How are clusters weighted in global queue?

In other words, what’s the most effective way to setup proper interleaving in global queue - documents, references, or tags? Will it treat documents as “decks” and therefore override everything else? Or pick a tag and drill down on it even if it contains rem from multiple docs? Or follow a long chain of references? Is the intention to toggle the booleans in the settings until only one is active or is there an explicit hierarchy when all are? I suppose this is the stuff best understood through looking the code itself, but since that is not an option, I have to ask here.

1 Like