Meet the people making science more open and accessible to everyone, including scientists themselves
We tend to think of science as magic. It diagnoses and cures the world of its ills and provides both the building blocks and the blueprint for such incredible feats as the eradication of polio, the rise of the Internet and space travel. But magic is incalculable. It’s immaterial, irrational. Science, by contrast, is precise, articulate, considered and based entirely upon evidence and verifiability.
Yet much of the field’s output sits locked away behind expensive paywalls and research institution intranets. Scientists, who seek to discover and release the secrets of the universe, from the tiniest atom through every organism and all the way up to the great cosmos, tend to be ironically closed-off and distant from both the public eye and each other. No longer will this be the case, however, if the growing open-science movement gets its way.
Scientists all around the world are breaking free from the shackles of the old system in pursuit of something better. They dream of a world of openness, where every paper, every draft, every tweak to an experimental algorithm can be checked, explored and reproduced by others.
Noon Silk, the organiser of the Open Science Workshop in Melbourne
“I really like the idea of speeding up the understanding of a paper by having it [be] interactive, so I can just go in and change things and learn what they’re trying to tell me a lot faster,” says Noon Silk, a 10-year software veteran who left a journeyman life of coding to study physics and mathematics (and eventually, he hopes, to do research in quantum computing).
He’s the organiser of a new Melbourne initiative called Open Science Workshops. It’s designed to help scientists of all stripes become familiar with the concepts and tools of open science — things like GitHub, which is a service that stores multiple versions of code and documents in a public repository that anyone can “fork” and expand in a “cloned” version all their own. “It occurred to me that a lot of my programming friends know about these version control systems,” Silk explains, “but a lot of people at the university and [the] researchers I know don’t know about them.”
Noon Silk explains to Richard Moss how he uses GitHub, and what makes it a useful tool to a research student.
There’s something of a high barrier to entry for people without a computer science background, however, so Silk thought he could “give a venue for people to come along [and] have everything de-jargonified.”
Still, even with events like this and Mozilla’s Software Carpentry workshops, the very fact that the tools are hard to learn is a huge barrier to the democratisation that open science entails. And that’s a big concern to Ian Mulvany, the head of technology at eLife Sciences Publications and an editorial board member at the Journal of Open Research Software. “Constructing tools which by definition only experts can use and clouding them in a language of expertise is a really good way to stop people getting involved,” Mulvany tells me at the Open Science Workshop.
Ian Mulvany (right) chats to an attendee at the Open Science Workshop
His role in the war for open science is currently focused on the publishing side. He’s helped build the eLife Lens publishing platform, which seeks to put academic scientific papers online in a web-native format in which figures, charts, videos and references all appear in a column alongside the text at the click or tap of a link.
ELife is far from the only open science publisher. Others include the highly-regarded non-profit Public Library of Science, which was founded a decade ago after broad circulation of an open letter by three leading scientists — Nobel Prize winner Harold Varmus, Stanford University biochemist Patrick O. Brown and computational biologist Michael Eisen. And even the US military backs a number of open access publications and projects.
Open access has clear benefits, its proponents argue. The issue is not merely one of philosophy — that scientific knowledge should be available to the public. It’s a matter of practicality. Contrary to popular imagery, science is not often accomplished by lone geniuses locked away from the world but by collaboration between individuals and groups of varying backgrounds who share ideas and debate hypothesises. But much of that is getting lost in an environment of paywalls.
The public is missing out, too. Silk gives an example of someone who wants to find out about cancer treatments. Perhaps they have been diagnosed. Or maybe someone they care about. “They want to read all the papers they can,” he says, “but every paper costs them $30. So of course they’re not going to pay that.”
Photos from the Open Science Workshop.
Freeing the Data
There’s a lot more to open science than just open access. The other, tougher front in this war against the old system is open data coupled with what is sometimes described as open notebook — making the research material itself, from initial conception through every phase of development, experimentation and analysis all the way to drafting, re-drafting and finalising that one document that traditionally represents the whole body of research. Not everyone likes the idea of making all their data available.
“I think it kind of happens with programming,” Silk says. “People think, ‘Ah, I want to keep my code secret.’ And then these big-name guys just publish their stuff and everyone’s just like, ‘Oh, well, it’s acceptable now.’ So I kind of secretly hope that same thing will happen [with science].”
Get over that hurdle, though, and there’s still the matter of tech savviness coupled with the difficulty of accurate documentation. “Ideally, from the moment you begin your work, it’s structured in a way that would be instantly reproducible,” Mulvany says. “But reality is always a lot messier.”
He cites the example of Phil Bourne, a co-director of the Protein Data Bank — a key resource for biologists — and associate director for data at the National Institute for Health in the US. “His lab is one of the most open labs in the world,” Mulvany explains, “and they went to try and reproduce one of their own results [with] the original people who were working on the experiments. They had all the data. They had written the original paper. And it took them over a year to reproduce the original result.”
They had to figure it all out again, even with all of the original material. Science is not always as precise and easily reproducible as it would have you believe.
Ian Mulvany presents the eLife publishing platform
It gets worse, Mulvany explains, when you also consider the churn rate of the field. Only eight per cent of science graduates stay in the field as academic researchers, he says. “So you’ve got a 92 per cent shed rate. And as these people leave, their data and code just rots. It evaporates,” Mulvany continues. “Because it was on that laptop or on that folder that was managed by a sysadmin in the department but that person’s moved. This stuff just sort of evaporates.
“The world as we think it is as described by science is mostly held on trust and not on evidence. All of this [open-science movement] is about trying to move to a place where it’s a slightly better world than the one where we are trusting what’s in the literature.”
Silk echoes the sentiment. “You just make the world a better place, maybe, by doing more effective research,” he says. “Because maybe the research people are doing is useful to other people, and if they can learn about it and do it more effectively, life would be a bit better.
The charts above show media mentions of open science (left) and frequency of the keywords “open science” and “open access” in pages crawled by search engine Google since 2004 (right). The very idea of “open science” dates back less than a decade, with the term as it’s now used seemingly invented in November 2006 as a follow-on from a much older discussion about the ideas of science and open-source software. Notable media coverage started in 2008 (courtesy of the Boston Globe).
Democratised Research
Jon Stanger learns the basics of GitHub from Noon Silk
Once scientists come around to the idea of open data, there’s still one last thing to consider. Jon Stanger calls it democratic science. He’s a multi-skilled, cross-disciplinary scientist from New Zealand who moved to Melbourne to find work after his project of eight years — which essentially involved cloning new tissue for burn victims — lost its funding.
Science started out as patronage by very rich people, Stanger explains. “Then the government started to realise that the innovations that we produce by this research is really important to their economies. So they started to pour a bit of money in from government funds, but the problem is that the general cycle of funding is that you as an individual get taxed.”
A proportion of your income goes into funding scientific projects that fit whatever science agenda the current government creates. “Then all of the scientists have to try and hammer and shape their particular expertise and specialities and projects they’re working on to fit the shape the government has now decreed that for this particular year is the flavour of the year,” Stanger explains.
“They do some research, get the funding, pay some of the government’s money to publishing organisations to put all of their research behind paywalls. And then you as an individual in society basically has no access to any of what I’ve done. The only time you’re ever going to see any result of it is when a corporation exploits research that was funded by you to produce a product that you then buy.”
Something doesn’t quite add up here. And Stanger believes the solution is to put scientific funding in the hands of the masses, who could each chip in a couple of dollars a month — or whatever amount they choose — to support research they care about. It’s essentially an amalgamation of the crowd patronage website Patreon and crowd funding website Kickstarter, except in his vision it’s run by a panel of scientists who’ll rate an idea according to its blue sky-ness (that is, whether it is consistent with current scientific consensus and it’s believed to be an achievable project). And scientists seeking funding would have to write two descriptions — one brief layman’s explanation, plus one detailed scientific breakdown.
It’s necessary, Stanger argues, because existing funding models favour imbalance. And because science is a “black swan” endeavour. “You as a scientist have an idea,” he explains, “and you basically have small failures for long periods of time until you have your one moment of success.”
You don’t know when this moment will come, however. “So the best way to get the best success out of a large group of scientists is to keep a low level of funding sustaining many different people. Then when someone has a great idea that could revolutionise society with just a million dollars, then you inject a bunch of money.”
JOUR90008 