Introduction

Developed by the Institute for Quantitative Social Science, Harvard University (IQSS), this draft document will contain best practices and points to resources for developing statistical software that is:

• robust,

• user-friendly,

• persistent,

• attributable,

• enables reproducible research.

The document begins with general best practices that are broadly applicable to developing statistical software to achieve these goals. It then provides resources for how to implement the best practices that are specific to particular programming languages.

Using the best practices

The best practices should not be viewed as infallible dictates that must be followed slavishly. All software projects are different and have different requirements and expectations. The best practices should be instead treated as a series of questions that developers should ask when they begin and throughout the development of a statistical software project. For example, would the intended users of this project benefit from having an open source license? Would it be useful to create a website specifically for this piece of software detailing its full set of functions? In general the answer to these sorts of questions will be "yes".

Contributing

This document is a work in progress and contributions are highly encouraged! That's why we are developing it on GitBook.