In order to support hashMaps we have to come up with a plan for how
they'll be persisted into the graph. For example, should the hashmap
become a kind of "blank" node off of which all it's entries become
relationships?
Or should the hashmap be persisted into a binary bucket? There is a
bad side effect in that it'd not be searchable. For example
Taylor