Sixteen years after the end of the cold war, roughly 27,000 bombs and warheads are gathering dust in the hinterlands of the world's established nuclear powers. More than 95% of these weapons are in the United States or Russia, but France, China and Britain all have stockpiles of several hundred devices. Each nation spends several billion dollars a year to house and maintain the weapons and train physicists and engineers with weapons expertise. Yet their political leaders have given little thought as to why this approach makes sense, or where it will lead.

During the cold war, the warheads were supposed to deter either a conventional military attack or a nuclear first strike by the opposing side. Under the cheerfully and accurately named paradigm of mutually assured destruction, or MAD, they were made ready for use in large-scale retaliation, to annihilate cities and nations.

That era is, mercifully, behind us. Now the five main nuclear powers all wish to maintain their ageing stockpiles, while offering confused messages about what they are for, and how many are needed.

Business as usual meets neither military needs nor obligations on disarmament.

The United States' latest nuclear-weapons research programme, called the Reliable Replacement Warhead (RRW), captures the flavour of this era of nuclear befuddlement. As described on page 18, the programme's supporters envisage a future nuclear weapon that will be cheap to build and easy to maintain.

They portray the design as neither old nor new, but rather as an amalgam of previous designs. It will be assured to work, but, the designers envisage, will never be tested to see if it does. Calling the weapon 'reliable' angers critics, who hold that existing warheads will be reliable for many decades. They see the RRW as another jobs programme for the vast US weapons labs, and fear that its arrival will further undermine the Nuclear Non-proliferation Treaty.

The RRW's advocates believe it will allow scientists and engineers to confront some important issues. At the end of the cold war, the United States closed its huge facilities for producing nuclear weapons. At some stage, they say, the country has to decide how to replace its existing weapons.

What's lacking is a political vision to guide such decisions. The administration of George W. Bush has reduced its nuclear-weapons stockpile, as agreed with Russia in 2002. But it has also floated the desirability of nuclear strikes on buried targets. Were it not for its rejection by Congress, the administration would be pursuing new 'bunker-buster' weapons for this purpose. Critics are left wondering whether the RRW is a pathway to a smaller arsenal, or to a new nuclear weapon.

The United States is not alone in sending mixed messages. France is reconfiguring its arsenal to be 'flexible', while the British government seems intent upon replacing its fleet of Trident submarines. Russia, like the United States, is committed to cutting back its arsenal, although President Vladimir Putin has begun to significantly increase funds for its nuclear-weapons laboratories.

This business-as-usual approach meets neither the nuclear powers' military requirements, nor their obligations, under the 1968 non-proliferation treaty, to move towards nuclear disarmament.

It is past time for these governments to re-examine what their nuclear-weapons research and development capability, as well as their weapons stockpiles, are actually for. Once they've figured that out, it may even be possible for them to manage their arsenals in ways marginally saner than those necessitated by the cold war.