Introduction

To scientists, data is the basis of their career. Scientific discoveries follow are only as good as the data generated by experimental methods and analysis tools. Surely scientists would never willingly place their data in danger of unverifiable corruption. Yet scientists do this routinely whenever they use an analysis tool to which they do not have access to the underlying algorithm and its source code implementation. What occurs is a black box transformation the researcher hopes is correct. This is analogous to conducting a microarray expression experiment and not knowing the exact DNA sequences which are spotted on the array. We are forced to accept on faith that the manufacturer is incapable of error and cannot choose poor sequences or mix up the clone identities [1].

This article discusses an alternative to this situation, the Open Source Software (OSS) development model, which guarantees all users the right to examine any the algorithm and its underlying implementation.

Depending on who you are, we have different goals for this article.

For all readers: we hope this article helps you become more knowledgeable about the OSS movement and how it helps produce better software and better research, more efficiently, for everyone.

• For Software Users, this article should help you find software to solve your specific needs.
• For Software Developers, we will provide an analysis that better enables you to decide when releasing your project as OSS is more or less likely to benefit you.
• For Policy Makers, we argue that the OSS development model is important to the speed and growth of technological innovation so that you can help make OSS become more widespread, both in its use and production.

Because many of the issues presented in this article are complex, the authors have provided a companion website to the article, that provides more detail on many of the issues discussed herein, as well as a broader listing available OSS packages[2].