NATO's military campaign against Yugoslavia is undergoing subtle shifts.
The dispatch of additional aircraft designed to operate in inclement weather, coupled with the expansion in the use of cruise missiles against government buildings in the Yugoslav capital are likely to inflict terrible damage on President Milosevic's state. NATO planners are also encouraged by the fact that public support for the operation is holding up well in key countries such as Germany, Italy and the US, originally identified as potential problem areas. And, crucially, a new strategy is emerging. It is designed to introduce ground troops but not in a warfighting capacity, to protect the Albanians, but leave the question of Kosovo's final disposition wide open.
It is already clear that, in planning for the air campaign, NATO commanders made a number of key errors which the alliance is now seeking to correct. The first, and most important, was in the timing of the operation. Western intelligence sources were in possession of reliable information as early as September last year that Mr Milosevic was planning for a massive offensive against the Albanians of Kosovo. He knew that a guerrilla campaign, once started against his regime, is almost impossible to defeat. He therefore planned to nip the Kosovo Liberation Army in the bud by destroying its support base among the local population, at almost any human cost. A war plan involving large shifts of population was therefore hatched.
Interestingly, however, the plan was opposed by the then chief of the Yugoslav forces, Gen Momcilo Perisic, who managed to drop oblique hints on this issue to NATO commanders. Air strikes at that time could have had a deterrence effect and would have come at a time when the Yugoslav forces were particularly exposed.
But President Clinton's administration was mired in impeachment hearings in Congress, while a newly-elected German coalition government still had to be persuaded about the very need for a NATO operation. The result was a deal which the Americans negotiated with Mr Milosevic allowing for the introduction of Western "observers" on the ground.
NATO touted this deal as a great triumph; in fact, it simply allowed Mr Milosevic to gain time. Gen Perisic was dismissed, and large Yugoslav forces were poured into the province.
By February this year, alliance commanders knew that air strikes, on their own, were unlikely to deter a humanitarian disaster. Nevertheless, NATO was stuck with the policy of air strikes essentially because the US administration was persuaded that its Congress would not tolerate any other alternative. Planners in Brussels hoped that Mr Milosevic would blink at the last moment. They subsequently assumed that, even if air strikes had to be launched, their impact would be swift, and he would sue for peace. The idea that the leader of a relatively small European country could call the bluff of the mightiest alliance in the world and survive to tell the tale two weeks thereafter surprised many in Brussels.
When the humanitarian disaster duly came, alliance leaders were lost in a labyrinth of contradictory positions. They could not claim to know nothing of the possibility that ethnic Albanians would be expelled: preventing their expulsion was actually NATO's original justification for the operation. Yet Western politicians could not explain how they thought that air strikes alone could achieve a humanitarian purpose, or why they failed to plan the necessary civil infrastructure to aid the anticipated waves of refugees.
In essence, NATO asserted a need to act urgently in order to avert a disaster for which it was patently unprepared. Governments ultimately settled for an intermediary position which sought to square the circle. They admitted previous knowledge of Mr Milosevic's murderous plans, but claimed to be "surprised" by the Yugoslav leader's "ferocity". It was an uncomfortable position which is now changing slowly. The changes are subtle, but they will soon become evident.
A ground operation against Yugoslavia is still ruled out, at least officially. Military planners calculate - correctly - that there is no point in threatening a ground offensive as long as the necessary troops are not stationed in the area. But preparations for this offensive are now continuing, under the guise of protecting refugees in neighbouring Macedonia and Albania.
Meanwhile, the Alliance is also refining its war aims. Publicly, NATO is still committed to upholding the Rambouillet peace treaty devised during negotiations in France earlier this February. In practice, however, everyone knows that this treaty is dead.
Even if Mr Milosevic returned to the negotiating table today, the Albanians will no longer accept mere autonomy within Yugoslavia, as they were originally promised. Yet NATO has to be careful about committing itself to an independent Kosovo partly because this could inflame public opinion in Macedonia against the alliance, and partly because Brussels is still not sure that the Americans are prepared to commit troops to a major offensive. The result is a new NATO position which, yet again, seeks to square various Balkan circles.
The dispersal of refugees throughout Europe is designed to reassure the Macedonians, who have their own Albanian minority. Meanwhile, the troop build-up in the region is aimed at frightening Mr Milosevic.
And, finally, NATO is now committed to a policy which demands the withdrawal of all Yugoslav troops from Kosovo, but gingerly avoids the thorny issue of what would happen to the province thereafter.
Governments know that Mr Milosevic will never consent to a formal withdrawal of his troops. But they hope that, under the weight of air strikes, he would be forced to consolidate his hold over the northern parts of the province where his supply lines are shorter, and abandon southern Kosovo. NATO troops would therefore be able to enter the region virtually unopposed.
There will be no peace deal and probably no ceasefire either. But NATO could claim to have introduced troops on the ground in order defend the Albanians, while Mr Milosevic could claim victory in resisting the West.
Jonathan Eyal is Director of Studies at the Royal United Services Institute in London