This page describes the main phases of the Mr. Tree algorithm. A short overview of all processors that operate in between these phases can be found here.
...
It should for example look for cycles in a tree and remove that cycles. This cycle breaking is done through eliminating the edge that destroys the tree property and putting that edge into a list. So this phase builds up a list with removed edges, so that the Tree tree algorithm can operate on the newly constructed graph which is now a tree. The previously removed edges could be reinserted during a post-processing.
...
This phase orders the nodes of each level by seperating separating the children of nodes into leaves and inner nodes. And It then fill whitespace fills whitespaces in the levels with corresponding leaves. It starts two levels above the deepest level, because the deepest level contains only nodes and therefore no reordering is necessary. And the level above the deepest level contains only children of the level above, which are ordered by the their parents.
...
Two tree traversals are used to produce the final x-coordinate of a node. The first traversal assigns the preliminary x-coordinate and modifier fields for each node; the second traversal computes the final x-coordinate of each node by summing the node's preliminary x-coordinate with the modifier fields of all of its ancestors. The final y-coordinate of the node is the height of the node's ancestors levels and the height nodes 's level and the adjust of the root location.
...