Property 2 allows an efficient binary search to be used to find the intersection points of two monotone chains. For many types of real-world data, these properties eliminate a large number of segment comparisons, producing substantial speed gains.
One of the goals of this implementation of MonotoneChains is to be as space and time efficient as possible. One design choice that aids this is that a MonotoneChain is based on a subarray of a list of points. This means that new arrays of points (potentially very large) do not have to be allocated.
MonotoneChains support the following kinds of queries:
index
The monotone chain search algorithm attempts to optimize performance by not calling the select action on chain segments which it can determine are not in the search envelope. However, it *may* call the select action on segments which do not intersect the search envelope. This saves on the overhead of checking envelope intersection each time, since clients may be able to do this more efficiently.
The monotone chain search algorithm attempts to optimize performance by not calling the overlap action on chain segments which it can determine do not overlap. However, it *may* call the overlap action on segments which do not actually interact. This saves on the overhead of checking intersection each time, since clients may be able to do this more efficiently.