(1) There is some nice geometry going on here. The tree appears to be embedded in hyperbolic space (is that what "hypertree" means?), visualized using the Poinc

(1) There is some nice geometry going on here. The tree appears to be embedded in hyperbolic space (is that what "hypertree" means?), visualized using the Poincare disk model. Edges of the tree are drawn as geodesics. Node-to-node moves are accomplished via translations of hyperbolic space. One result of this is that you can demonstrate to yourself that a rotation in hyperbolic space can be realized as the composition of translations -- which is not true in Euclidean space. Pick three nodes, repeatedly click on them in order (1, 2, 3, 1, 2, 3, etc.), and watch how the space slowly rotates. (2) There is a bit of a problem with the UI details, at least in my browser (Firefox). The clickable area for each node appears to be roughly a square whose top is the horizontal diameter of the red disc representing the node. Thus, clicking in the upper half of a red disc accomplishes nothing. And trying to click in the center of a disc gets you what you want about half of the time.