Innovator Story: Wrestling the Octopus

While working in media, I was used to ‘looking for the story’ and then trying to make it as persuasive, engaging, emotional and relevant to the audience as possible. That work was always under scrutiny – we were judged on how many people watched, and how much they liked our programmes. These were felt to be appropriate measures of success: entertaining people was core to the job.

I was rather shocked to find that academic researchers were being judged on exactly the same metrics. Researchers are under huge pressure to ‘find a story’ through their data, to write it up in a way that persuades the reader that this story is in some way ‘the truth’, and that it is both new and vitally important.

Watching this on the one hand, and on the other seeing the debates about ‘questionable research practices’ (means of eliminating data that doesn’t fit a hypothesis), misconduct and a toxic research culture in many institutions, it struck me that these problems stemmed from the pressures to publish papers in the format of stories, as if written by a journalist. By contrast, though, for researchers, these measures of success are entirely inappropriate. A good story is not the same as good science.

When Charles Ebikeme of the International Council for Science visited Cambridge in April 2017 and posed the question, “If we could restart the entire scientific system, what would it look like?”, my answer was: start again with the scientific publishing system so that it serves the needs of science and scientists, not those of publishers – and I wrote down the structure that I now call Octopus in a couple of hours’ hurried typing. Looking around at the academic world it seems that so many problems have their root in the way we currently publish work: as complete ‘stories’ based around ‘proving’ a hypothesis. Octopus is based on the premise that we throw all of that away and start again.

Charles’ reaction? “That sounds great – so how are you going to do it?”

Looking around at the academic world it seems that so many problems have their root in the way we currently publish work: as complete ‘stories’ based around ‘proving’ a hypothesis. Octopus is based on the premise that we throw all of that away and start again.

– Alex Freeman

Octopus is:

1. A fully digital-first platform (i.e. all text is hypertext, with no PDFs, to make it fully searchable and taggable with embedded metadata). Knowledge is inherently a web of interconnected ideas, and the internet can represent that in a way that printed text can’t. We need to abandon the legacy of the 18th and 19th-century ‘journals’ and use 21st-century tools to their utmost.
2. Completely language-agnostic (using automatic language translation), so that every user reads and writes in their language of choice, maximising access and broadening the base of people contributing to global scientific knowledge.
3. Free to read, and free to publish. This is just fundamental: scientific knowledge is a public good.
4. And Octopus upholds that the unit of publication is not a ‘paper’ but instead one of eight shorter forms:

• Scientific problem
• Hypothesis
• Method/protocol
• Results/data
• Analysis
• Interpretation
• Implementation/translation
• Review

This is where Octopus’ structure becomes different from most other scientific publication platforms.

Breaking up the unit of publication allows smaller author groups and hence more accessibility for researchers with limited resources, meritocracy/accountability for work done and greater recognition of the need for specialisation (e.g. for statisticians for analysis or technicians in protocol design).

It also encourages fast publication of each piece of scientific work (to establish priority). It’s just ridiculous that researchers have to hold on to a good idea for years whilst they try to collect data, force that data to support the hypothesis as strongly as possible, and then waste months of time writing it up (a long, wasteful and tortuous process). In Octopus, publication can be instant.

Crucially, this structure also values each part of the scientific process regardless of subsequent or previous work. For example, a hypothesis can be valued regardless of whether subsequent results support or refute it; a data set can be published and valued regardless of size or what it suggests about previous work. When I did my doctorate, I had an idea (which I still think was a good one) just as I was writing up: I had collected a small amount of data, but I couldn’t get that published, so I put my thesis online as the only way to share my work. Removing the need for data to support a hypothesis before that hypothesis can be published removes the pressures for questionable research practices.

Treating reviews as an equal form of publication as the other seven forms, listed above, values perceptive reviewing, and hence collaboration as a vital part of the scientific process. It also means that readers can benefit from the insights of others who have reviewed a particular publication. There is potential to publish new versions of publications in the light of reviews, with the old version always accessible and the opportunity to include new authors, such as any reviewer who had significant input.

Octopus aims to reward good science rather than good storytelling, so there is also the rating of each publication by readers on each of three predefined criteria chosen to represent best practice in each kind of publication. This allows work to be judged on the most appropriate criteria and hence true quality recognised. I think this is crucial – things that we incentivise have to be the things that we really want to value. It allows us as a community to define what ‘good science’ means for each kind of work. ‘Originality’ is key to a good hypothesis – but not to good data!

And what of journals? I see their role as changing from being the publishers of primary research to being the home of the narrative. I don’t read scientific papers for entertainment – I read them to try to extract the facts from them. That’s why I think that removing the narrative structure from primary research is good. I would, though, be very happy to read a well-written editorial which brought facts together, out of the primary research home of Octopus and into a narrative structure – an article that told me the latest findings in a field, and gave an expert perspective on them.

That, then, is how I see Octopus changing science.

My vision for the future is one in which scientific research is faster, more collaborative, more meritocratic, less wasteful, and done in a more egalitarian and hence happy environment – all through changing the way research is shared.

– Alex Freeman

Developing Octopus was my first introduction to the open-science community: for several months, I nervously approached people with the idea, expecting it to be dismissed – either because there was a fundamental flaw in the concept, or because it was too naively ambitious. I quickly found how supportive and encouraging people were, despite my naivety and lack of knowledge of the world of open-science platforms and tools. Getting a place at the eLife Innovation Sprint in 2018 was the first big step for me and Octopus.

One of the responses I've had to my approach of being completely open about my ideas and my plan for all the code to be open-source is that this could be a death knell for the project.

“What's stopping one of the big players from stealing your idea and building a paywalled version of it?”

Well, absolutely nothing, of course...except economics: building commercial software to do something this radically different from anything they currently do is a huge barrier for existing publishers. They have the problem of legacy. Maintaining and developing the software, so that it works for all users of whatever speciality, would also be a vast undertaking: harnessing the work done by so many open-science and open-source geniuses is far more efficient. Instead of trying to own and control it all, let others develop tools and make improvements so that it is a community-owned venture.

Secondly, though, if someone built a system identical to Octopus except for a charge for users then that would, in my view, be better than not having a system like this at all. Although Octopus is at its core a free resource designed to open access to scientific research to a far broader community, it is more than that. The structure of it is designed to make scientific research more effective, more efficient and more meritocratic – and adding commerciality wouldn't destroy those aspects.

So I'm committed to being open in every respect, and will be interested to see where that journey leads me. So far, I've been amazed at the generosity, friendliness and talent within the open-source community. To a non-technical person like me, it was a daunting step reaching out into the unknown and wondering who, if anyone, would respond – but people have, and they have freely given their time, their advice and their expertise, and that is an incredibly valuable resource. I have been emboldened by continuous encouragement from almost everyone I’ve met, and continue to meet (although there’s always a healthy dose of realism), and I am more determined than ever to make Octopus a reality.

Wrestling Octopus into a system that can replace the current scientific publishing establishment is undoubtedly a challenge. Technically, I think it’s perfectly achievable. Politically, it will take a lot of work.

I’m well aware of the need to find the ideal user bases for a first iteration of Octopus, and have been so pleased to be inundated with emails from people who are sitting on troves of ideas, protocols and data for which they have no other publishing outlet, and from people in developing countries who have little access to traditional journals. I also know the key role that funders and institutions may play in driving researchers to publish in Octopus, if they adopt it as a means of judging people’s contributions to science.

What I have been most surprised about, though, is finding so many people even within the commercial worlds of scientific publishing who also share this vision and are excited at what Octopus could bring. Although it might mean a huge change to the institutions within which they work, many people care deeply about science itself and the role it has to play in the world – and that is, for many, more important than anything else.

So, at last, my love of storytelling may have been trumped by my passion for science. Both are important forces in the world, but the power of narrative can be too strong for a seedling scientific idea. In order to let that idea branch and flourish in its own directions, we need to avoid tying it to a stake and pruning it hard to keep it straight. And we need to avoid turning research scientists into journal-ists.

#

We welcome comments, questions and feedback. Please annotate publicly on the article or contact us at innovation [at] elifesciences [dot] org.

Are you interested in contributing to open-source projects like Octopus to drive forward open science? Register your interest for the eLife Innovation Sprint in September 2019.

Do you have an idea or innovation to share? Send a short outline for a Labs blogpost to innovation [at] elifesciences [dot] org.

For the latest in innovation, eLife Labs and new open-source tools, sign up for our technology and innovation newsletter. You can also follow @eLifeInnovation on Twitter.