Introduction

The ideas we present here are actually more far-reaching. We can apply them to streamline processes that involve running mathematical computations and producing documents that display the computations and their results.

Python vs C++

In our framework, Python does the

• Managing (of complex tasks, processes and computations),
• Talking (to a wide variety of systems, 3rd party model libraries, and data sources),
• Thinking (translating, e.g., financial contract logic into computational tasks, sending the tasks to our C++ model libraries),
• Light lifting (light computational tasks), and
• Reporting (aggregating results and producing reports).

C++ does the

• Heavy lifting - carying out tasks sent from Python (Monte Carlo and/or PDE model libraries implemented in C++).

Live documentation

Section Using Python and C++ libraries is an example of live documentation. The Python scripts that produced the plots are invoked from within the document source code and are run when building the document. The results are automatically inserted in the html document.