It is much more likely that SpaceX does extensive end to end tests since they write all software in house and are much more vertically integrated compared to Boeing.
Two things
- I think its important to align ‘ETE’ testing. There’s no way SpaceX is doing extensive ETE software testing at the launch processing facility because there’s minimal value in doing so. Since all of their other philosophies are forward leaning and generally based in first principals, its pretty clear their V&V approach is also on-brand. SpaceX will no doubt perform extensive software testing in the lab (as everyone does, apparently some more than others…); at the assembled launcher level testing will be designed to exercise representative touch points, not exhaustive functionality.
- Other than the fact that SpaceX builds both the launcher and the crew capsule, the “much more vertically integrated” aspect is pretty non-applicable in context. Boeing is definitely writing all their code in house for Starliner; ULA—practically a wing of Boeing—has the launcher side.
And finally, they completely lost telemetry for quite a period with the rocket. Isn't that also bad engineering?
First order, no. That’s the whole point of software running a rocket and not some dude in a driver’s seat.
There’s plenty of situations where there’s no communications with a rocket (or satellite). For instance, one of my launches a few years ago the flight director was literally reading a timed script that said “satellite 1 should have just deployed”, “satellite 2 should have just deployed” etc, because we didn’t have tlm coverage in the region where the satellites were scheduled to deploy. We actually used successful satellite acquisition as each satellite flew over a ground station some minutes after scheduled separation to confirm successful deployment from the launcher. We could have put a boat out in the middle of the pacific to get tlm during the sep events but there was minimal value in spending a couple hundred grand on something that, at best, would have confirmed that there was in fact a launcher failure and the whole pile headed into the drink. Without tlm the best we could do is speculate that had happened, had we not received tlm from the deployed satellites.
That said it is certainly fair to suggest that, in context, a crewed mission have higher standards than a satellite launch (and thus demands more communication coverage), but a transient launch event still demands pre-determined anomaly identification and, in most cases, recovery. Even then, had the [fairly easy to check] clock offset been identified, the mission outcome would probably have been the same.