Sunday, September 18, 2005

Starting a new bushfire

As the dispute between the The American Chemical Society (ACS) and the National Institutes of Health (NIH) rumbles on it is becoming increasingly apparent that providing open access to publicly-funded data is more complicated, and more controversial, than anyone could have anticipated. Why else, after all, is the Open Access Movement — 10+ years after Stevan Harnad's Subversive Proposal — still struggling to make scientific research papers freely available on the Web?

On the surface the decision by NIH to create PubChem — a freely accessible database containing chemical structures of small organic molecules and information on their biological activities — seemed simple, straightforward, and eminently desirable.

The move, however, immediately pushed NIH into an apparently endless dispute with ACS, which claims that the planned open-access database would threaten ACS' revenue from Chemical Abstracts Service (CAS) — the highly profitable division of the ACS that sells access to its CAS Registry. In a public statement dated May 23rd CAS claimed that PubChem will over time "pose an insurmountable threat to CAS' survival" since it is "a mini-replica of the CAS Registry, and a replica poised to expand."

Since then there has been a continuous, and at times bitter, debate about NIH's plan, with contradictory claims over the extent to which PubChem would or would not duplicate what CAS offers, and questions raised as to whether building such a database was even an appropriate task for NIH to undertake.

One of CAS' early tactics was to argue that it was wrong for NIH to use public money to build PubChem. However, critics were quick to point out that the allegedly threatened CAS Registry database was itself built with taxpayers' money — in the form of a grant from the National Science Foundation (NSF). They also argued that as a non-profit organisation CAS has over the years benefited substantially from the public purse in the form of tax concessions — making its objections to PubChem sound somewhat hypocritical.

Presumably sensing that it was losing the PR war, in August the ACS offered to develop and support a freely available database that would include data from NIH screening centres, as well as other compounds with associated bioassay data. To support the project it pledged $10 million and 15 staff members over five years.

The offer, however, was rejected by NIH Director Elias Zerhouni. He did so for a number of reasons, not least — as he pointed out in a letter to ACS President William Carroll dated August 22nd — because under the requirements of the Federal Acquisitions Regulation, NIH "would not be able to enter into an exclusive bilateral relationship with ACS without such an opportunity being made available to other private sector suppliers of chemical information. "More importantly, the letter stated, if NIH accepted ACS' proposal "some of the most critical aspects of PubChem would be lost."

In the same letter, however, Zerhouni proposed a six-part "alternative structure". Amongst other things this would include collaboration between CAS and NIH to assign registry numbers for PubChem structures, and a promise that PubChem "will not disseminate information on chemical reactions, measured properties, methods, patents and applications, markush structures, or conference information."

In offering the patent concession, however, Zerhouni is in danger of starting a new bushfire, since the proposal is sure to inflame the passions of patent information searchers.

In a posting to the Patent Information Users Group mailing list on September 17th, for instance, self-styled patent buster Greg Aharonian — who also runs the Patnews  mailing list — complained: "If this is right, the public funded NIH that is building a public funded chemical database is agreeing not to include publicly funded information from publicly available patents. This is wrong, and people (led by the PTO) should contact the NIH to lobby that they include as much information as possible from patents into PubChem."

He added: "Over time, PubMed + Pubchem will become an ever more useful tool for PTO examiners to search through [in order] to issue high quality patents. And higher quality patents is a service the government should be providing."

We should feel some pity for Zerhouni: in seeking to get ACS off his back he is in danger of attracting the wrath of one of the more voluble and argumentative information user communities.

Certainly no one would willingly put themselves in the sniper sights of Aharonian. As Wired magazine commented a few years ago, when describing Aharonian's endless battle against "crappy" software patents, "Aharonian isn't just a loose cannon — he's a carpet bomber. Patnews, a sort of Drudge Report for the patent world, targets corporations, patent attorneys, bad patents, and, invariably, the PTO for the failures of the patent system. Patnews' 3,900 subscribers, and the many more who read forwarded copies, view him alternately as a self-serving wonk, a tireless public advocate, and a pain in the ass."

No comments: