How is scientific information distributed?
What is the best way to share, and spread, knowledge?
Page 1 of 2
Following the advent of modern academic publishing at the Royal Society in the 17th century, a profitable industry formed around peer-reviewed academic publishingwhere only paying subscribers could access scientific journalsand the flow of information between scientists ebbed to a trickle. Today, digital publishing is turning this model upside down, and the landscape of science is shifting beneath exponentially accumulating digital data, much of it available for free online. Although the final outcome of these changes is unclear and some believe peer review itself is threatened, one benefit is apparent: Scientists are discovering new ways to collaborate and communicate.
Since 2002, Sweden's Lund University has managed the Directory of Open Access Journals. Promoting standards and enabling access to this new kind of publishing, the Directory has grown steadily each year, but its members, which now number 3,663, are still well below the 16,000 or so active journals working on the traditional pay-for-content model.
How has science publishing changed during your career?
I marvel at the difference between now and 10 years ago. We now have ready access to an increasing breadth of content. The great frustration is that it's also hit or missour wealth of information has increased our expectations, but the sophistication of our tools has yet to match our aspirations.
What do you mean?
We're mostly just using the electronic infrastructure as a new way to distribute scholarly communications; the underlying document formats haven't sufficiently evolved to exploit the new opportunities. For instance, most documents aren't automatically queryableit's difficult to find their origins, histories, and dependencies. Marvin Minsky once made a comment about future libraries: "Can you imagine they used to have libraries where books didn't talk to each other?" Soon, we may be asking, "Can you imagine they used to have an Internet where authors, databases, articles, and readers didn't talk to each other?"
How do we get there?
Most of it is already here but unrecognized. Policy debates over open access and whether federally funded researchers should be required to make their findings freely available are just distractions from the extraordinary benefits to be gained by having the freedom not only to read but also to mine and manipulate information.
Is open access publishing compatible with the peer review process?
If we somehow had had the Internet before the printing press, someone would have still invented peer review. Just the notion that an article has been read and certified is extraordinarily usefulthe arXiv is parasitic on this credentialing process. Peer review defends scientific communication against the introduction of nonscientific ideas, but it needs to be a more bottom-up approach to quality control.
What will the arXiv be like in ten years?
When you read an article, automated tools will show you not only the things the author is explicitly referring to, but also things the author should have referred to. These will be found on the basis of text similarity, or through automated services harvesting co-citation, co-readership, and other shared properties. Also, if we provide incentives for voluntary participation, we'll have actively curated linkages in addition to passive collaborative filtering inferred from logs of user activities. Together, these filters will give us a system where novel things are quickly noticed, highlighted, and glued into our growing knowledge network. Interviewed by Lee Billings
Page 1 of 2
The Fundamentals: Publishing
Posted November 20, 2008
Originally appeared in Seed 19