My tests keep failing until I fix all of my code, then we deploy to production. If code fails in production than that's a problem.
We could say that rockets are not code. A test run of a Spaceship surely cost much more than a test run of any software on my laptop but tests are still tests. They are very likely to fail and there are things to learn from their failures.
You test components in isolation, you test integration of components, you run simulations of the entire rocket, and finally you test the rocket launch.
You’ll catch issues along the way, but you can’t catch all of them before a full launch test. That’s why there are launch tests.
This has been SpaceX’s methodology for a long time now and has gotten them to the point where they have the most reliable western launch vehicles ever launching record amounts of mass to orbit each year at record low prices.
I truly hope that if you ever design a rocket yourself, that you will test it. I have no idea why you'd think testing is a terrible thing to do if it has to do with rockets.
Even NASA years into their existence has suffered catastrophic fatal failures. Even with the best and most knowledgeable experts working on it we are ultimately still in the infancy of space flight. Just like airlines every incident we try and understand the cause and prevent it from happening again. Lastly what they are doing is incredibly difficult with probably thousands of things that could go wrong. I think they are doing an amazing job and hope one day, even if I miss it, that space flight becomes acceptable to all who wish to go to space.
I you are referring to the two Space Shuttle accidents, both of them could have been avoided with just a little bit more care - not launching in freezing temperatures for Challenger, and making sure insulation foam doesn't fall off the tank for Columbia.
The history of rocketry goes much further back than the space shuttle. The shuttle was supposed to be a step towards reusability but didn’t succeed or progress the way they thought it would. Starship is continuing that dream of full reusability and their approach is working. You can’t plan everything on paper when it comes to hardware especially when attempting things that have never been done before, you just don’t have the data in that case. You have to build prototypes and test them to destruction. All manufacturers do this.
Can you name a space company with less failures? Also I think it is unfair to even compare SpaceX to anything else, because of the insane amount of starts / tests combined unparalleled creativity.
According to this website their current success rate is 99,18%. That's a good number I guess? Considering other companies did not even land their stages for years.
There's more to "overall success" then launch failure rate. Cost and time are very important, which are the other dimensions they are optimizing for here.
It says right there in your source that that figure refers to Falcon in particular. For comparison, Starship's current track record is 3/7 launch failures (+1 landing failure).
There's an order of magnitude difference between them. If they were cars, it'd be like comparing the smallest car you can think of vs one of the biggest tanks ever made.
I ignored those, since the starship at this stage can be considered a prototype. I am just trying to argue, that calling SpaceX unreliable, especially compared to its competitors and time to market, is bold.
The usual definition of success for a rocket is getting the payload to the intended orbit. Since Starship doesn't have a payload yet, at least not a real one, its "success rate" is not measuring the same thing.
I'd say that only the 7th mission was legitimately a failure, because there was some rerouting of flights outside the exclusion zone. The other six missions were successful tests since nothing other than the rocket itself was affected.
You cannot compare a mature product to something that is still under initial development.
That would be like comparing a 1-y.o.'s ability to run to a 10-y.o.'s. Of course the younger kid doesn't yet control their legs, but that doesn't mean it's going to stumble and fall forever.
The taxpayer money is for r&d. We should be very tolerant of failure. Aggressively testing with real hardware is a key part of how we learn to make a more robust systems. Fear of failure and waste will slow down progress.
It sounds like he's talking to investors and not to general public.
In my experience in corporate america you communicate efficiency by proclaiming a checklist of things to do - plausible, but not necessarily accurate things - and then let engineers figure it out.
Nobody cares of the original checklist as long as the problem gets resolved. It's weird but it seems very hard to utter statement "I don't have specific answers but we have very capable engineers, I'm sure they will figure it out". It's always better to say (from the top of your head) "To resolve A, we will do X,Y and Z!". Then when A get's resolved, everyone praises the effort. Then when they query what actually was done it's "well we found out in fact what were amiss were I, J K".
Test flights.
My tests keep failing until I fix all of my code, then we deploy to production. If code fails in production than that's a problem.
We could say that rockets are not code. A test run of a Spaceship surely cost much more than a test run of any software on my laptop but tests are still tests. They are very likely to fail and there are things to learn from their failures.
Running a code test doesn't require firing a rocket.
How would you test a rocket?
You test components in isolation, you test integration of components, you run simulations of the entire rocket, and finally you test the rocket launch.
You’ll catch issues along the way, but you can’t catch all of them before a full launch test. That’s why there are launch tests.
5 replies →
Test code by running it.
Test a rocket by launching it.
9 replies →
Thank god you're not building rockets.
Testing to failure is pretty common in rocketry. If you don’t push the limits you’ll never really know where the limits are.
This has been SpaceX’s methodology for a long time now and has gotten them to the point where they have the most reliable western launch vehicles ever launching record amounts of mass to orbit each year at record low prices.
I truly hope that if you ever design a rocket yourself, that you will test it. I have no idea why you'd think testing is a terrible thing to do if it has to do with rockets.
2 replies →
Why precisely? Can you elaborate?
He just means MORE checking for leaks.
They already implemented a whole host of changes to the vehicles after the first test back in 2023. There's a list of corrective actions here.
https://imgur.com/a/Y9dd43o
Even NASA years into their existence has suffered catastrophic fatal failures. Even with the best and most knowledgeable experts working on it we are ultimately still in the infancy of space flight. Just like airlines every incident we try and understand the cause and prevent it from happening again. Lastly what they are doing is incredibly difficult with probably thousands of things that could go wrong. I think they are doing an amazing job and hope one day, even if I miss it, that space flight becomes acceptable to all who wish to go to space.
I you are referring to the two Space Shuttle accidents, both of them could have been avoided with just a little bit more care - not launching in freezing temperatures for Challenger, and making sure insulation foam doesn't fall off the tank for Columbia.
The history of rocketry goes much further back than the space shuttle. The shuttle was supposed to be a step towards reusability but didn’t succeed or progress the way they thought it would. Starship is continuing that dream of full reusability and their approach is working. You can’t plan everything on paper when it comes to hardware especially when attempting things that have never been done before, you just don’t have the data in that case. You have to build prototypes and test them to destruction. All manufacturers do this.
In hindsight yes. The trick is knowing which of the thousands of things to do are necessary. And yes, that’s how you end up with preflight checklists
Can you name a space company with less failures? Also I think it is unfair to even compare SpaceX to anything else, because of the insane amount of starts / tests combined unparalleled creativity.
According to this website their current success rate is 99,18%. That's a good number I guess? Considering other companies did not even land their stages for years.
https://spaceinsider.tech/2024/07/31/ula-vs-spacex/#:~:text=....
Success rate isn’t a great metric for efficient initial work: it will keep improving as more launches are done, regardless of the initial work.
There's more to "overall success" then launch failure rate. Cost and time are very important, which are the other dimensions they are optimizing for here.
It says right there in your source that that figure refers to Falcon in particular. For comparison, Starship's current track record is 3/7 launch failures (+1 landing failure).
There's an order of magnitude difference between them. If they were cars, it'd be like comparing the smallest car you can think of vs one of the biggest tanks ever made.
I ignored those, since the starship at this stage can be considered a prototype. I am just trying to argue, that calling SpaceX unreliable, especially compared to its competitors and time to market, is bold.
The usual definition of success for a rocket is getting the payload to the intended orbit. Since Starship doesn't have a payload yet, at least not a real one, its "success rate" is not measuring the same thing.
I'd say that only the 7th mission was legitimately a failure, because there was some rerouting of flights outside the exclusion zone. The other six missions were successful tests since nothing other than the rocket itself was affected.
It’s like comparing the reliability of the Model 3 and the Cybertruck.
You cannot compare a mature product to something that is still under initial development.
That would be like comparing a 1-y.o.'s ability to run to a 10-y.o.'s. Of course the younger kid doesn't yet control their legs, but that doesn't mean it's going to stumble and fall forever.
It's just taxpayer money they're blowing up, so it doesn't really matter.
The taxpayer money is for r&d. We should be very tolerant of failure. Aggressively testing with real hardware is a key part of how we learn to make a more robust systems. Fear of failure and waste will slow down progress.
They're blowing up their own money, unless you still count it as being the taxpayer's after the government pays them for launch services.
R&D for starship has a several-billion-dollar NASA grant. Something like 30-50% of the money being blown up on this program is taxpayer money.
7 replies →
Starship program is funded in part by NASA as part of Artemis program. So some of this money is ours.
1 reply →
It sounds like he's talking to investors and not to general public.
In my experience in corporate america you communicate efficiency by proclaiming a checklist of things to do - plausible, but not necessarily accurate things - and then let engineers figure it out.
Nobody cares of the original checklist as long as the problem gets resolved. It's weird but it seems very hard to utter statement "I don't have specific answers but we have very capable engineers, I'm sure they will figure it out". It's always better to say (from the top of your head) "To resolve A, we will do X,Y and Z!". Then when A get's resolved, everyone praises the effort. Then when they query what actually was done it's "well we found out in fact what were amiss were I, J K".
He's talking to the FAA, because this will trigger an investigation and would usually mean months of no launches.
Fortunately (for him) he'll be President on Monday and can then order the FAA to let him do whatever he wants.
1 reply →