Elsevier and the executable paper

Have you ever wanted to make your paper come alive with data? The results of a recent competition may help that to become a reality.

Elsevier are fond of tapping into the scientific community’s ideas on the future of publishing. A few years back they issued a grand challenge on ‘Knowledge enhancement in the life sciences‘, and also had a contest about the article of the future, or ‘Article 2.0’ as they called it. They have another competition running right now to develop library and searching apps on their SciVerse platform. And they have just announced the winners of the Executable Paper Grand Challenge.

Now, I’m sure there are plenty of papers that should be executed, or made to disappear, but this challenge was not about those. The competition was intended to address the problem that computer science research results can be difficult to reproduce. Paper is a static medium and most current electronic journals reproduce that quality by relying on the PDF format. The Beyond the PDF project is looking to address this problem, and various publishers are trying things out too.

The first prize (US$10k plus an iPad) in Elsevier’s competition went to a Polish team, with a US collaborator, who developed a system (called the Collage Authoring Environment) which “allows researchers to create papers by combining narrative discussion with snippets of executable code”. Second prize went to a Dutch-German team that developed a “web portal for creating and sharing executable research papers”. Third prize went to a US team for their “universal identifier for computational results”. Papers describing the winning entries have been published in an Elsevier journal (see links on the competition home page).

Although the competition sounds at first to be just about publishing computer science, I was interested to see that the winning entry does address itself to the wider world of publishing scientific research and the question of research data accessibility. The introduction to their paper says

Furthermore, the actual data which is the result of (and frequently the basis for) published research is not preserved and made accessible in a coherent manner – thus, the reader of a scientific paper must often take the author’s word that the professed results and conclusions are, in fact, valid. Clearly, the scientific paper itself no longer conveys sufficient information to enable reviewers and other readers to judge it on its own merits.

The paper also includes a nice review of current trends in scientific publishing, including a poke at Web 2.0, noting tartly that

…naïve approaches to sharing and reusing are insufficient and inappropriate in the realm of science.

I do not know enough about other work going on this area to be able to judge whether Collage will be the system that finally makes it possible to combine narrative, data and processing all in one place, but I look forward to the day when this can happen.