4 ms·

Your point is completely unfair. Sure if our goal was to mass manufacture a "cheap and reliable JWST" we could, with a lot of money, create an entire assembly

Your point is completely unfair. Sure if our goal was to mass manufacture a "cheap and reliable JWST" we could, with a lot of money, create an entire assembly line and benefit from economies of scale. And then what? What are we supposed to do with 100s of JWST? It's cool they only cost 200M$ each now, but we don't need them. What we need after JWST is a different telescope with different technologies to answer questions that JWST cannot answer and to follow up on the discoveries from JWST. SpaceX and Boeing are responding to a scale problem: getting as much mass as possible into space, getting as much people as possible from point A to B. You solve a scale problem with scale. JWST is science instrument, looking for answer to specific questions, once those answer are found we will want a different instrument to answer different questions. Scale does not help. Multiple JWSTs cannot see what a single one can't.

WalterBright 5y ago
Is it unfair? In Spring 2020, in this forum, the vaccine experts here told me in no uncertain terms that a vaccine would take 18 months to develop. I said it could be done in 6 if all the slack was removed, and things done serially were done concurrently. I was told that was all completely unreasonable and impossible. 6 months later, the vaccine was released. The JWST has what, 390 single points of failure, and cost $10B. I.e., it cannot afford to fail, so $10B was spent to ensure it would not fail. And it worked! But suppose for $1B one could build a less reliable JWST. They fail like Musk's early rockets, but since they're cheaper the failure is not career ending, and you can iterate the design fixing the things that broke rather than trying to make everything perfect. You wind up in the end spending a lot less money. BTW, the airplane industry long ago gave up trying to make parts that could not fail. (390 parts that could fail and end the mission is pretty darn risky.) Instead, one makes redundant systems. It is far, far cheaper, and yet more reliable than going for perfection. Yes, it'll be heavier, and that will cost more. It makes airliners heavier, too, and it indeed costs more. But in the end it costs less, much less. Do I know how to build satellite telescopes? Nope. But I do know how airliners are designed and built. And I know that Musk upended how rockets were designed and built. As for different technology telescopes, I bet a lot of the telescope could be the same from design to design, just changing the instrument package.
- rcxdude 5y ago
  > Yes, it'll be heavier, and that will cost more. It won't just cost more, it'll be impossible. Aircraft care a lot about weight, spacecraft are obsessed with it, because it's often an absolute limit based on what launch technology you have available and where you're trying to get. JWST doesn't have any spare weight to try to build more redundancy. > But suppose for $1B one could build a less reliable JWST Well, the calculus of many single failure points makes this questionable. Even if you're 99% confident in each point of failure, with 390 failure points you have a 2% chance of success. You need extremely reliable components. Also, it's not clear that the majority of the budget is being spent on increasing that reliability: certainly a lot goes into testing, but all of that testing is generally a lot cheaper than a launch and rebuild.
- sailingparrot 5y ago
  What I am saying is unfair is your insistence on trying to think of astronomy as a problem that can be resolved by scale. > that a vaccine would take 18 months to develop. I said it could be done in 6 Cool anecdote. I was also part of those who thought it would be done earlier, we agreed on this. This is not comparable at all. > so $10B was spent to ensure it would not fail. And it worked! But suppose for $1B one could build a less reliable JWST. That was the plan, JWST was supposed to cost 1B$, not 10B$. And then we were supposed to start building other telescope based on the lesson learned with that design such as: https://en.wikipedia.org/wiki/Large_Ultraviolet_Optical_Infrared_Surveyor https://en.wikipedia.org/wiki/Large_Ultraviolet_Optical_Infr.... But it ended costing 10B$, not because that's what anyone wanted, but because that's what happens when you try to push the enveloppe. Just like the A380 program ended up costing 25B instead of 9 planed. > Instead, one makes redundant systems. It is far, far cheaper, and yet more reliable than going for perfection The JWST does not have 390 single point of failures like the media like to say, a lot of those are highly preferable but not make or break: If the mirror wings failed to deploy, JWST would still work at a lower res. If some of the sunshade layers didn't deploy correctly, it would still work. If the momentum flap didn't deploy, it would still work but require more fuel to keep orbit stable, lowering life expectancy of the missions, Latching mechanisms and release bolts where all designed with tolerance for some of them to fails etc. There are was actually just a handful of true single point of failures with no redundancy. > But I do know how airliners are designed and built Yes I think by now we are all aware since you keep mentioning it in every single one of your comment. And I don't doubt that you know what you are talking about when discussing airliners, but it also seems to me that your experience is a hammer and now everything you see looks like a nail to you. What I find irritating, is that the team that build the JWST is made of many, many highly intelligent individuals that thought about about how to best respond to the need of astronomy over 25 years, and you seem to think that they are complete idiots that didn't think about solutions and problem that you can think of in 5 minutes on hacker news. They have. A lot of those people come from the world of airplanes. One the biggest contractor for the JWST is Lockheed, which knows a thing or two about designing airplanes and asssembly line. Hubble, was actually built like that, as it's actually a repurposed KH-11 reconnaissance satellite. Lockheed built 18 of those. Only one of them was ever ordered for astronomy. So do not worry, JWST is built like it is, not because no one thought about your ideas before, but because they don't help solve what we actually care about.

WalterBright 5y ago

Is it unfair? In Spring 2020, in this forum, the vaccine experts here told me in no uncertain terms that a vaccine would take 18 months to develop. I said it could be done in 6 if all the slack was removed, and things done serially were done concurrently. I was told that was all completely unreasonable and impossible. 6 months later, the vaccine was released. The JWST has what, 390 single points of failure, and cost $10B. I.e., it cannot afford to fail, so $10B was spent to ensure it would not fail. And it worked! But suppose for $1B one could build a less reliable JWST. They fail like Musk's early rockets, but since they're cheaper the failure is not career ending, and you can iterate the design fixing the things that broke rather than trying to make everything perfect. You wind up in the end spending a lot less money. BTW, the airplane industry long ago gave up trying to make parts that could not fail. (390 parts that could fail and end the mission is pretty darn risky.) Instead, one makes redundant systems. It is far, far cheaper, and yet more reliable than going for perfection. Yes, it'll be heavier, and that will cost more. It makes airliners heavier, too, and it indeed costs more. But in the end it costs less, much less. Do I know how to build satellite telescopes? Nope. But I do know how airliners are designed and built. And I know that Musk upended how rockets were designed and built. As for different technology telescopes, I bet a lot of the telescope could be the same from design to design, just changing the instrument package.

rcxdude 5y ago
> Yes, it'll be heavier, and that will cost more. It won't just cost more, it'll be impossible. Aircraft care a lot about weight, spacecraft are obsessed with it, because it's often an absolute limit based on what launch technology you have available and where you're trying to get. JWST doesn't have any spare weight to try to build more redundancy. > But suppose for $1B one could build a less reliable JWST Well, the calculus of many single failure points makes this questionable. Even if you're 99% confident in each point of failure, with 390 failure points you have a 2% chance of success. You need extremely reliable components. Also, it's not clear that the majority of the budget is being spent on increasing that reliability: certainly a lot goes into testing, but all of that testing is generally a lot cheaper than a launch and rebuild.
sailingparrot 5y ago
What I am saying is unfair is your insistence on trying to think of astronomy as a problem that can be resolved by scale. > that a vaccine would take 18 months to develop. I said it could be done in 6 Cool anecdote. I was also part of those who thought it would be done earlier, we agreed on this. This is not comparable at all. > so $10B was spent to ensure it would not fail. And it worked! But suppose for $1B one could build a less reliable JWST. That was the plan, JWST was supposed to cost 1B$, not 10B$. And then we were supposed to start building other telescope based on the lesson learned with that design such as: https://en.wikipedia.org/wiki/Large_Ultraviolet_Optical_Infrared_Surveyor https://en.wikipedia.org/wiki/Large_Ultraviolet_Optical_Infr.... But it ended costing 10B$, not because that's what anyone wanted, but because that's what happens when you try to push the enveloppe. Just like the A380 program ended up costing 25B instead of 9 planed. > Instead, one makes redundant systems. It is far, far cheaper, and yet more reliable than going for perfection The JWST does not have 390 single point of failures like the media like to say, a lot of those are highly preferable but not make or break: If the mirror wings failed to deploy, JWST would still work at a lower res. If some of the sunshade layers didn't deploy correctly, it would still work. If the momentum flap didn't deploy, it would still work but require more fuel to keep orbit stable, lowering life expectancy of the missions, Latching mechanisms and release bolts where all designed with tolerance for some of them to fails etc. There are was actually just a handful of true single point of failures with no redundancy. > But I do know how airliners are designed and built Yes I think by now we are all aware since you keep mentioning it in every single one of your comment. And I don't doubt that you know what you are talking about when discussing airliners, but it also seems to me that your experience is a hammer and now everything you see looks like a nail to you. What I find irritating, is that the team that build the JWST is made of many, many highly intelligent individuals that thought about about how to best respond to the need of astronomy over 25 years, and you seem to think that they are complete idiots that didn't think about solutions and problem that you can think of in 5 minutes on hacker news. They have. A lot of those people come from the world of airplanes. One the biggest contractor for the JWST is Lockheed, which knows a thing or two about designing airplanes and asssembly line. Hubble, was actually built like that, as it's actually a repurposed KH-11 reconnaissance satellite. Lockheed built 18 of those. Only one of them was ever ordered for astronomy. So do not worry, JWST is built like it is, not because no one thought about your ideas before, but because they don't help solve what we actually care about.