Launching a brand new rocket every time is like manually deploying a cross-team app to production bugfree without any testing except unit tests. The only way you can have success is by building up organizational experience and discipline. I think this is why legacy companies like Boeing are hitting generational turbulence and why startups like this one are prone to disruption by staffing retention.
More and more "flight tested" rockets are seeming like they should be the standard if not requirement for human space flight.
SpaceX is only "flight testing" half of their rockets. The upper stages, which are equally as complex albeit smaller than lower stages, remain single use devices. And SpaceX has blown up plenty of rockets learning how to launch them. That a Japan has lost one doesn't speak ill of anything or anyone.
Boeing isn’t a great comparison for your point because their planes are reused for each flight and their recent drop in safety is not due to staff retention (or at least and drop in staff retention is another symptom, like their safety record, of the larger management problems behind Boeing)
Boeings recent aircraft failures have been their newest planes off the line encountering issues, not their flight proven ones, but actually I was referring to Boeing/ULA starliner trials moreso.
Because of flight proven boosters, dragon was flown on 26 missions before humans were introduced.. meanwhile starliner is about to launch 2 humans on a fully unflown stack after 1 successful unmanned flight out of 4 attempts.
Boeing isn't hitting generational turbulence, it's hitting the consequences of serious mismanagement over the past 2 decades. Right now is the least turbulent point in terms of availability of rocket engineering talent in the US, considering all the different companies out there that have built or are building rockets.
Everyone managing durable technology deals with this issue, from government leaders deciding on nuclear-powered submarines to each person with their phone: What has greater ROI: building/buying a new thing or repairing and maintaining an old thing?
Both terms of the ROI (R/I) expression differ for the new thing and old thing: the return, new and old will have different features and performance; and the investment, build/buy vs repair/maintain.
Reliability is one cost - the cost of building or of repairing the machine to perform at reliability level X. Certainly, we manufacture very many reliable new things - most people expect their new car or phone or house to have lower reliability costs (i.e., greater reliability) than an old one. Old servers have higher maintenance costs than new ones, generally speaking. Famously, when you drive your brand new car off the lot and its status flag switches to 'used', its value drops by half.
Is there evidence that reliability in particular costs less on old rockets, or that there is any consistency in those numbers across rocket types? Rockets vary greatly between vendors.
And yes, I know about the cost advantages of reusing rockets - how could anyone on HN miss it? That doesn't mean that the used rocket is more reliable than the unused one, only that its reliability costs are low enough that the overall costs net out in favor of the reused first stage - and then only up to a certain number of uses. Will we ever have discussion of anything space-related without this endless SpaceX-is-better stuff?
I feel like maybe your idea of reliability is different from what is typically referred to, but the idea is that a reused rocket has already proven that it can do the job, so with sufficient bathtub curve statistics, the risk of losing a payload is lowered.
A brand new rocket doesn't have that prior, you're left to assume that the entire thing has been assembled perfectly. While this assumption is often still fair, it'd be hard to call something genuinely reusable if it got less reliable after the first use.
The example of a car doesn't make a lot of sense, as even a purchased new car has been driven, while a new rocket has never flown.
> your idea of reliability is different from what is typically referred to
How does it differ, in your view? I think my view is standard among professionals.
> a reused rocket has already proven that it can do the job
It's the same with any used equipment, but as described, used equipment isn't generally considered more reliable.
> it'd be hard to call something genuinely reusable if it got less reliable after the first use.
All durable technology is reusable - cars, phones, etc. But almost none (that I can think of) are considered more reliable after they are used. Reliability is just one component of cost, and the overall cost of reuse is much less than buying new - buying a new car for each trip wouldn't be worth the cost.
There is some uncertainty about reliability in brand new things, but that is easily reduced through testing.
SpaceX fans seem to live, like so many today, in an alternate reality where they repeat their 'team' propaganda without the filter of skepticism. Just think about it for a moment - is a used car necessarily more reliable?
I figured you had a different idea of reliability because I figure you couldn't seriously be saying that something that has been used is less reliable than something that has literally never done the thing it's supposed to do.
Cars, phones etc all have to do the thing they're designed to do before they're sold to consumers. In that way they're all "flight proven" like SpaceX's boosters. By the point of a used sale, the price drops because you don't know how a random consumer will have treated the item.
There's a lot of irony in you accusing SpaceX fans of living in an alternate reality and repeating team propaganda without skepticism.
> Cars, phones etc all have to do the thing they're designed to do before they're sold to consumers.
Do you mean that someone individually uses each product sold? I don't think so.
Things wear down in use. It's just reality. Not only does the market value drop, but accounting systems recognize it through depreciation. You've tried to redefine 'used' as superior reliability, but that's only in the rhetorical world where 'war is peace', 'freedom is slavery', etc.
There is uncertainty to brand new, untested technology, but that's not what we're talking about. There is uncertainty to a more limited extent in unused individual products, but I've already addressed that.
Japan has lots of exceptional engineers, lots of people and institutions skilled in managing highly successful engineering projects, and lots of money. Why are there so many problems with space technology (or is that my imagination)?
I'm not looking for a cultural answer, which is usually nonsense. Is it just not a priority? Doesn't pay well?
You can't really test a rocket until you fly it. You can ground test subsystems and components all you want, but that's not a system-level test of a rocket. The only true system-level test of a rocket is to fly it. When you deal with volatile hypergolic substances subjected to extreme conditions all flowing through a complex system-of-systems that have never been through integrated testing together, you are going to have some spectacular failures. Doesn't matter how much money you have or how clever your engineers are or what a quality organization you've implemented.
Is it clear that they do have more problems? The U.S. Government blew up a majority of its initial rockets. U.S. private sector companies almost always have a failed first launch. SpaceX's Falcon 1 failed in 3 of its first 5 launches. Japan having a failed first flight is 100% what one would expect—it would be exceptional had it not failed.
21 comments
[ 4.0 ms ] story [ 60.2 ms ] threadMore and more "flight tested" rockets are seeming like they should be the standard if not requirement for human space flight.
Because of flight proven boosters, dragon was flown on 26 missions before humans were introduced.. meanwhile starliner is about to launch 2 humans on a fully unflown stack after 1 successful unmanned flight out of 4 attempts.
Both terms of the ROI (R/I) expression differ for the new thing and old thing: the return, new and old will have different features and performance; and the investment, build/buy vs repair/maintain.
Reliability is one cost - the cost of building or of repairing the machine to perform at reliability level X. Certainly, we manufacture very many reliable new things - most people expect their new car or phone or house to have lower reliability costs (i.e., greater reliability) than an old one. Old servers have higher maintenance costs than new ones, generally speaking. Famously, when you drive your brand new car off the lot and its status flag switches to 'used', its value drops by half.
Is there evidence that reliability in particular costs less on old rockets, or that there is any consistency in those numbers across rocket types? Rockets vary greatly between vendors.
And yes, I know about the cost advantages of reusing rockets - how could anyone on HN miss it? That doesn't mean that the used rocket is more reliable than the unused one, only that its reliability costs are low enough that the overall costs net out in favor of the reused first stage - and then only up to a certain number of uses. Will we ever have discussion of anything space-related without this endless SpaceX-is-better stuff?
A brand new rocket doesn't have that prior, you're left to assume that the entire thing has been assembled perfectly. While this assumption is often still fair, it'd be hard to call something genuinely reusable if it got less reliable after the first use.
The example of a car doesn't make a lot of sense, as even a purchased new car has been driven, while a new rocket has never flown.
How does it differ, in your view? I think my view is standard among professionals.
> a reused rocket has already proven that it can do the job
It's the same with any used equipment, but as described, used equipment isn't generally considered more reliable.
> it'd be hard to call something genuinely reusable if it got less reliable after the first use.
All durable technology is reusable - cars, phones, etc. But almost none (that I can think of) are considered more reliable after they are used. Reliability is just one component of cost, and the overall cost of reuse is much less than buying new - buying a new car for each trip wouldn't be worth the cost.
There is some uncertainty about reliability in brand new things, but that is easily reduced through testing.
SpaceX fans seem to live, like so many today, in an alternate reality where they repeat their 'team' propaganda without the filter of skepticism. Just think about it for a moment - is a used car necessarily more reliable?
Cars, phones etc all have to do the thing they're designed to do before they're sold to consumers. In that way they're all "flight proven" like SpaceX's boosters. By the point of a used sale, the price drops because you don't know how a random consumer will have treated the item.
There's a lot of irony in you accusing SpaceX fans of living in an alternate reality and repeating team propaganda without skepticism.
Do you mean that someone individually uses each product sold? I don't think so.
Things wear down in use. It's just reality. Not only does the market value drop, but accounting systems recognize it through depreciation. You've tried to redefine 'used' as superior reliability, but that's only in the rhetorical world where 'war is peace', 'freedom is slavery', etc.
There is uncertainty to brand new, untested technology, but that's not what we're talking about. There is uncertainty to a more limited extent in unused individual products, but I've already addressed that.
I don't know what team you think I'm on?
I'm not looking for a cultural answer, which is usually nonsense. Is it just not a priority? Doesn't pay well?