Last Friday night, fellow blogger Seth, some of our friends, and I had a discussion (with what David Foster Wallace called “that special intensity that comes after about the fourth beer”) about some interesting results from the CDF collaboration. After a while, the conversation wandered towards discussing the role of open repositories versus refereed journals. This is an issue of interest for science in general. However, high energy physics is an especially visible example of the contrasts between the two, because of our field’s reliance on a culture of preprints. To quote an interesting study (itself available on the open arXiv repository, established in 1991) on HEP publishing practices:
The preprint culture in HEP pioneered the free distribution of scientific results. For decades, theoretical physicists and scientific collaborations, eager to disseminate their findings in a way faster than the distribution of scholarly publications, printed and mailed hundreds, even thousands, of copies of their manuscripts before submitting them to peer-reviewed journals. This preprint culture tended, however, to favour the large laboratories and universities that could afford mailing large numbers of preprints while receiving comprehensive regular mailings. The spread of the Internet and the inception of the arXiv repository ushered a new era for the preprint culture, offering all scientists a level playing field. In its current implementation, arXiv allows researchers to submit their preprints and browse or receive regular feeds on recent submissions in their area of interest.
Open repositories like the arXiv have some major advantages over traditional refereed journals. The first and most obvious advantage is that access to the materials is free (at least to the end user). Instead of a costly journal subscription (which may run to many thousands per year), anyone with a working internet connection can download and read the results for themselves. Another key advantage is speed. Papers on the arXiv are instantly available around the world, which is important in a fast-moving and competitive field. The same results can take months to appear in refereed journals. And as author lists for large HEP collaborations swell, free electronic distribution avoids the high printing costs and page count of reprinting author lists with each paper1.
However, open repositories have a number of notable drawbacks when compared with traditional journals. There is no review process, so unlike a published journal, there is no panel of referees (experts in the field) reading and checking the results. Not every paper posted on the arXiv makes it into a journal (around sixty percent of the articles eventually make it into print). And even in the case where the paper is accepted and published by a journal, the version left up on the arXiv does not always reflect the final published version (though authors can and do upload revised versions to reflect these changes). Journals are also important because of the value to a scientist of having a published article: publications are one of the primary factors used by universities, funding agencies, and other sources of support to discriminate between candidates for grants or promotions.
The best solution would be one that combined the advantages of the open repository (free availability, speed of communication, and avoidance of paper printing costs and limits) with those of the refereed journals (peer review, authoritative versions of studies, and meaning for employers and funding agencies). Fortunately, such a solution has been proposed for academia in general and for high energy physics in particular. Open access publishing (OAP) is an initiative to make the results of scientific research freely available, using the Internet as a distribution mechanism. Not only does this make papers accessible to the worldwide research community, it also provides access for the billions of taxpayers around the world which subsidize research.
In many ways, the HEP field is a natural candidate to lead the initiative toward OAP. The aforementioned preprint culture and the establishment of the arXiv were both steps toward the open access model. HEP also gave birth to the first two peer-reviewed electronic journals: the Journal of High Energy Physics (1997) and the Physical Review Special Topics Accelerators and Beams (1998). The internationally connected nature of the discipline makes it ideal for rapid and distributed electronic communication. With a relatively small number of scientists (not more than a few tens of thousands at most), and a correspondingly small publishing output (87% of HEP articles appear in just six journals), the community can easily and quickly adapt to an open access model. In fact, the following statement has been adopted by each of the LHC collaborations:
“We, the CMS Collaboration, strongly encourage the usage of electronic publishing methods for CMS publications and support the principles of Open Access Publishing, which includes granting free access of our CMS publications to all. Furthermore, we encourage all CMS members to publish papers in easily accessible journals, following the principles of the Open Access Paradigm.”
where CMS may be replaced by ALICE/ATLAS/LHCb.
Though open access is free in the sense of unlimited access, it is not free in the sense of costs nothing, since the review process, web hosting, editorial services, etc. all take money to maintain even for an open access publication. Since the HEP community already pays a for journal access (multiply tens of thousands for a subscription times the hundreds of institutes in the field), publishing services already cost the field millions per year ($13M by one estimate). The SCOAP3 proposal (Sponsoring Consortium for Open Access Publishing in Particle Physics), which is gaining support from increasing numbers of institutions around the world, is to redirect this subscription money towards sponsoring open access for HEP journals. In addition, the proposal will eliminate author-pays publishing fees that act as a barrier to entry for scientists from countries or institutes without substantial HEP funding. So far nearly 60% of the needed funds have been pledged by members from 20 countries (and CERN). In the meantime, several publishers have offered some or all of their HEP content for free under an open access model in anticipation of SCOAP3 being fully funded. The future of the open-access model, which combines the strengths of the existing open repositories and peer-reviewed journals, seems bright.
Elsewhere on the internets….
- The Gauge Connection points out a strength of another sort of open access resource, and *Not Even Wrong* points out a related weakness.
- A Quantum Diaries Survivor offers an amusing survey. [My answers: A1, B6, C2, D6, E2, F4, G2, H4, I5, J3, K3.]
1 Conventions of authorship in HEP are themselves a rather interesting subject, see for example this report comparing the approach of HEP collaborations with other fields like astrophysics and the biosciences, and suggesting some possible options to adapt HEP authorship to the era of multi-thousand-author collaborations.
