Open Source

Because you cannot see how most commercial code works, that is, you cannot examine the source code, the commercial approach made it difficult for programmers to share source code & advance the field.

In the 1980s, Richard Stallman founded the “free” or “open source” software movement in reaction to the commercial approach.

The open source movement defined legal protections for sharing source code: Copyleft allows software to be used, modified, and distributed freely as long as anything derived from it is bound by the same condition. The GNU General Public Licenses are a series of copyleft licenses. Note that not all free software is open source, but most open source software is free. That is, just because you don't pay money for some software, does not make it open source if you don't have access to the human-readable source code. Access to the source code, and not just free executables, is what defines the Open Source movement.

The creator of Linux said this about open source:

The more widely available the source code is for public testing, scrutiny, and experimentation, the more rapidly all the bugs will be discovered. Linus Torvalds

The cost of software is determined by its license. Much (but not all!) open source software is distributed with permissive licenses which impose no costs. It's important to know what the software license requires. For example, you may need to credit the software author or disclaim liability. Finally, there are a variety of open source licenses, some less restrictive than the GNU Public License. Poldrack expresses a strong preference for minimally restrictive licensing to support reproducibility:

In the context of software, [we express] a strong preference for minimally restrictive licensing [e.g., using the permissive BSD (Berkeley Software Distribution), MIT, or Apache licenses rather than licenses with restrictions on commercial use or viral licenses such as the GNU Public License]

Poldrack RA, Gorgolewski KJ, Varoquaux G (2019) Computational and informatic advances for reproducible data analysis in neuroimaging. Annual Review of Biomedical Data Science 2: 119-138.

Watch this 3-minute video clip

Programming 1 (10)