As an Ars commenter observed <https://arstechnica.com/space/2024/08/nasa-acknowledges-it-c...>, it is possible that the real issue isn't whether Starliner is safe to return with humans. If that were the question two months of debate are, as you said, by itself enough to say "no". Return Wilmore and Williams on Crew Dragon. Done.
The commenter posited that the real issue is that NASA does not trust Boeing's software to undock Starliner autonomously. We know that Wilmore had to take manual control on the way up because of the thruster issues. NASA may fear that if thrusters fail again, Starliner software may again not be able to handle them, and the spacecraft might ram ISS. Thus, the agency wants a human to be able to take over if necessary. *That* is the dilemma. This is something that I and others had mentioned over the past couple of weeks, but the Ars commenter is I think the first outside NASA to put it so starkly.
[1] https://simpleflying.com/boeing-starliner-undocking-software...
The SLS[1] stans (as Berger described on Twitter) are now focusing on how parameter change != software update thus Berger was wrong all along. It'd be one thing if said changes took a day or two to do. But assuming that it is the four weeks Berger reported, that absolutely means that it is the same thing as "software needs replacing"/"functionality was removed". To the client, NASA in this case, it doesn't matter whether the weeks to implement a new feature is (one week of uploading parameters and three weeks to validate said parameters), or (3.5 weeks of uploading new software and 0.5 weeks of validating said new software). The end result is needing four weeks.
[1] Space Launch System, intended to return the US to the moon. Cost: $24 billion and rising fast
[1] https://futurism.com/stranded-astronauts-spacex-boeing-space...
If I were a decision maker at NASA and I simply "trust" Boeing at its job and something goes wrong, I would likely end up fired at best.
"Essentially, what we're asking the team is to go back two years in time and resurrect the software parameters that are required to give automatic responses to breakouts near the ISS should we have a problem in close to ISS, which the software now allows them to do manually," Stich said. "The team is always updating these mission data loads as different things change."
- https://arstechnica.com/space/2024/08/nasa-official-acknowle...
If you don't have crew onboard, you don't have any choice but to rely on the software's judgement.
I think the big problem is that the changeover is going to take several weeks, since Starliner should be able to fallback to flying autonomously if the crew happens to become incapacitated for some reason. The changeover should've just been a matter of sending a command from the ground or via a relay satellite.
In parallel, my pet theory is that NASA has probably already made up its mind (astronauts are not going to return on Starliner), and have been dragging this out mostly to make it look like they aren't just going to throw their contractors under the bus (even if they deserve it). Boeing has declared cold feet over fixed cost contracts (in general, not just with NASA), and I think NASA wants to keep the rest of the contractor pool at least at ease that, okay maybe NASA might start being stingier with the money and contracts, but they aren't going just throw you under the bus when issues appear.
They have an idea. "Tests of a similar control jet on the ground suggested a Teflon seal in an internal valve could swell at higher temperatures, restricting the flow of propellant to the thruster." That is the idea they have and they are checking if it matches with the measured signals / explains the observations.
But yeah, about the only thing more irritating than having a problem occur for reasons you don't understand is having it disappear for reasons you also don't understand.
I'd been a space fan for long enough that I know "improper O/F ratio" usually goes under a picture of a fireball or an artist's impression file image. There were multiple spacecrafts that at least blew off engines that way. The JAXA SLIM mission just this January had it that way. The fact that NASA/Boeing Starliner team keep triggering it and getting away with it is probably technically magnificent.
With that prior knowledge, one way I can interpret those corpospeak is "we aren't sure if we can continue to do that and not finally kill the engines, or worse yet, turn everything into a bomb with people inside or around". Many are instead receiving "we technically know such and such [unintelligible] but we aren't sure of anything and we have no idea". That's a great demonstration of public relations skills.
That's what's making the risk difficult to quantify.
The longer NASA pushes out a decision on what to do with Starliner, the more it becomes likely that people within the Biden administration don't want to go with the obvious choice of bringing home the crew on a Crew Dragon, because they don't want the resulting headlines shouting "Elon Musk rescues astronauts from space".
That’s beside the point, though. NASA is an agency of the United States Government. The idea that politics do not come into play is naive at best.
1. Return uncrewed and burn up. Nasa complemented for its caution.
2. Return uncrewed successfully. Nasa complemented for "excess of caution".
3. Return crewed successfully. Major concern remains over the craft. People continue to question the decision making for years.
4. Craft kills the crew.
Either way you are not getting a certified spacecraft out of this.
Of course there is a further possibility that the departing craft will cause risk or damage to the iss.
I mean I'm sure that's still happening to some degree but this process of modeling sounds a lot more formal.
They have weekly status update conferences, and just cancelling those conferences might be more of a PR risk than just keeping the conference and announcing that they're delaying making a decision.
It's like if I have a service outage, maybe I might spend a few minutes to collect debugging information, but my priority would be to bring the service back up via rollbacks or whatever to restore a previously known good state. Currently they are debugging Starliner with people stranded, but maybe they should prioritize on getting those people back home first.
Or maybe everyone involved don't consider being stranded for months in space as a bad state.
If the system is so complex that an extra week does yield some major new insights, that’s way too complex to use.