Efficient and Flexible Incremental Parsing

Requires a context-free grammar.

Runs in O(t+s lg N) time, where t is terminal symbols, s is modification sites, and N is number of trees in the node.

Node reuse: nodes which were parsed in a previous parse, should be reused where the same values exist in the new version of the document.

Manages a parsing stack of trees which are manipulated to build the final parse tree.