I just found this column in Nature discussing the need for scientists to publish the code they used. I’m still amazed that this is not getting more attention. If a proper-written Methods section is mandatory in a paper, why not the code that produced the results?
Among the excuses for not publishing the code (and the reasons why they are not valid most of the time) that the column identifies are:
- It is not common practice.
- People will pick holes and demand support and bug fixes.
- The code is valuable intellectual property that belongs to my institution.
- It is too much work to polish the code.
Even when the Methods section may be enough to be able to reproduce your results, why condemn someone else to go through the whole process of debugging some code to make it work? The code may not be pretty, may require some obscure software, or may be more convoluted than it has to be, but it is incredibly valuable and a time-saver for researchers. This kind of code is also very useful for students, it lets them learn how research is done in their area.
Of course, this requires some planning and careful file management, hopefully universities and societies will start promoting code publishing to make this practice more common.
Barnes, Nick. 2010. Publish your computer code: it is good enough. Nature 467: 753. doi:10.1038/467753a