A new section of Nature examines the software and websites that make research easier.
Asked to list the essential tools of a scientist’s trade, most would probably think first of hardware: the microscope, the telescope, the mass spectrometer, the genome sequencer, the test-tube. But just as important to today’s data-wranglers are software — Excel, ChemDraw, MATLAB — and the programming languages used to create it, such as Python, R and SQL. Such tools are integral to modern research practice, whether for analysing or visualizing data, sharing files, collaborating, writing up papers, publishing, searching the literature or simply organizing one’s work. And although software engineers have often overlooked science in favour of more lucrative markets — think Flappy Bird, Instagram and iTunes — software, websites and apps designed specifically for researchers are blossoming.
Partly in response to this flourishing sector, Nature this week introduces a new section to help readers keep up to date. The Toolbox pages will collect the journal’s writing on software tools and websites that researchers use to work more efficiently, or in new ways. Find them online at nature.com/toolbox, and monthly in print.
Barely a week goes by without the appearance of a website offering improved research productivity, or the launch of a start-up firm hoping that its unique idea will change scientific workflow. Toolbox will aim to guide the perplexed through the maze of sites and programs, discussing their similarities and distinguishing points. But it will be a community-driven resource, with scientists in various fields who work heavily with data or programs offering thoughts on their most commonly used software. The site will also collect Nature’s writing on the broader context of online research — from open data to citizen science and crowd-funding.
In this issue, for instance, the section reviews a recent trend: the emergence of ‘recommendation engines’ that sift the flood of literature so that scientists can find relevant papers and information (see page 129). Future articles will include a look at the iPython interactive-computing project and its applications for scientists, and an examination of websites that promise to help researchers to collaborate on research papers.
A little software literacy can make any researcher’s daily life more efficient. Version-control systems such as Git, for instance, help to record changes made to files, to allow recall and analysis of past and evolving work. Websites such as GitHub (a favourite of software engineers, but increasingly of scientists too) build on these platforms to help researchers to work collaboratively on a research paper, or to ensure that data analysis is clear and reproducible.
Such programming tools are already the daily bread of the data scientists, bioinformaticians and climate modellers among Nature’s readers. But jargon can make them off-putting for non-coders. At the same time, it can be hard to tell which of the more polished, graphically friendly software packages are worth investing your time in. So on the Toolbox website, scientists will be able to share their recommendations for particular software (both commercial and free). As a taster, this week the Software Carpentry movement, which teaches basic software skills to researchers, explains its motivations and operations. A bad workman may blame his tools; a good scientist needs to keep track of them.