Truly open-source science

Back to index page

Next article: A web of human relationships

Science, in some ways, is the ultimate open-source project. Just as programmers anywhere can re-use free or open-source code, scientists anywhere can build on published results. But with computer code, replication is as simple as cutting and pasting text – whereas in the lab, reproducing a result, making a copy of a specific tool, or adapting a method is fraught with problems. Published science focuses on the results: the process is described, but the nitty-gritty background details are often missing. How many times was the experiment tried before it worked? What brand of tube was used? How important is the timing? What do the raw data look like? Such details can be crucial when re-using published results, but they are rarely found in scientific papers.

Worse still, most research is never published: most experiments fail, give negative results, or provide information that 'doesn't fit' into the final publication. This information is nearly impossible to publish in scientific journals, but can be incredibly valuable to a researcher thinking 'why hasn't someone tried this?' as they decide whether to spend the next six months giving it a go. An enormous body of scientific knowledge remains unavailable to those who need it most. It exists only in the minds of individual researchers—and in their notebooks.

Most scientists, even today, keep a paper notebook, but an increasing number are adopting a range of electronic systems. This raises the possibility of something truly radical: making the entire primary research record available by putting lab notebooks on the web. Jean-Claude Bradley, a chemist at Drexel University, was the first to promote this approach and has coined the term 'Open Notebook Science' to describe it. A small but growing number of scientists, including my research group, are starting to explore opening their notebooks to the world. The slogan 'no insider information' captures what we are trying to do: make all the information that is available to the researcher, available to everyone else.

This raises a number of technical and social issues.  Blogs or Wikis can be used as lab journals, as can documents simply placed on the web, but it can be difficult to make a complete record available in a form that is useful to other researchers. Files are often in proprietary formats, setups can be difficult to describe or even photograph, data often require extensive annotation, and so on. Overcoming these problems improves the quality of the record, but much more effort is needed to create it.

These technical issues are just that: issues to be solved. The bigger challenge is overcoming the social and emotional hurdle of exposing your work to the outside world. Scientific practice has become extremely secretive, due to the growth of interest in commercial exploitation and to increasing competition for places on the career ladder.  Concerns of 'being scooped' or having results or ideas stolen are not to be taken lightly. Radical sharing is probably not for everyone and is clearly not appropriate where patenting is important, or where private information is used.

There is great potential for a more open practice to lead to a more efficient scientific community, where less time is wasted obtaining information that has already been recorded, or trying things that are known not to work. As with anything, finding the right balance for each case and indeed for each person is important. Those of us who are adopting Open Notebook Science in our research groups would argue that the balance needs to move more towards the open.

Cameron Neylon, STFC.

www.tinyurl.com/5wxng7