Free Science, One Paper at a Time (Neuron Culture Moving Party Track 6)


Yet if a more open scientific publishing landscape may seem inevitable, it’s hardly clear how to get there. Talk of inevitability hasn’t much helped Jonathan Eisen get his father’s papers out in the open. He has struggled to find the right leverage point, or perhaps the right tool, to lift them onto a platform any more prominent than his own web page.

A few months ago, however, Jonathan ran into a tool that added some leverage —  and just might chip away as well at the calcified matrix in which his father’s and others’ scientific work  has been stuck.

It is the simplest of academic tools: a desktop reference manager called Mendeley. Yet it comes with an extra dimension: a website at which you can share papers you like, creating a metadata-rich index that can lead other users to your user profile and papers, and vice-versa. You load your bibliography up — all those papers on cognitive neuroscience, say, or dark energy, or, if you’re Jonathan Eisen, evolutionary biology and extremophile bacteria — and Mendeley’s algorithms link you up with papers you might have overlooked and the researchers who wrote, read, or collected them. Maybe, Jonathan wondered aloud on Twitter, he could create a posthumous profile for his dad and post his papers up there. Mendeley promptly told him he could. He did. Howard Eisen now has his own Mendeley page, with all 41 of his papers listed and 24 of them uploaded  as PDFs. Now you, as well as anyone in the research community who takes a minute to sign up at Mendeley, can find and read them, add them to your libraries. Since Mendeley now has over 800,000 members and is growing at an accelerating rates, this puts Howard Eisen and his work if not in science’s mainstream, then in a sizeable and fast-growing tributary.

Jonathan also likes Mendeley because it seems to advance the larger open-science agenda. It’s a sort of friendly Trojan horse. You download a reference manager — a good one, and free — and suddenly have a tool that can help open science.

“Smart,” says Leslie Carr, director of the Web Science Training Center at the University of Southampton. “Most people who’ve tried to create software to drive open science have started off on the web and tried to encourage people to share. Mendeley starts where the researchers already are, with a tool researchers need, which is a desktop reference manager to manage their bibliography and  organize their thoughts. Then the very act of looking for more papers leads them into an open science model based on sharing.”

Mendeley exists because its CEO and co-founder, Victor Henning, needed a tool to understand better the cross-disciplinary mountain of literature he’d compiled for his thesis at Bauhaus-University of Weimar. “I had this huge trove of papers from different disciplines,” Henning told me, “and I wanted to see where the connections and overlaps and gaps were.” But when he looked for software to do this, he found nothing he liked.

“This was 2004, 2005. was happening. By then I was collaborating a lot, and we were talking about doing a couple different people’s data. Then we realized you could do a social version of this. Why not map a bunch of people’s data? That would give an even better picture of the ideas in play.” By this time, being a business student, he was thinking: startup. He also realized he didn’t like most of the reference managers on the market. So he thought: let’s roll that in too.

Thus emerged Mendeley. The name rose from its dual mission: Mendeleev was the Russian chemist who created the periodic table, which organized the known elements into a structure that suggests the properties of other elements still to be found. Mendel was the 19th-century monk and botanist who saw that crossing two packets of information could yield a third packet that derived but differed from the first two. Two nice models of how science works.

The focus on the paper came of pure necessity. The  American bank robber Willie Sutton said he robbed banks because “that’s where the money is.” At least for now, the paper is where the data are. But the people at Mendeley know quite well that a) the paper will get unbundled and in many functions displaced and b) they’re now grasping a bundle with a bunch more stuff in it. But they’re most interested in the threads that run from paper to paper. They mean to charge not for the bundled information but for helping people find the connections between the bundles. The company offers a free version that accommodates smallish libraries. If your library runs bigger than 500 MB, you can pay a $10 a month to run the company’s algorithms and store a copy of your data and papers in the cloud. If you’re a company or a department or simply someone who wants to run some highly sophisticated or customized analyses on aggregated scientific data — and on the all-important hivemind indications of what’s newly hot — you can pay more, providing Mendeley another income stream. Mendeley also talks of striking a deal, maybe, with publishers to make papers available on a rough iTunes model: a buck a paper, perhaps, with algorithms running in the background to help you find papers you’d like but don’t know about.

Many feel this model holds a lot of potential not only to make papers available more freely (or cheaply) but to help unbundle and redistribute the functions now unnecessarily bundled with the paper. But can Mendeley do this? It seems to possess the vision and flexibility of mind. While Mendeley is necessarily focused on the PDF right now, for instance, there’s no reason it can’t adjust its databases and algorithms to index, share, and analyze the importance of contributions other than traditionally published papers; they can do new metrics. And the company’s advanced programming interface, or API, recently published, should allow outside developers to create modules and add-ons to track new metrics, including author identifiers such as ORCID.

The chassis, then, can accommodate changes under the hood. The trickier part may be getting the steering right — that is, creating a UI that offers a powerful and full-featured but easy and intuitive way to use both the traditional reference manager and the broader social, sharing, and analytical tools.

They’re still working on that. “Most of the people I talk to who’ve used this,” says Leslie Carr, “think that the desktop and the web sharing aren’t as well integrated as they could be, from a software perspective, and that the analytic tools aren’t as accessible or transparent as they should be.” I find the same thing myself: Many of the metrics and connections between papers aren’t accessible on the desktop, presumably because they require the server’s data and processing power, and finding them on the web interface feels vaguely opaque. Even when you find some relationships, you worry you’re missing something.

Yet the company seems both open and responsive. When users pressed last summer for more hivemind information and more fluid sharing, the company substantially upgraded the website’s social-sharing module. It was quick to produce iPhone and iPad versions of its software. In general it seems fairly nimble and eager to meet user needs.

On the other hand, a lot could stop them. They could run out of money; with over 40 employees, their burn rate is high, but then again, their funding angels seem both confident and deeply pocketed. They could get sued. They could fail to add features fast enough to satisfy demanding users.  They could not quite create the magic that software needs to be transformative. In short, they’ll need what any gamechanger needs: a good concept, some serious programming and promotional chops, and luck.

Mendeley chief scientist Jason Hoyt thinks the real killer app in open science will not be software but … the researcher. He recently made the call in a blog post titled “Dear researcher, which side of history will you be on?”

For the past three centuries, he noted, technology has prevented us from fulfilling Panizzi’s dream of fast, free science. But the technology is there now, and so are the business models, as PLoS has shown. So what is the revolution waiting for.

It is waiting, wrote Hoyt, “for us, the researchers.”

We could choose to publish in only Open Access. We could choose to reward tenure for Open Data. We could choose to only reward publications or data that are proven to be reused and make either a marked economic or research impact. Instead, we choose to follow a model that promotes prestige as the primary objective. …

“The future, I suspect, will look upon our society and practice with regards to scientific knowledge-share as we similarly do now with the Dark Ages. Each time we hold back data or publish research that isn’t immediately open to all, we have chosen to be on the wrong side of history.

He has a point. It’s interesting, for instance, to imagine what would happen if researchers and university librarians got together and created a global version of the sort of revolt that the University of California librarians threatened. “You get all the librarians together on this,” says Cameron Neylon, a director at the UK’s Science and Technology Facilities Council and an academic editor at PLoS, “and this is pretty much over.” And Librarians at the Ramparts sure makes a nice image.

Jonathan Eisen, too, thinks that opening science will require the researchers to step up. But he suspects they won’t step up in number until reward systems offer some incentive more tangible than being on history’s good side. Only then will the upslope ease. In the meantime, Jonathan continues to push his father’s papers up that hill, and he waits to see how well Mendeley, among his other efforts, can help pull them up into the open. Jonathan tends to push hard in strong spurts around Father’s Day, make some progress, then set the load down a while before resuming.

“It’s one of those things that’s just going to take some time,” he says. “I didn’t think it would be quite so hard. But we’ll get there.”

~ ~ ~

Copyright 2011 David Dobbs. All rights reserved. You may excerpt short sections, as per fair use, as long as you link back to this article. For permission to reprint in whole, please drop me a line.

*Disclosure: I sometimes write for Nature Publishing Group and have friends both there and at PLoS.

NOTE: In the week or so after this published, Jonathan Eisen was inspired to substantially complete the job of assembling his father’s publications at Mendeley. See my short follow-up post here.


May 12, 2011: • Fixed some typos. • Changed pounds to dollars. • The original version made it sound as if all PLoS journals evaluated submitted papers based only on method, rather than method plus significance. The current version is corrected to state that only PLoS’s flagship journal, PLoS One, uses that streamlined method of peer review.

May 13, 2011: • Corrected i.d. of Jonathan Eisen’s position at PLoS, where is an academic editor-in-chief of PLoS Biology. • Prior version called PLoS One the “flagship journal” of PLoS. A couple people differed. Changed to note that it is PLoS’s most innovative and cross-disciplinary journal. • Clarified criteria by which PLoS One referees papers. • Corrected description of KnowledgeBlog, which was created by Phillip Lord and Robert Stevens, not Peter Murray-Rust (who told me about it).


Resources & more reading — a post of links I put together

Jonathan Eisen’s page on trying to find his dad’s papers

Jonathan Eisen Frees (Almost All) His Father’s Papers (a follow-up) | Neuron Culture

How to Crack Open Science – from ScienceOnline

A TED Talk to Open Your Eyes to Open Science

Google to Host Terabytes of Open-Source Science Data

Open Access | Wired Science | (all Wired Science stories tagged “open science”)

Open Data | Wired Science | (all Wired Science stories tagged “open data”)

Open-Access Debate: Public Library of Science Responds ( story from 2007)

Leave a Reply

Your email address will not be published. Required fields are marked *