How could they possibly screw up so badly that a SW update would brick the thing? Did they forget the logic to just reboot to last good/factory FW in case of failure?
This feels like the story they're telling isn't the real story because that's just basic, basic shit that gets beaten on all the time during development.
You're used to consumer software updates that have been tested over literally billions of updates. And they're brilliant - they mostly just work well enough that you don't expect your phone or computer to get bricked during one. But there are rare cases where something goes wrong and they do get bricked. For something like the first ever software update in space, I'd expect the chances of something catastrophic happening would be much higher.
I'm used to writing bootloaders and firmware update systems for heavy equipment with impatient users who do terrible, terrible things to their hardware. Remote software updates fail all the time, for all sorts of reasons, but the things you need to do to come back up 'no worse than when you started' has been known for years.
If you make any sort of device and your updates can brick the device, you suck at what you do.
Not for my users. And there are enough of them out there that if there was some incredibly rare problem that only happened in 1/50000 installs, you would be getting a lot of complaints. Seriously, this is something that can be made bulletproof, if one cares to.
Unless your users are hundreds of kilometers in space and at risk of decompression from something going wrong, then no, your stuff is nowhere near as risky.
A docking system has many, many moving parts under significant stress, and not all of them are tested under the assumption that arbitrary software could drive their positions/torques beyond assumed limits. You can’t factory-reset metal fatigue after it has happened. It’s quite reasonable for NASA to say “this OTA update has to be tested and QA’d to the most rigorous standards.”
My theory, given that this claim makes no sense and the article doesn't cite any sources that say anything like this, that this is just bog standard G/O Media blogspam, with bit of misinformation to increase views.
> Three separate, well-placed sources have confirmed to Ars that the current flight software on board Starliner cannot perform an automated undocking from the space station and entry into Earth’s atmosphere.
The reasons why the current software can't perform this operation when it could during ground testing are supposition though, and it seems Jolopnik ran with that a little bit.
are we? I know that word has "shit" in it, and this is a degredation of things, and so we want to use the new word so sound like we're cool and hip to the new lingo the kids are using, but the term, as coined by Cory Doctorow, refers to when a platform ends up in late stage capitalism and goes off extracting value from users at the cost of their experience. Like Facebook. Is that what's going on here, or is it just a more general things are shitty?
> when a platform ends up in late stage capitalism and goes off extracting value from users at the cost of their experience.
That's Boeing's mission statement in general I think, but specifically here I think there's a strong indication that they deployed some software for their stakeholders that they promised would do a thing, then they literally un-deployed it after they got the sign-off.
Doctorow's own definition of enshittification can drift [1][2][3] but I called it that because Boeing sold this system as one thing, but then apprently unilaterally pulled a core feature. In other words they reduced the value of the system _after_ they extracted what they could from it by scoring the contract. It might not reach the exact definition, but it rhymes.
Honestly I thought "enshittification" as something "hip" had already had its time. It just seemed to apply.
[2] https://doctorow.medium.com/the-specific-process-by-which-go... "[T]he ability to change the underlying functions of the business from moment to moment and user to user, allowing for the rapid transfer of value between business customers, end users and shareholders"
[3] https://pluralistic.net/2023/01/21/potemkin-ai/
first, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves. Then, they die.
This feels like the story they're telling isn't the real story because that's just basic, basic shit that gets beaten on all the time during development.