I was very pleased last week to read that Nature published an editorial that argued for open sourcing software that had been used in the research leading up to a publication.
The ground principle is very simple: in order for claimed scientific results to be credible, it must be possible to verify those results. The key to doing that properly is the ability to reproduce these results. And if there is some piece of software code used to create these results, that is not made available to the scientific community, it is not possible for the wider community to reproduce the results.
Varies initiatives have been taken to ensure this academic principle is followed. For example, a conference like Sigmod installs a repeatability committee that will need the software used for the creation of accepted papers. Although this is good to ensure Sigmod’s papers have been thoroughly checked, it will not enable other researchers in the field to verify the results or to build on them.
Luckily, many scientists see open sourcing their code as a normal practice in their research, such as Daniel Lemire. The software project provides a solid basis for collaboration, and as such is an example of open innovation in academic communities.
One of the barriers that is cited against open sourcing code, is that the university may see commercial value in it and wish to commercialise it. It is important to realise, however, that there are many business models available for institutions that go far beyond just selling the raw outputs of a software projects. All of these still allow the institutions to create a viable business by adding value to what is available on the download page of a code hosting website. The true value of a software project is never just in the raw code.
At OSS Watch, we work with academic projects that develop software as part of their research and provide free support to the UK academic sector. So if you have a question about how to open source your code or how to deal with licences, we welcome you to contact us.