In recent years, players from many corners of the publishing and research fields have been asking what can be done to make journal articles more affordable and accessible. For Arfon Smith, a co-founder of the citizen science site Zooniverse and now resident “science guy” at Github, the question he and a few collaborators wanted to answer was more basic—what makes a journal article, and how can that process be simplified? As an answer, or at least the beginning of one, Smith and his colleagues built The Open Journal, a software prototype built on top of the arXiv, an open physics research repository, that lets researchers and peer reviewers discuss edits inside a browser-based system and publish the results. The result eschews publishers—and copyediting, typesetting, and marketing—altogether, letting researchers bring their work directly to readers with a minimum of fuss. Smith spoke with LJ about the issues The Open Journal can address for researchers and how the model could develop in the future. What was the inspiration behind The Open Journal? One of the things that came up in conversations with my colleague Chris Lintott is there’s a conflation of a whole lot of a things that happen in a journal. There’s peer review, which is a crucial activity, but then there’s the other things, from editorial boards to typesetting. A lot of those things journals do today are reflections of the process of book publishing, but the idea of publishing today is changing. The act of getting something on the web is now free or very low cost, something that is demonstrated well by arXiv, where you can get your paper up and your peers can get a look at it. What’s missing from arXiv being a proper journal is peer review. We wanted to ask ‘What does it look like to have a lightweight service that just facilitates review, that doesn’t try to manage the editorial process, but recognizes that content lives wherever it needs to live, and there’s value in a service that just ingests content and presents it to colleagues?’ Would a journal program like this need more fields to adopt open data repositories to be really useful? In physics, journals generally don’t mind authors putting up preprint versions prior to publication. That’s not standard across all fields, though we are seeing others come closer. For example, you now have BiorXiv, a preprint server for biology research. The idea of preprint servers, though, is good for everyone. Except maybe people who want to charge for content. You have a background on citizen science projects like Zooniverse. Do you see a connection between that work and The Open Journal? Zooniverse is a way to ask what scientific collaboration will look like going forward. Because you had such a mix of people involved, it was a really interesting experiment in what academic work might look like in the future. It also raised some access issues, as there are a million people who have contributed time to Zooniverse work, and a vanishingly small number of those people have regular access to scientific journals. That’s a problem, and it’s where I think there’s a connection:  the concern over access to information. Making peer review easy and attractive seems pretty key to The Open Journal . Would readers be able to look at that back and forth and how the paper changed between its first trip to the repository and final publication? If you got to the arXiv you can download all revisions for any paper there. As for readers seeing peer review notes, maybe. That’s a conversation an editor would want to have with an author. Right now the goal of the review interface is to make the process as web-native as possible, and to improve over the current typical workflow of reading a pdf, emailing comments, rereading the updated PDF… that is not great. Our goal was to make the revisions happen in the browser. Making that process open to the public is not necessarily what we set out to achieve, but it also would not be difficult to achieve on this platform. Not to sell your work short, but the demo didn’t make this seem like it was all that much effort on the part of you and your colleagues. Is this a model that could be at least tested pretty easily? One of the big rules of software development is that the first 80 percent is easy and the last 20 percent is really hard. We have a functional prototype, and we’re trying to get some papers in the arXiv through review this summer as a testing ground. Ultimately, it’s not clear what comes of that, though. My goal is not to run a journal. We’d like to get other people involved. Have you talked with other organizations about partnering to get this model off the ground? That would be my dream scenario, that some kind of value comes out of this that lets us collaborate on taking it forward with people whose job it is to do this stuff. And hopefully that’s something that’s not just applicable to journals. The model of reviewing an entity that’s posted on the web and letting people help you refine it and give it status is something that applies more widely than to the arXiv. It could be interesting to ask “What does this look like on Github?’