STARSHIP — USA, 2023? If they manage to build this ship in passenger configuration, it’ll have around 1000 cubic meters of pressurized space — the size of a big airliner, or the entire International Space Station. They envision a Mars vessel with forty little cabins holding 2.5 passengers each (or possibly more if you want to pack poor Irish families into steerage) and a big chunk of common space. Their mockups show large windows at the front; if I were a passenger I’d feel a bit nervous about having a big weak point like that. They have not mentioned anything like, say, making the cabins have airtight doors in case of a blowout. One important bit they’re planning to include will be a solar storm shelter; if there’s a flare or something, everybody will have to crowd together into a bunker in the middle of the ship until it passes. Musk claimed that the Starship can fly under its own power with no booster, and could even reach orbit with a light enough payload, achieving the pointless but impressive milestone known as “SSTO”. This made more sense once he revealed that the initial version would not have any vacuum-optimized engines, just seven sea-level Raptors... but once they changed their minds back again, the SSTO idea is probably no longer feasible. According to Musk, the big heat shield on the belly may be the most difficult part to make, as he wants it to withstand reentry from interplanetary velocities. That reentry will try to lose as much velocity as possible at as high an altitude as possible: though the shape is streamlined for going up, they want the opposite of that when coming down, so they will do the descent belly-first. The ship has protruding fins at both ends which are not wings or flaps, but air-brakes whose purpose is to maximize drag. Even once it has lost its horizontal speed, they plan to maintain this same kind of approach and have it belly-flop straight down through the lower atmosphere, using the fins at both ends the way skydivers use their limbs, to balance the ship on a level plane. This would be a slower descent than in any normal capsule or spaceplane, though it still might be faster than the terminal velocity of a falling person. Then, in the final seconds, it would light the engines and go vertical to land on its tail. (This is tricky enough in Earth’s thick atmosphere, and they’re planning to try the same thing on Mars, which will be tougher since there’s so much less drag at lower speeds.) Their original idea was to use a conventional ablative heat shield. Once they switched to making the ship’s skin out of steel, they dropped this and said they would just leave the metal exposed and polish it very brightly. This would reflect away most heat (which is radiated by glowing air in the shock front) and they would use liquid cooling to dissipate the rest — the liquid in question being unburned propellant, which means that if you don’t have tons to spare, you can’t come back down. (TMBG prophesied this when they sang “They run out of gas, the plane can never land.”) The fuel would not return to the tank, but boil out through pinholes in the shield surface. By their calculations this wastage would still come out lighter than an ablative shield over carbon fiber. I soon realized that this approach would face tough challenges at edges and corners, and especially at protecting the hinges for the fins. And that even the flat areas would have a tough time keeping the evaporative cooling uniform and consistent, without hot spots. And unlike an ablative shield, any hot spot could rapidly start absorbing more heat in a vicious cycle, because once an object is warm enough to glow, even if just in the infrared, it also absorbs radiation in the same frequencies that it is emitting. It loses its reflectivity. A few clogged pores, or even a region of below average pressure in the coolant, could lead to a warm spot that just keeps heating until it melts through. In a conventional ablative heat shield, a divot or other weak spot can be somewhat sheltered by the good areas around it, but in this approach, any local imperfection in shedding heat just becomes a magnet that draws in even more heat! This inherent fragility substantially undermined my confidence in the success of the BFR/Starship idea, which sounded great and revolutionary back in 2017. But then the SpaceX engineers sobered up, and Elon said they were moving away from the evaporative cooling idea, and now working on a ceramic tile which would protect the steel. The challenge was to give it the right emissivity to minimize absorbtion and maximize dissipation of radiant heat. But since the steel behind the tile is still okay with being heated to a thousand degrees, the tile can be very thin and light. But Musk says they might still use a bit of evaporative cooling in a few critical spots, such as the flap hinges. But if they get the heat shielding solution to a dependable state (perhaps around the same time the booster is ready), the capability to refuel in orbit will make this a step beyond all other capsules and spaceplanes. It would be in every sense a true spaceship — humanity’s first. They started building the factory for the Starship and its superheavy booster in the Port of Los Angeles, which they picked so the rockets could be transported by sea. They already started making some large pieces of carbon fiber cylinder before the building was built there... and then they tore down their preliminary construction, and scrapped at least one piece of expensive tooling which they had purchased to wrap carbon fiber around. They then said the manufacturing of the steel version will be done in Florida and Texas. At their existing Texas facility, they built a “hopper” to test taking off and landing in a fake Starship, shorter and lighter than the real thing but with the full nine meter diameter. They flew this just once (with a rough landing) before immediately moving on to two full sized prototypes. Then, as the first true flyable Starship was taking shape, they decided to reopen the Los Angeles facility. Then they apparently forgot about it again, concentrating all their efforts on expanding their test facility in Boca Chica (near the mouth of the Rio Grande) into a full shipyard.