A principled editorial policy, implemented with rigour, can lose a journal some excellent papers to other journals that are more lax. What is more, journals are in a poor position to wag the dog of community behaviour. Nevertheless, there are occasions when they can have a significant influence, especially if competitiveness can be sidestepped. One such occasion may be approaching in respect of the vexed issue of the availability, on publication, of crystallographic coordinates of biomolecular structures.

Most researchers believe that, once a piece of science has been published, the methods and materials on which it is based as well as the results themselves should be publicly available, immediately. Many structural biologists disagree. For them, it is a common practice, on publication, to deposit crystallographic atomic coordinates in a repository such as the protein database at Brookhaven National Laboratory, but to keep them inaccessible for a year.

In this context, Nature's policy is to publish when (and only when) a database accession number is supplied, but to accept the principle of the one-year embargo. But that acceptance is increasingly reluctant. As the use of structural data becomes ever greater across biology, that practice is ever more frustrating to the progress of science, not to say offensive to the broader community.

One obstacle to change is the claim by some structural biologists (see Nature Structural Biology 5, 83; 1998) that researchers are owed the breathing space of a one-year hold to exploit results that have been obtained by dint of investment of many months of work, with the substantial risk that the analysis will prove unsuccessful. Researchers in other fields may well view such a claim as special pleading. Another obstacle is the knowledge that such data may not be patentable (although that possibility is being investigated by lawyers), and also that industry, with major quantities of information up its sleeve that non-industrial researchers can make good use of, will refuse to collaborate if immediate openness is required.

The extent to which all of these considerations should determine policy is highly debatable. Nevertheless, they clearly win the argument within the International Union of Crystallographers, whose continuing support for the one-year hold has considerable influence within the structural biological community.

Nature is a service provider. If the journal singlehandedly insists on immediate public access, and enforces that as rigorously as it does, for example, for geneticists, it could lose some excellent papers — or, to put it less selfishly, could cease to be considered as an appropriate service provider. But whose interests are we serving? It seems that increasingly, in accepting the one-year hold, we are serving industry at the expense of science.

We believe that the best way forward is to consult but also to minimize the use of such policy as a competitive weapon between journals. Accordingly, Nature and Science have agreed to collaborate in consulting the community and, potentially, reaching a simultaneous and mutually consistent decision on any change of policy. (Other journals are also being consulted.) Both journals are encouraging readers to express their opinions via a survey being conducted via the website of Nature Structural Biology (http://structbio.nature.com). Either by that means or by correspondence (corres@nature.com), we would welcome readers' opinions.