Why aren't more researchers using open source?

Academic researchers would benefit by ditching their expensive, slow, and non-specific commercial software for open source versions.
447 readers like this
447 readers like this
a magnifying glass looking at a brain illustration

Opensource.com

Academic researchers depend on a variety of highly specialized software to power their studies. The commercial software options in common use are expensive; either investigators must purchase a large number of licenses for common applications like data analysis tools, or they have to buy costly single licenses for specialized software, such as an application for a specific laboratory device.

It's surprising, then, that so few researchers are using open source instead of expensive commercial software. Not only does open source produce cost savings over commercial options—money that can be invested back into research—it also provides researchers the opportunity to develop software for their own specific needs, then share it with others doing similar work.

Assume your research group invents a new technology or methodology to advance your field and publishes your findings in a notable journal. In order for commercial software to be developed for your technology, someone in a software company has to read your paper, trigger the development process, and bring the software to market maturity. Depending on the project's complexity, this could take several years.

You can shortcut the process significantly if you develop the software at the same time you develop your new technology. Because you don't have to wait for commercial software to come to market, you can start widely using your innovation immediately after (or even before, in some cases) your research is published. And, once data is available on how many people are using your open source software, commercial software vendors can determine whether it's a viable market for them, then view your source code to learn the requirements for developing their own version.

Here's a more specific example. Surface chemists use a pendant drop tensiometer to determine the surface tension of fluids (the reason why water clusters into drops and insects can stride over it). For research and industrial applications, it is important to know the surface tension of specific fluids, since it influences the ability to form emulsions, foams, etc. Many pendant drop tensiometer vendors also provide commercial software that analyzes the shape of droplets to determine surface tension. Our research group developed an open source software called OpenCapsule that not only performs the same kind of analysis, it also determines the elastic properties of a membrane coating a droplet, which was impossible before. OpenCapsule provides researchers a new, free, GPL-licensed software tool to characterize elastic nanomaterials. We're hoping that new contributors will help us optimize OpenCapsule and push the project further.

Open source offers many advantages for researchers, including ready-to-use, cutting-edge technology; cost savings; and faster development. No commercial vendor can know your needs as well as you know them yourself, so why not leverage your own research and become independent from software vendors by developing your own open source software?

Tags
Jonas Hegemann is a physicist from Germany. In February 2018 he finished his dissertation in theoretical biophysics including computer simulations on cell biology and video analyses of elastic materials.

8 Comments

Great article, Jonas. I agree that open source software is an important part of reproducibility. From my experience working at an American university, the general reason is because commercial vendors provide the software at steep discount to get graduate students hooked. The students graduate and get jobs that will need to buy full-price licenses.

Thanks for your comment, Ben.

The reproducibility issue you address is very important I think. Often it is difficult to reproduce the findings of a publication, because the corresponding code is not published with the paper. Things will be much easier and efficient if more research related code is published. This also concerns code with a limited application field.

As a student, I got several licenses of Microsoft products for free, so the discount thing you describe also exists in germany. But the more had to work with them, the more I realized that I Linux is the better choice. It is probably the responsibility of the university to equally hint at open source alternatives.

In reply to by Ben Cotton

I think a big reason is that most science researchers are NOT programmers and don't want to be. Anymore than most chemists blow their own glassware or most biologists make their own microscopes. While writing software might require fewer tools than making one's own lab equipment, it isn't clear that it requires less skill. And even if there are already written open source alternatives, they are probably harder to find. If you have written software that might be useful in other labs, I would suggest that you think about how to publicize it. Depending on your field, you might be able to do a poster at a poster session at a conference that you are attending for other reasons.

Many researchers go on to work in the companies that develop the software. Developing the software is dependent on having people who understand the science, not on being able to develop software. Various journals and funding bodies already require the provision of source code, this needs to both become mandatory and a requirement for further funding.

Great article. Same reasons I feel research use open source software. I discover many researchers not know about open source software available. They follow what other research people use and not think.

The author perhaps speaks for his own lab or environment. As an academic of many years in the related area (bioinformatics), I can say that globally bioinformatics research is done almost exclusively open source software tools. Specifically I would mention Linux, Python, Perl, R (and many open source packages developed on this ecosystem), but the list does not stop there.

This is a rather uninformed article. Or rather opinion.

I'm not talking about my own group since we're running nearly exclusively on open source four a long time, with only a few exceptions like e.g. wolfram mathematica. But I also know many groups which run completely on commercial products like origin, comsol multiphysics or similar software packages. Open source is, as you correctly say, actually widely spread in the academic field, but there are still many places where it should definitely be used more intense!

In reply to by NoBull1 (not verified)

I guess one professor I know uses commercial Matlab just because she learned it first and wants to concentrate on writing papers rather than learning new softwares. I also met one lecturer or researcher at the university who said said that he uses SPSS instead of R because if there is a bug or program needs some new property, there has to be a commercial support so the property will be implemented.

Creative Commons LicenseThis work is licensed under a Creative Commons Attribution-Share Alike 4.0 International License.