[armada651]

> if they would get lucky and none of the issues would manifest, they could assume that design is perfect and do not try to improve it. so it's good if whatever can go wrong goes wrong at test flight

I agree that the success should be measured in lessons learned from the launch. However we learned next to nothing about the second stage because the first stage failed for reasons that weren't unexpected.

So from that perspective the launch could be considered a failure, because they traded getting data on the second stage for learning a lesson that was already known.

[Szpadel]

assuming that this failure was caused by failed engines, I think this is a very important lesson, engines can fail during real missing and if their algorithms couldn't account correctly that, this is a very important thing to improve.

and remember, as you already mentioned, that stage 2 won't even start until stage 1 succeeds, so they need to perfect that first.

but I see your point, if they would manage to separate before stage 1 failure they could gather more data

[AstralStorm]

Engines are not supposed to fail ever. Super Heavy does not have sufficient margin to handle multiple failures (maybe one, maybe), and Starship has the same engines, where one failing means it is liable to just fall from the sky.

Unless on the Moon, but then it would have to LEO or lunar orbit abort if it does not burn up on the spot and require a recovery mission.