NASA got men to the moon by blowing a snotload of stuff up (see e.g. the combustion instability problems of the F-1 engine, which were solved by trial and error over a large number of iterations). The difference is that they did it all on test stands, before committing to a first flight.
SpaceX decided to go from test stands to flight testing early on, because the hard part of Starship (returning the two stages to land) cannot be tested well on a test stand. It's an approach that has worked for them before (developing the first stage landing and recovery for Falcon 9).
This approach has also been used by others: the USSR did it for the Proton and N-1, needing 14 Proton launches before the rocket was ready for operational use.

Definitely. Scott Manley did a comprehensive video on the engineering issues.
Oceangate made the hull by laying up a series of CF layers about 2 cm thick. Their process was bad enough that at this thickness they'd have a bunch of waves in the layers, adding extra thickness. Then they'd machine the cylinder to flatten out those waves, then they'd stack up more CF, creating new waves. The photos of the wreckage show the hull has delaminated on the planes where they machined the hull.

I recently read An Atomic Empire, by C.N. Hill. This is a history of the British nuclear power program.

They invested massive amounts of money in both initial R&D and in a series of commercial reactors. The path they chose was a bit unusual (the Magnox reactors were gas-cooled), but they ended up with standardized designs that were built in decent numbers and provided up to 26% of the nation's power needs.

That changed in the 1980s, when the power generation industry was privatized. This scuttled investment in the next generation of reactors, and only one nuclear powerplant was built since.

This is a pattern in post-WW2 UK: build an early lead, then squander any advantage they had through lack of follow-through.

Lead. Lead as in the opposite of 'follow' is written as 'lead' everywhere. 'lede' was used to avoid confusion with lead (Pb), the stuff newspaper type was made of. Since nobody has used Pb to cast newspaper type for decades, this use is archaic and unneccessary, making it pretentious hipsterism and not a serious word.

In my experience, a restart WILL occur overnight without notification. My company sets a deadline for restarts. When that's in place, it's going to restart no matter what.

My home computer (MacOS) has been able to do restarts and system upgrades *without* losing any work for years now. Clearly this is a solvable problem.

Summary: " taikonauts (Chinese astronauts)"

The Chinese space agency calls them astronauts, so should we.

Windows 11 still has "fuck you" style restarts today. I have notification switched on, and if I very carefully manage it I can prevent data loss by saving and closing my applications manually. I never get the computer back in the state I had before the restart though.
I've had occasions where the notification popped up, with the "restart now and lose all your work" button in front of a button I was about to click, and guess what happened.

And yes, if a restart does happen because you left your computer running a process overnight and it fell idle at 3 AM, guess what: you lose all unsaved work.

is continuity across restarts. Get rid of the 'fuck you' style restarts that happen now and:
- save all open documents
- save the state of each application
- after the restart, reopen every application and every document

We're programmers. Proofreading is for pussies. Write once, commit, then let Support pick up the pieces.

I think you're misinterpreting that.

" the background image bitmap code" is the code that handled using bitmaps as background images, not code embedded in a bitmap.

"A coalition of labor rights groups, 404 Media, is pushing to mandate Wi-Fi on ships"

That should read:

I meant that the other way round: Driver61's doing it the hard way.
For decades, we've had racecars that produce enough downforce that they could theoretically drive on the ceiling, but nobody's done it IRL. Driver61 is designing a track where this can be done.

Driver61 is working on a project to drive upside down without using a fan, using downforce generated by driving the car at high speed. The main problem was finding a place to do this: road tunnels have too much infrastructure hanging from the ceiling.

No, the problem isn't me. The problem is that a spec is used that is useless on its own and requires calculation (of a type that not everybody is familiar with, and that requires data not everybody has) before it is useful for any purpose at all.