Dense and Sparse Matrix Classes Using the C++ Standard Template Library

The C++ programming language has undergone signiicant changes since its inception in the 1980's, but has now reached a relatively steady state. Standard C++ now includes a general library of container classes, the Standard Template Library (STL). These developments are rapidly changing the styles used in C++ class programming. The paper has dual purposes: It provides an introduction to STL for C++ programmers, and it develops an eecient matrix class library , built upon STL, which provides functionality useful in areas such as computational economics, nance, mathematical programming and statistics. This library, which is freely available, comprises a full set of vector and matrix operations using both dense and sparse implementations. The paper discusses approaches towards and pitfalls in constructing C++ concrete data types, and has references for further on-line information.