The RS21007 document could have featured in a John le Carré spy novel. The author of its six drily written pages discusses a plan to turn nuclear submarines into cruise-missile carriers. It is a detailed analysis, probing issues such as exactly which firing tubes would need to be modified. In Britain, it is unlikely that such matters would be discussed outside classified government meetings or clandestine get-togethers between le Carré's fictional spies.

Yet the origins of RS21007 are mundane: its authors are a group of uncontroversial public officials working at the US Congressional Research Service (CRS). The level of detail is not unusual for such a report. In the United States, RS21007 is seen as a useful contribution to the debate about the country's military capabilities. Had the document been written in Britain, it might never have seen the light of day. And that's a problem — without such documents it is impossible to properly scrutinize government proposals.

The proposal that matters right now, and which is currently being studied under conditions of information poverty (see page 464), is an important one. Britain's Trident submarines carry the country's nuclear weapons. The end of their design life is around 15 years away. What comes next? The government favours building a new fleet, at a cost of up to £20 billion (US$40 billion), and a parliamentary vote is expected next month.

Submarine builders and the military say that replacing the fleet will ensure that Britain can deter nuclear attacks and avoid the expensive repair costs associated with old vessels. Lining up against them are various weapons experts and think-tanks who argue that the submarines' working lives could be extended by up to 20 years. That would give the government time to better assess whether a system designed for the cold war is really the right defence for a world where rogue nuclear states and terrorists are the biggest threats.

Which side is correct? Unhappily, Nature has to admit ignorance on this point. Just a few pages of the relevant White Paper were devoted to the issue of replacement versus repair. Attempts by think-tanks to prise information from the Ministry of Defence have failed. A document like RS21007 would have gone a long way towards helping assess the options. Yet no such document exists and there are no plans for one to be published before the vote takes place.

The debate would be more robust if Britain took the US approach, of which the CRS is a relatively minor component. More important are the scientific experts who are given security clearance. Congressional committees have access to information, can interrogate officials and pass judgements, in public and in private, on classified programmes. Non-profit groups such as the Natural Resources Defense Council have accumulated considerable knowledge on nuclear matters and frequently engage the federal government in lively and informed public debate. The National Academy of Sciences is also used by the government as a sounding board for security issues, and many reports are published in unclassified form.

In Britain, the gulf between the Ministry of Defence and academia is far wider, partly because of the over-secretive culture of the civil service. But there is little reason why this should be so. If Britain is to properly evaluate the threat it faces, outside experts need to join the debate. Groups such as the Royal Society and the Royal Academy of Engineering would be obvious first points of contact, and both organizations could do more to make the case for their being involved.

The United States has shown that public scrutiny of critical defence expenditures needn't hand its enemies critical secrets. Britain can learn to do the same.