In the current specification the tree:remainingItems is put on a tree:Node. I wonder whether it makes more sense to put this in the relation instead.
Semantically:
Then we would make remainingItems relative to the relation, and not an inherent property of a Node. This means a Node, depending on which client arrived on that node in a certain way, may have a different number of remainingItems.
Algorithmically:
It would be very useful for clients to understand the selectivity of a certain relation when performing joins
In the current specification the tree:remainingItems is put on a tree:Node. I wonder whether it makes more sense to put this in the relation instead.
Semantically:
Algorithmically: