Their vulnerability to submarines' torpedoes or to missiles flying at them from subs, other ships or ground launchers has been worrying strategists for decades.

The ships are getting defences that are ever more elaborate - and costly. But the missile threat has been surging ahead.

The government needs expert advice on whether surface ships have just about had their day, preferably not from naval officers committed professionally and even culturally to building them.

Ukraine's destruction of the Russian cruiser Moskva on April 14 does not prove that the surface-warship game is over. That ship should have fended off what our navy would regard as a weak attack, so something went badly wrong on the Russian side.

Nonetheless, the engagement illustrates the threat to ships - and the great complexity of costly systems they need for survival. We can use it to understand the problem.

The following description of the attack is partly conjectural, and it assumes that Moskva promptly used all its defensive systems. In fact, there's good reason to think not all were used.

The scene opens with two Ukrainian Neptune cruise missiles flying towards Moskva at about 900 km/h - 250 metres per second - and maybe only 2 metres above the water. They're staying below the horizon to give Moskva as little time as possible for defending itself.

The Neptunes appear on the horizon perhaps 29 km, or 116 seconds, from the ship. Knowing it is not far away, they switch on radars in their noses, locate it precisely and adjust course.

After 15 seconds or so, they're picked up by a big radar that's perched on Moskva's tallest mast, for a long view; it alerts the computerised command centre deep in the hull.

A jammer on the ship aims at the Neptunes, trying to blind their radars - like shining a torch in their eyes. But they have ways to avoid the problem.

Meanwhile, Moskva's main air-defence complex, called S-300F, has been ordered into action.

About 20 seconds pass as its radar powers up, turns towards the Neptunes and finds them. Interceptor missiles, meanwhile, are made ready for flight.

With the Neptunes 78 seconds away, the first interceptor roars up from its vertical tube. It shoots towards one Neptune, accelerating to six times the speed of sound to save precious time.

Three more interceptors successively shoot out. By now the Neptunes may have begun twisting and turning, trying to ruin the S-300F's aim.

Nonetheless, each interceptor has a good chance of hitting. With two heading for each Neptune, the chance should be excellent. Yet they all miss.

By the time the command system has realised this, the Neptunes are 58 seconds from impact. The S-300F fires four more interceptors.

They miss, too. With 40 seconds to go, the S-300F may have time to try again before the Neptunes get too close for it. If it does, it fails again.

The ship launches clouds of aluminum foil strips to confuse the Neptunes' radars. Maybe a radio system called a spoofer tries to deceive them about where the ship is.

All this fails.

A secondary, short-range defensive missile system called Osa-M takes over the engagement. It can fire only one interceptor at each Neptune, however, and not with great chance of success.

It fails, tries again and fails again. Time to impact: 15 seconds.

Moskva is down to its last-ditch defences. Four 30mm cannons fire torrents of bullets. They have a high chance of hitting - but fail.

Impact.

Naval experts would expect a warship as big as Moskva to cope with blasts inflicted by two missiles of the size of Neptunes. But warship damage and survival is a chancy business, and in this case something goes badly wrong; the ship burns and eventually sinks.

We don't know how many people died.

MORE AGE OF THE DRAGON:

Australia and other countries have warships, usually called destroyers, that have the same defensive layers as Moskva. But their systems, especially their long-range S-300F equivalent, called Aegis, mostly work faster, improving chances of survival.

Combat ships without the long-range layer are usually called frigates, though nine "frigates" that we're planning will have it, obviously because our navy thinks they wouldn't be safe otherwise.

The threats that our ships may face are vastly worse than what Moskva (probably incompetently) failed to resist. China and other countries have anti-ship missiles several times faster than Neptunes, greatly compressing that nail-biting period in which the ship fights for survival.

Faster missiles are harder to shoot down, too.

Some new missiles are not fast but stealthy. They're not detected until they're close, and they're also harder to hit.

The attacking force might be not two missiles but 20, or even more. These things cost only about 1/1000 as much as the ships they try to destroy.

Probably two or more interceptors will be fired to shoot down each incoming missile - so the defence may cost more than the attack even if the ship survives.

And enough interceptors have to be aboard the ships to deal with repeated attacks. Our three destroyers don't carry enough, and the frigates will have even fewer.

The great hope for navies is shooting down missiles with lasers. But the technology is not ready and may never be adequate.

All this means that sooner or later we must start thinking about whether and when to stop buying surface warships. Let's start thinking now.