What's the Latest Development?
An editorial in the scientific journal Nature argues that software used to interpret data from scientific experiments must remain open source so that other scientists can attempt to repeat the experiment. The authors point out that, today, scientists and engineers rely heavily on computer software to interpret data before arriving at conclusions. And since reproducing the results of an experiment is essential to confirming its validity, scientists must have access to more than a description of the code used in the course of an experiment.
What's the Big Idea?
Currently, many scientific journals simply request descriptions of computer code used in experiments. However, "ambiguity in the descriptions and errors in the code" are cited as two reasons why this practice is insufficient. In some cases, however, the institutions which fund experiments may prevent their researchers from releasing code because of copyright protections. Another reason given for why code is not presently released is that scientists stand to be embarrassed if errors are found in their procedures.
Photo credit: shutterstock.com