Funding of the arXiv preprint server must now be shared by more of its users, says Nature Physics in its March Editorial (6, 147; 2010) From the Editorial:

The arXiv preprint server has become central to the working lives of most physicists: ‘checking the arXiv’ is the starting point of many a daily routine. Founded by Paul Ginsparg, the arXiv has expanded to include not only physics — astrophysics, condensed matter, and high-energy physics being heavily represented — but also mathematics, computer science, quantitative biology and even quantitative finance. The arXiv now hosts nearly 600,000 preprints from 101,000 registered submitters in 200 countries, and serves more than 2.5 million article downloads every month to 400,000 users.

The statistics are remarkable. And it is also remarkable that, having initially been hosted at the Los Alamos National Laboratory, the server has in recent years been funded and managed by a single institute, Cornell University. Now the operating costs of the ever-growing arXiv match the entire Cornell library budget for physics and astronomy, and the university has made a plea for help in funding it.

As a short-term solution, Cornell is approaching the top 200 user-institutions of the arXiv — who account for 75% of institutional downloads — for a financial contribution to the maintenance of the arXiv. It is heartening that help has already been promised from several large universities and laboratories, but wider support is still being sought. For the longer term, Cornell will assess, with those who come forward to support the arXiv, what the options are in developing a sustainable model for the future.

Secure, ongoing funding of the preprint server is vital, and surely deserves at least national endorsement from US funding bodies, if not some international arrangement. The fast communication of results — data or theory — between experts that the arXiv facilitates is a boon to physics, and one well recognized by Nature Publishing Group: any submission to Nature Physics or its sister journals may be posted, in that original submitted form, on the preprint server (although we do ask that the final, revised and accepted version is not posted until six months after publication in the journal; the published version, in the Nature Physics layout, may not be posted).

It’s up to physicists now to sustain their arXiv, encouraging their institutions’ libraries in particular to engage in its development. More information is available; and comments and suggestions may be sent by e-mail.

Nature journal policies on posting material before and after publication.

Roy Welch and Laura Welch of Syracuse University, New York, examine why researchers seem reluctant to be more directly involved in the annotation of microbial genomes in the February issue of Nature Reviews Microbiology (7, 90; 2009). They write:

“To annotate an organism’s genome, biological information about the organism must be matched to the genes and genetic elements in the sequenced genome. The process is iterative and open-ended: new information is constantly incorporated into the annotation. It can also be recursive: analysis of the annotation may provide insight about the organism that in turn leads to changes to the annotation. Unfortunately, the generation of new information and annotation of the genome are at present completely separate processes. Often new information does not become incorporated into the annotation in a timely manner, a costly loss for those who rely on it to advance their research.

The community of expert researchers who study an organism produce most of the information that becomes part of the annotation and are also the primary group of end-users. It is therefore curious that the annotation process is circuitous and inefficient: researchers communicate new information not as direct updates to the annotation, but as research papers that must later be interpreted and incorporated into the annotation separately — most often by a third party! Indeed, some information never finds its way into the annotation. It would be far more efficient for the research community to contribute directly to genome annotation. Yet the life science community as a whole remains stuck in the old, inefficient paradigm.”

The authors go on to argue that technology is not the impediment, given the wide availability of wikis (collaborative editing websites) and the databases that have been created using these technologies, including EcoliWiki, GONUTS, Myxopedia and Wikipathways. Rather, state the authors, the impediment seems to be sociological: until contributions to a genome-annotation collaborative information repository can be credited by inclusion in a PhD thesis, curriculum vitae, tenure application or grant proposal, direct collaborative annotations are unlikely to fulfil their promise and potential to accelerate scientific achievement.