Geometry Template Library for STL-like 2D Operations

There is a proliferation of geometric algorithms and data types with no existing mechanism to unify geometric programming in C++. The geometry template library (GTL) provides geometry concepts and concept mapping through traits as well as algorithms parameterized by conceptual geometric data type to provide a unified library of fundamental geometric algorithms that is interoperable with existing geometric data types without the need for data copy conversion. Specific concepts and algorithms provided in GTL focus on high performance/capacity 2D polygon manipulation, especially polygon clipping. The applicationprogramming interface (API) for invoking algorithms is based on overloading of generic free functions by concepts. Overloaded generic operator syntax for polygon clipping Booleans (see Figure 1) and the supporting operator templates are provided to make the API highly productive and abstract away the details of algorithms from their usage. The library was implemented in Intel Corporation to converge the programming of geometric manipulations in C++ while providing best in class runtime and memory performance for Booleans operations. This paper discusses the specific needs of generic geometry programming and how those needs are met by the concepts-based type system that makes the generic API possible.