Once I know that, I can regenerate a set-tree implementation from nothing. The articles linked to come from an extremely SQL orientation which doesn't suit the way I (currently) think. Makes them a hard read, but I'll perserve because I want to learn how the coversion process works.
Fortunately, you spent time a year or so ago beating set-trees it into my head and for that I thank you! It's the kind of elegant design I like implementing.
Wade.