid of these tickets or create new ones with tighter focus.