The next time you feel moved to comment on an article in the open-access online journal eLife, be prepared for a different user experience.
On 31 January, eLife announced it had adopted the open-source annotation service, Hypothesis, replacing its traditional commenting system. That’s the result of a year-long effort between the two services to make Hypothesis more amenable to the scholarly publishing community.
As I detailed in a 2015 Toolbox article, Hypothesis allows readers to mark up HTML pages and PDFs inline — as opposed to at the bottom of the page — as if they were using a highlighter pen and Post-It notes. Those annotations can be private, public, or restricted to groups, and they are persistent, popping up every time the reader revisits a page. In practice the tool is akin to the annotation systems on sites like medium.com and ReadCube, but where those tools work only on their host sites, Hypothesis works anywhere.
The system has some obvious benefits. It’s free and open-source, meaning anyone can use it. And it’s far easier to have a discussion near to the text in question than to require readers to scroll up from the bottom of the page. But several features limited its utility to scholarly publishers, says Giuliano Maciocci, Head of Product at eLife. For one thing, Hypothesis required users to have an account on the Hypothesis service. Comments could come from anyone, and publishers had no mechanism for identifying who those people were or moderating what they said. And, the user interface differed from that of the publishers themselves.
The new system rectifies those shortcomings, Maciocci says.
To view annotations — which can be made on any eLife content — users click the ‘annotations’ link at the top of each page. They can add new annotations simply by selecting the relevant text, logging in with their ORCiD ID. Publishers can add their own content and moderate conversations, and tweak the user interface to better match their site design.
Other features strengthen the tool’s utility for personal record-keeping and communication. Users can review their old annotations by clicking the user profile icon at the top right of each page. Each annotation has its own unique and sharable URL. And, by clicking on the username behind any public comment, users can learn more about the people with whom they are interacting.
According to Maciocci, the ability to moderate conversations was a key development, eliminating the “wild west free-for-all” that effectively required publishers “to abdicate … any significant level of control over the tone and appropriateness of the discussion that might have gone on on their site through the Hypothesis plugin.”
Maciocci acknowledges that requiring users to authenticate themselves may limit commentary, as some users may be disinclined to comment publicly. But Maciocci says the journal editors felt that people who are commenting publicly on research, should be willing to stand behind those statements. “If they’re uncomfortable expressing those views, then a public commenting system is not the place for them. They can reach out to the corresponding author, they can approach things in a quiet or more discrete manner.” Comments are not pre-screened, Maciocci emphasizes, but users can flag objectionable content.
For Dan Whaley, founder and CEO of Hypothesis, the eLife announcement represents “a huge milestone.”
“Our working thesis has always been that annotation was a powerful evolution of what people have previously implemented, an evolution of comments,” Whaley says. “That they enable more fine-grained comment and discussion and attribution; that they can exist in layers for different purposes; and that we could create tools for publishers that are more powerful and more focused on the scholarly paradigm than you get from a more generic system.”
“For us it’s a huge milestone that a publisher like eLife has not only finally put this out and made this available to the users, but has really invested so much of themselves over the last year in helping us get to this point.”
Hypothesis software is open-source and freely available to other publishers, and several have indicated they plan to adopt it. These include MIT Press, the American Geophysical Union, bioRxiv, and the 17 preprint services which are hosted on Open Science Framework Preprints service, not to mention journals on both the Ingenta and PubFactory platforms. (Readers can also annotate pages of journals that have not yet adopted the platform, using the standalone Hypothesis plugin.)
At the moment, Hypothesis users can “listen in” on comments from the public and the hosting journal. A planned upgrade for 2018 will allow users to tune in to other annotation sources as well, Whaley says.
The eLife announcement comes just a day before the NIH announced it was shutting down the commenting function on PubMed Commons due to lack of use. But on 9 February, Hypothesis announced that it had imported those comments, making them available as annotations.
For Maciocci, commenting represents just the beginning of eLife‘s expectations for the annotation service. For one thing, the publisher is now exploring the possibility of using Hypothesis as a platform for peer review as well as commenting (as the AGU has done). “It’s a pretty good bet that, if not Hypothesis, then an implementation that might be based on Hypothesis or something very close to it would be the answer in that sort of scenario,” he says.
If nothing else, Hypothesis provides a forum for authors and readers to interact, he says, for instance to request clarification on protocols or reagents. Indeed, the journal is encouraging new authors to get the ball rolling by making the first annotations.
Time will tell how extensive those conversations will be.
Jeffrey Perkel is Technology Editor, Nature
Correction (12 Feb 2018): This post has been updated to more accurately reflect eLife’s testing of and expectations for Hypothesis. The original version suggested that eLife was actively testing Hypothesis for use in peer review; in fact, the journal is exploring the concept. The original version also stated that eLife hoped to see users adopting Hypothesis to create private journal clubs; in fact, the journal simply hopes to see journal clubs adopt Hypothesis, as it is not currently possible to create private groups with eLife’s built-in Hypothesis client.