Breakout 1: What is a scientific paper? (Science Online London 2009)

Lee-Ann Coleman, Katharine Barnes, Enrico Balli, Theo Bloom

Is the traditional paper format derived from the printed paper still appropriate today? How can new kinds of content such as audio, video, 3D structures, etc. be integrated into a research paper? Can a scientific paper contain just datasets or descriptions of a method? And how does free access to a paper change the way we use the information contained in a paper?

From Katharine Barnes (Nature Protocols):

Is the traditional paper format derived from the printed paper still appropriate today? How can new kinds of content such as audio, video, 3D structures, etc. be integrated into a research paper? Can a scientific paper contain just datasets or descriptions of a method? And how does free access to a paper change the way we use the information contained in a paper?

What is a scientific paper at Nature Protocols? Peer-reviewed and edited articles. Network Protocols are not papers, but are more like blogs in that they’re published and comments are invited. They believe movies about protocols are very important. Specifically, JOVE is a good resource.

Other innovations at NPG include: nature precedings, nature chemistry, and thinking of improvements to articles. For instance, the presentation of the article could be improved, stats about the article could be visible, etc. They maintain a traditional view of the basic unit of what constitutes a published paper (peer reviewed and edited). However, they are keen to enhance the basic paper as much as possible with additional material. Her question to us is how far can we go from the model of the traditional paper?

From Theo Bloom (PLoS Biology):

Science publishing has come a long way in the past 50 years, but the current definition of “paper” no longer really works in her opinion. She thinks it’s time for a radical overhaul. For instance, in the “Finishing the euchromatic sequence of the human chromosome”, Nature simply couldn’t fit all the authors on the first page. Also, sometimes the results are just too big for the traditional paper format, or indeed for any journal to host. Perhaps central databases could provide a snapshot of the data at the point in time associated with publishing of an article.

There’s also much that can be done for visualization of figures in papers and display of specialized media types. How best to match the data to the experiment? You need to datestamp and store results appropriately. In an era of machine-readable factoids, how and where does the author express a view? A Crick and Watson style 1-page view of the data? The time is ripe to integrate references with databases for real-time analysis. Specifically, look at what Shotton et al. did for a paper originally published in PLoS Tropical Diseases.

From Enrico Balli:

SISSA started publishing scientific content in 1991, and since then have had to consider reprints, archiving and more. In 1997 they started publishing journals, which was not the original intent (if I am interpreting the talk correctly). What they published was online-only journals, and since it is not physical, the type of content can be very different from what otherwise could be published.

They have normal proceedings, but they also publish posters. Is a poster a paper? Can it be reviewed? They also publish lectures (video and/or slides). They publish some in collaboration with the British Institute of Physics. Manuals are also published by them, which are not “papers”. How can you really review and edit manuals for software? Another instance of the author problem happened with the ATLAS research at the CERN Large Hadron Collider, where there were more than 5000 people working on the project. What does it mean to have such author lists?

They are thinking about a new project due out next year called “The Journal of Stuff”. All the details haven’t been worked out yet. It is a kind of “un-Journal”.


Creating enhanced content is more costly than just typing. So, who should be covering the burden of those costs? Should you attempt to preserve that information long-term? How do you future proof it? The people who fund the research should fund the dissemination of that research: this is the standard method for open-access journals (by TB). For some reason, some researchers find it very difficult to find the original data, which isn’t how it should be (by KB). Publishers are not obliged to provide digital libraries their information.

Audience: It can be very hard to organize the data in a way that others can understand, e.g. when there are petabytes of data. It also seems like what constitutes a paper in the biological field (as opposed to physics) is much more constrained.

Peter Murray-Rust: It’s good, but what’s happening is not nearly enough. He feels the scientific paper is appalling at communicating in the modern world: dense text, for example, is not very communicative. Universities cannot afford to innovate because they have to publish with conventional, high-impact journals. This means that publishers are actually holding back innovation.

Phil Lord: The idea that publishing the data in a way that is plausibly useful is cheap to do is wrong: it can be quite expensive. Of course, it’s completely worth it to publish the data to help both others and you in a couple of years. There’s this bifurcation of data: on the one hand a relatively content-free paper for the RAEs, and on the other hand the data in a database. We don’t really know how to link these two things together. At the moment, we still have to publish, as the RAEs and similar demand it.

(Didn’t catch the name): Agreed with Phil: the description of the research is a very important item to be archived. What really needs to be mentioned is corrections – corrections need to be logged and tracked. What we really need are clear, agreed, annotated databases for all these papers.

TB: Giving credit back to the source of information, e.g. the original paper describing a new knockout mouse, isn’t always done and should be. This often happens with journals that do not allow unlimited references.

Cameron Neylon: The problem isn’t with the paper (publishing a summary or discussion of some research), but with the journal. Filtering and peer review are useful, but the use of the “legacy” paper format is not useful.

Question: Will open source and open access come together in the publishing world? With PLoS, their publishing platform is open source, and they try to use open-source software where possible. Any software it published has to be open source (Answered by TB).

Question: What about redefining papers as open-source software? This would be a situation where the paper is constantly changing and undergoing version changes, as software is. TB: The versions MUST be date stamped.

FriendFeed Discussion

Please note that this post is merely my notes on the presentation. They are not guaranteed to be correct, and unless explicitly stated are not my opinions. They do not reflect the opinions of my employers. Any errors you can happily assume to be mine and no-one else’s. I’m happy to correct any errors you may spot – just let me know!

2 thoughts on “Breakout 1: What is a scientific paper? (Science Online London 2009)”

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s