Viewpoint: No easy endgame in Libya
In early January 1943, Field Marshal Montgomery decided "there was only one thing to do - crash on to Tripoli". Drawing on three divisions from Benghazi, and attacking from multiple directions, his Eighth Army took the city by the end of the month.
Nearly 70 years on, British forces find themselves once more bombarding Tripoli and its environs in anticipation of a final push. But it is Libyans who have surrounded the capital, poised to crash on to their final prize.
The rebels, having learnt from disorganised advances earlier in the war, are fighting methodically, consolidating their grip over territory before rushing forwards.
Regime forces, stretched across too many fronts, are pinned down in hopeless positions for fear of Nato air strikes during retreat over open road.
In Tripoli, rolling blackouts and fuel shortages will only worsen as supply lines are progressively severed. A pincer movement from the coastal roads is inevitable.
In the meantime, Tripoli residents have pre-empted this by starting their local revolts. Even if these don't tip the balance themselves, they should divert regime forces and so weaken defences elsewhere.
But a march on Tripoli would not be simple.
Col Muammar Gaddafi's elite forces have not disintegrated even in the face of ultimately hopeless odds, and they can be expected to contest each of the 40km (25 miles) between Zawiya and the capital.
Once opposition fighters arrive at the edges of the city, the urban terrain will offer advantages to the defender.
As in Zlitan, the regime will exploit civilian buildings, including mosques and schools, as fixed points of defence for rocket launchers and snipers.
Nato will not stop its raids, but it will be forced to take much greater care in targeting. And support from the local population could be uneven, with armed volunteers slowing the way.
A second problem might arise if Col Gaddafi lashes out in desperation.
The regime possesses between 80 and a few hundred Scud missiles, one of which was fired ineffectually towards Brega last week.
Could these be used during a last stand, to terrorise Nato warships or rebel territory?
Between 1988 and 1992, for instance, Soviet forces fired nearly 2,000 Scuds against the Afghan mujahideen, killing large numbers of civilians.
The risk is manageable. Scuds are highly inaccurate, and their long preparation time of one hour (which stems from their liquid fuel) means that Nato aircraft would probably spot and destroy any large concentrations of missiles moved into the open.
Nor are chemical weapons a serious threat.
Col Gaddafi may have around 10 tonnes of mustard gas, but Western special forces are likely to have monitored or even secured the storage site (south of Sirte) and it would anyway be prohibitively difficult to weaponise these stocks.
More likely is a simpler form of terrorism.
Operations against Western countries would be difficult, given the intensive monitoring of Libyan intelligence officers abroad, and the diversion of the state security apparatus towards the uprising.
But key oil and transport infrastructure within Libya, as well as civilian targets in Tripoli, could be targets of bombings.
As with Saddam Hussein's burning of Kuwaiti oil wells, these would do little but delay the inevitable. They could, however, make a political transition all the more chaotic.
Finally, then, what are the challenges of such a transition?
First, the Transitional National Council (TNC) must physically enter Tripoli at some point.
Its well trained Tripoli Brigade of several hundred fighters has been active in the west, and it enjoys international recognition as the sole legitimate representative of the Libyan people.
But if other rebels seize government buildings in the course of an urban insurrection, the Benghazi-based committee might be sidelined by self-declared competitors.
The TNC has broad regional representation, but many westerners see this as a facade and resent the perceived imbalance in sacrifices between the two parts of the country.
Second, what of the tribal question? This should not be treated in either alarmist or dismissive fashion.
Urban areas and western Libya have much weaker tribal identities than rural areas and the east, and tribes are not autonomous armed groupings.
But it is naive to imagine that long-simmering tribal grievances, some of which have been considerably sharpened by the war, will not prove incredibly divisive.
The Warfalla, Tarhuna, Magarha and Warshafana tribes have all made enemies, but their real or imagined mistreatment by the TNC during a post-conflict period would be highly destabilising.
Lastly, could Islamists hijack the revolution? This is almost certainly an overblown fear.
Islamists are fighting alongside rebel forces - given their history in eastern Libya, and their combat experience from Afghanistan and Iraq, it would be surprising if they were not.
It is absurd, however, to compare them to the battle-hardened and well-funded Islamists of Afghanistan or to imply that they have substantive ties to al-Qaeda.
Disarming these and other militias will be a crucial task for any provisional government, but no more a priority than the re-establishment of economic normality and fuel exports.
These are serious challenges.
Confident rebel assurances of an unproblematic transition should be taken with due scepticism. But it is untrue to suggest that no planning has taken place or that an Islamist coup and opposition infighting is inevitable.
Much depends on the choices taken as the rebel shadow over Tripoli lengthens.
Shashank Joshi is an Associate Fellow at the Royal United Services Institute (RUSI), a defence think tank, and a doctoral student of international relations at Harvard University.