[dredmorbius]

You don't need the aircraft to survive.

You need the passengers to survive.

And that's eminently possible, even where the hull itself is destroyed (several cases of fuselages splitting in two or three with no or minimal fatalities):

<https://news.ycombinator.com/item?id=42606790>

[bagels]

The choice of keeping the craft intact is much more likely likely to lead to a safer outcome, isn't it?

[dredmorbius]

It's a vastly more challenging goal, with higher engineering, financial, and land-use requirements.

Passenger survivial: Decelleration g-forces kept within a given threshold, evacuation slides operable, passengers cleared within 90 seconds. Hull is sacrificed.

Airframe survival: No significant damage to aircraft structure or systems.

Humans in this case are substantially more robust than aircraft.

You'll find a similar situation in, e.g., earthquake safety construction. The goal isn't for structural reuse, but for inhabitant survivability. Structures may be renovated in some case but are generally demolished and replaced. They did their job in saving lives.

Steel-reinforced concrete buildings can still sustain considerable damage, possibly to the point that they will be unusable after the quake. This has to do with the way governments set building codes, which tell engineers how to design a building to withstand a certain level of earthquake shaking. Codes, including those in the U.S. and Turkey, generally require that a building achieves what is called “life safety” under a given maximum expected earthquake in an area. “Our seismic codes are only a minimum requirement,” says Sissy Nikolaou, research earthquake engineer at the National Institute of Standards and Technology. “You just want these buildings at least to give you the chance to get out of it alive when the big one happens, under the assumption that they may be seriously damaged.” The situation is akin to a car that crumples in a crash: the vehicle absorbs the impact to protect passengers, but it is totaled.

<https://www.scientificamerican.com/article/how-to-engineer-b...>

The automobile-safety example given above is also apt. A car can be destroyed in a crash, what's key is that its occupants survive.

[bagels]

When the wings come off, you get a pretty big fire though. Obviously they cannot be made infinitely strong, but I'd rather be in the plane that isn't disintegrating and in fire.

[dredmorbius]

Armouring wings against fuel-system penetration (tanks, hoses, valves, etc.) in a mishap is an absolutely insane level of engineering, and all but certainly infeasible.

The much more reasonable alternative is to have 1) fuel-dumping systems which rid the craft of excess fuel prior to an emergency landing (another reason, BTW, to remain aloft as long as possible after an emergency has been declared), and 2) to have the means to evacuate the aircraft quickly. Modern standard is 90 seconds, with only half the exits in use. Initial fireballs, whilst impressive, usually occur largely above the aircraft and direct heat upwards. It's the subsequent ground- and cabin-based fires which are most lethal, and those tend not to develop for about two minutes.

<https://simpleflying.com/aircraft-90-second-evacuation/>

After that, you want the actual fuselage and passenger restraint systems (seats, seat belts) to provide maximum protection against injury in the crash itself. They largely do this. At least when aircraft aren't encountering Muan Murder Walls at speeds well in excess of 100 kt and disintegrating fully.