Just as debate is raging in the life sciences over models for enhanced access to the full text of the scientific literature, a budget cut recommended by a US congressional subcommittee on energy and water development, of all things, threatens the most basic of researchers' services: search functions across authors, titles and abstracts. The committee singled out PubScience, developed by the Department of Energy to provide such services across the physical sciences and modelled on the biomedical service PubMed. The committee decried PubScience as an undesirable duplication of activities already carried out by the private sector. If enacted, its recommendation will mark the death knell of PubScience.

PubScience, barely two years old, has yet to establish itself. There is little doubt that political lobbying by large secondary publishers (see page 980) influenced the recommendation, and that this exercise has been a practice run for a subsequent challenge on the more established PubMed. Given PubMed's strong bipartisan support across US politics, a challenge seems unlikely to succeed, and seems rather to be aimed at preventing any further expansion of PubMed.

Proponents of a strong government role in scientific information and in a free-access archive to the entire literature should take note of congressional opposition to their position. Supporters and beneficiaries of a free and competitive market, including this journal, might be tempted to salute the congressional recommendation, which upholds a principle that efficiency and public interests are generally best served by governments outsourcing to a competitive private sector, rather than trying to emulate it.

But PubMed, run by the National Library of Medicine, can be appreciated by anyone in biology as a service that works, not only for researchers but also for the public. Publishers who contribute abstracts to both PubMed and PubScience do so voluntarily, and both services drive traffic to publishers' subscription sites. These services provide no-frills access across the literature that, if left to the private sector, would have been obtained more slowly and at a greater cost to the research enterprise.

Those dismayed by the latest development would be wrong simply to blame the publishers, who have as much right to lobby Congress as high-profile researchers. More to the point, a resistance to government incursions on free enterprise is common ground between Democrats and Republicans. But the real issue is where lines should be drawn, and this requires a subtle judgement of the public interest, which is why a bill being introduced that would support PubScience has a chance of success.