Actually, the whole *point* of this particular new line of ARM chips is that they become as integrated as your average AVR processor, with all of the usual trappings. There isn't any external memory or peripheral bus. 48-64 pins is the high-end of the AVR pin-count.
As a mere amateur (but then, this is Slashdot) I'd say that this particular design will mean that your board will be roughly equal in size to one built around an AVR. Except that there's more CPU power and memory to work with, which can be then used so that you have an easier time building the software.
And that's nothing to sneeze at. There's a definate potential to replace external parts with more software logic. Like moving the power conversion/battery control logic onto the chip instead of having it external.
There's the definate potential if you can get a 32-bit Atmel ARM for the same price as a PIC or AVR for people to pick the ARM and squeeze more functionality in, even if all it means is that the system has self-diagnostics that it wouldn't otherwise or is written in C instead of Assembler and uses a real embedded OS instead of a few handy subroutines.
The big thing is that Atmel is going after Microchip's PIC processors. I'd view it as a coup for them. The ARM is a "real" embedded processor that Atmel has been working with for a while. This is an act that it will be harder for Microchip to follow.
Remapping the Stinkpad keyboard so that my fingers fall in the right place would probably require a certain amount of hardware bending. Otherwise, it would be about as useful as trying to get a old-style Sun-layout keyboard to act like a PC keyboard.
The problem is that my fingers expect a Windows key, an Alt key, and a Control key to the left of the spacebar. Other laptops accomidate this perfectly well. The other problem is that the keyboard controller itself is poorly designed. When I am using a full-size keyboard, I like to have the numeric keypad readily available as number keys -- numlock key on. As we all know, all of the numeric keypad functionality with numlock off is already accessible via other keys. However, IBM, in their infinite wisdom, decided that the numlock key not only means that I want numbers instead of cursor movement, but that it also means that I want the keyboard-overlay numeric keypad instead of alphabetic keys.
Between that and the still-not-fixed-as-of-a-few-months-back utility software means that there's an annoying ritual that I need to go through every single time I dock and undock my laptop.
And the thing is, IBM makes big noises about doing *research* about user interface and being *good* at it. Yet in so many cases, they've clearly just made things worse.
Ummm.. There's no provisions for external memory. This is aimed at AVR designers who want more oomph, so all memory and flash is internal. No address bus.
The problem, of course, is that a TQFP package is not quite as hobyist-hackable as the old DIP packages because it requires you to have etched PCBs or a prototype adapter, which makes breadboarding harder.
Dude, you make it sound like IBM makes good hardware. But, in fact, hate my Stinkpad. Apparently, there's an astonishingly failure rate on the T30 mainboards. And we can't forget about IBM's 75GXP failure fun.
Remember, the AT&T vs. BSD agreement was never completely disclosed. Just because it's a public company doesn't mean that every piece of relevant information about the operations of the company needs to be disclosed to the shareholders.
They can't afford to take the risk of a $1b judgement. That's why they settled. And the details of the settlement are probably not available for disclosure because it's in neither Sun nor Kodak's best interest to disclose them.
The shuttle costs too much because it takes a lot of effort to recondition everything, and if anything breaks, you've just lost a crew, so you'd better make sure it's right.
The big thing that it's showing is that space startups have finally learned their lesson.
There have been startups since the 70s trying to get to orbit for cheap, but they've either been squeezed out by the existing players or run out of money before they actually prove their point, never even finishing the prototypes.
The 2000s startups are much smarter and have managed to get flying and actually produce real hardware.
The biggest contribution to progress, I think, is showing that if you don't run out of money, you *can* deliver on your promises. Which really tends to help other, newer, startups find investors so that they can have a little better assurance that they won't be pouring money down the drain.
Armadillo still has the option to go either way. They have a mixed-monopropellant engine that's good for cheap suborbital flights -- a little bit of methanol mixed in with hydrogen peroxide gives you a reasonable amount of boost. And they are also working on a conventional bipropellant engine. They don't need to decide on where to put their effort for another year, according to the latest updates.
Also don't forget SpaceX, who is completely not an X-prize contender but is simply concentrating on reducing the cost of payload to orbit, but by a large margin. Or XCOR, who has been spending a lot of time on economical and safe liquid-fueled engines.
Rutan did next to nothing as far as propellant development goes. Two hybrid-rocket startups competed for the job. So he doesn't care about weather there's a HTPB+NOS engine or if there's an XCOR liquid rocket, as long as it fits in the bolt-on propulsion section.
There's a number of interesting ways to get down from space that haven't been adequately explored. The denser your vehicle, the hotter it is on reentry, so simply carying more fuel in the upper stage, or having an SSTO, means that you can use lighter weight heat shielding. Or you can just have a carbon-carbon base on your craft and Burt's orange ablative covering on the rest and fly a feathered approach like SpaceShipOne.
Remember, the VSS Enterprise is a new design. Same aerodynamic shapings, but they are going to have to redesign it to be larger, so they will probably be fixing the roll issue at that point.
This is normal. A prototype will have various little problems here and there that need to be fixed. They end up adding a little fin or adjust the controls or things like that to fix the quirk and move on. Most aircraft require some tweaks after the first few flights, even when they are designed by genius designers.
Remember, the roll was happening when there wasn't very much atmosphere, so they were well within the stress limits of the craft. At no point was it in danger; once SS1 was feathered, it didn't matter what the orientation was, it could have been upside-down, backwards, and rolling and it would still get down OK. They intentionally re-entered upside-down on a test flight to prove that it was, in fact, tollerant of trouble.
No, there was no vehicle changes. They just knew where it would have a propensity to roll and flew to avoid it. You have to remember that it was nowhere near the point of structural damage, so as long as Mike didn't black out, the rolling motion would go away when reentry started.
The problem is that the amount of energy dissapated is much much higher.
So while you can probably use at least some of that technique to make a orbital vehicle a little more "trouble free", you still need to dissipate more energy, either by having an orbital shuttle be not especially dense and getting a little hotter, or being prepared to have your vehicle a *lot* hotter and build it out of high-temperature materials like the X-33 or space shuttle.
It's that Burt and his famous predecessors in the same veign (Lockheed's Skunk Works is one of these) are content to have a design that's 80% of the best it can be yet can be made quickly and efficently, instead of trying to be 100% perfect.
Sometimes, that means that you stick-and-rudder fly it, but sometimes it doesn't. Remember that Black Armadillo is equally pragmatic, yet they are 100% computerized.
Rutan's *always* been running his mouth and complaining about stuff. He first worked for the government. Didn't like the bureacracy. Then he started making tiny homebuilt aircraft. Got sued by the family of an "escort" who had the poor choice to go flying in the back of a Rutan-designed homebuilt that was piloted by some drunk rich dude. Of course he's not going to be designing an aircraft that would be safe in the hands of a drunk pilot, but he got sued anyway. He built a prototype and helped with the design of a Raytheon/Beech aircraft that is currently being purchased up to be scrapped because Beech wants to bury the memory. Now he does experimental, revolutionary craft at a tiny fraction of the cost it would take a "normal" group of people. He, like every homebuilder, knows that it's expensive FAA requirements and a depressed general aviation market in general, that prevents him from selling fully complete LongEZs from a dealership right by the local landing strip.
The thing is, we came no closer to being a truly spacefaring people in the decades of 1980-2000. So he's not the only one who's annoyed at the promise of our "other" space program being squandered, sometimes because of the fault of the program itself, sometimes just because of Congress.
The thing to remember about being privately funded is that they have the same sort of usually self-preservation-related tendnacy to check, recheck, and document every tightened bolt of importance. I mean, every Helecopter has the "Jesus" bolt where if it comes loose, the rotors fall off the aircraft. Naturally, they are routinely inspected, carefully torqued down, monitored, etc. The trick is just to have fewer bolts to tighten, not to just wing it and hope nothing comes undone.
The thing to remember is that the shuttle is far from perfect. It's cheaper, for the same amount of payload, to put it atop an expendable booster and launch it that way. It was supposed to be the other way around. It's an accomplishment, but only in the same was as the NS Savannah was -- A technologically advanced form of transportation that just didn't make fiscal sense.
True, but, on the other hand, any not-currently-patented algorythms used in open-source will be patent-free in the future. Remember, as long as the source code is present, it can be used as prior art in any patent dispute.
So, it sucks to be you, but is good for everybody else.
However, in either case, you are dealing with a lot of modifications. They managed to convince the F-16's FBW system to work in the F-117 prototype and the F-15's to make the DC-X work. I don't know if anybody's tried that with Airbus avionics in the same way, but I imagine that a fighter's control computer would be much more able to handle the changes.
And these modifications would have to be tested.
It just adds time and money to the program. I think it's pretty cool that they've managed to do it with all simple controls. Remember, the Boeing "Bird of Prey" stealth aircraft prototype also doesn't require any fly-by-wire.
As an experimental aircraft, any avionics aren't required to be certified in the same way as a production aircraft. However, if it, or a derivative, makes it to the production line, it will require certification.
Remember, he's got a stick for subsonic flight. He's got trim for supersonic flight. And then he's got thrusters for space usage. Plus backup systems, which you have to know when they should be activated. So it can't just be an off-the-shelf system.
The thing is, if you *needed* the autopilot, you'd need to have redundancy and reliability and whatnot. If you don't *need* the autopilot, it's an added expense, a waste of time, and it takes up weight that can be used for something else. So, for an experimental aircraft that's going to be flown by Scaled's best pilots, why not?
The other problem is that the main folks who have an off-the-shelf flight computer that would be suitable is the Air Force. Who obviously isn't going to sell one to "just anyone", which means that an X-prize contender can't have it.
Yeah, it was nerve-wracking as all hell live. It looked, based on inadequate resolution, like not only was it rolling, it was pitching and yawing.
Thankfully it was "just" rolling.
The other part was that a bunch of folks were on IRC gabbing about it, which meant that some folks were 30 seconds less lagged, which meant that I had 30 extra seconds of tension.
Well, the next one to get a launch vehicle working will be able to compete with SpaceShipOne for the real payoff -- commercial spaceline companies. Especially if they can do it for less money, safer, or better.
Also, all of your equipment is generally certified for 1 ATM and the cooling available from that amount of atmosphere. Lower pressures require different components.
On the other hand, it does mean that you don't need to spend as much time pre-breathing before you go out in your low-pressure suit.
True, but is the ice going to be structurally sound enough and not evaporate? The advantage of an inflatable module is that it's compact and lightweight, but I'd say that if you want ease of repair for Things That Can't Be Replaced, you'd probably stick with a hard-shelled Aluminum case.
So a mars probe with an inflatable habitat would end up like the characters in Dark Star -- Sleeping in the meat locker.
They just handle space debris like all of the other modules. They use a series of hard-and-soft layers to deaccelerate and absorb the energy of impacts. In order to even be considered for use, it *must* be at least as good as the hard-sided modules. If you RTFA, you'd read that they already fired projectiles at orbital velocity at the module. They found out that it works just fine, except that it's cheaper.
You also have to remember that astronauts have never actually repaired any sort of hull puncture in space.
It's Kevlar fiber, generally, along with a variety of other materials, all of which have been tested in space or are currently on the space station. Just because a cheap injection-molded plastic toy breaks easily doesn't mean that all non-metallic materials are easily broken.
They've already worked the water supply angle out there. It depends on the project. The Transhab had the water supply going down the center, so you could stay in the shadow of it during radiation events.
Leakproofing is one of the problems they solve to make it all work. Just because a cheap rubber balloon can't hold pressure for years on end doesn't mean that all non-metallic materials have leaks.
A ballon is a few PSI of pressure higher than the surrounding atmosphere. A inflatable space module is 14.7 PSI higher than the surrounding atmosphere.
All space modules need to deal with being inflated at 14.7 PSI. It's just that the whole point of the inflatable space module, like the older Atlas boosters, are designed such that they work *with* the overpresure instead of working against it.
The main problem is if you leak out all of your atmosphere, it won't stay inflated. But generally, it not staying inflated is the least of your problems.
You are thinking that the walls will be like a rubber ballon.
The walls will be thick enough to provide the same level of orbital debris protection as existing space station modules. Remember, the Atlas booster had aluminum ballon fuel tanks -- it would actually collapse if the tanks were empty. When other engineers were suggesting it would never fly, they presurized it and gave them a sledgehammer to try to break the booster.
It rebounded and almost hurt the engineer swinging the hammer.
The instant-fill holes are harder than you'd think. Instant-fill tire stuff is designed to work in an atmosphere that you aren't breathing.
So you don't fly airliners because they experience turbulence and therefore aren't 100% reliable.
I'd love to see the ticket counter agents make fun of you after you leave if you tried to argue that you should be refunded your ticket on the grounds of too much turbulence.
Slashdot Mirror
Actually, the whole *point* of this particular new line of ARM chips is that they become as integrated as your average AVR processor, with all of the usual trappings. There isn't any external memory or peripheral bus. 48-64 pins is the high-end of the AVR pin-count.
As a mere amateur (but then, this is Slashdot) I'd say that this particular design will mean that your board will be roughly equal in size to one built around an AVR. Except that there's more CPU power and memory to work with, which can be then used so that you have an easier time building the software.
And that's nothing to sneeze at. There's a definate potential to replace external parts with more software logic. Like moving the power conversion/battery control logic onto the chip instead of having it external.
There's the definate potential if you can get a 32-bit Atmel ARM for the same price as a PIC or AVR for people to pick the ARM and squeeze more functionality in, even if all it means is that the system has self-diagnostics that it wouldn't otherwise or is written in C instead of Assembler and uses a real embedded OS instead of a few handy subroutines.
The big thing is that Atmel is going after Microchip's PIC processors. I'd view it as a coup for them. The ARM is a "real" embedded processor that Atmel has been working with for a while. This is an act that it will be harder for Microchip to follow.
Remapping the Stinkpad keyboard so that my fingers fall in the right place would probably require a certain amount of hardware bending. Otherwise, it would be about as useful as trying to get a old-style Sun-layout keyboard to act like a PC keyboard.
The problem is that my fingers expect a Windows key, an Alt key, and a Control key to the left of the spacebar. Other laptops accomidate this perfectly well. The other problem is that the keyboard controller itself is poorly designed. When I am using a full-size keyboard, I like to have the numeric keypad readily available as number keys -- numlock key on. As we all know, all of the numeric keypad functionality with numlock off is already accessible via other keys. However, IBM, in their infinite wisdom, decided that the numlock key not only means that I want numbers instead of cursor movement, but that it also means that I want the keyboard-overlay numeric keypad instead of alphabetic keys.
Between that and the still-not-fixed-as-of-a-few-months-back utility software means that there's an annoying ritual that I need to go through every single time I dock and undock my laptop.
And the thing is, IBM makes big noises about doing *research* about user interface and being *good* at it. Yet in so many cases, they've clearly just made things worse.
Ummm.. There's no provisions for external memory. This is aimed at AVR designers who want more oomph, so all memory and flash is internal. No address bus.
The problem, of course, is that a TQFP package is not quite as hobyist-hackable as the old DIP packages because it requires you to have etched PCBs or a prototype adapter, which makes breadboarding harder.
Dude, you make it sound like IBM makes good hardware. But, in fact, hate my Stinkpad. Apparently, there's an astonishingly failure rate on the T30 mainboards. And we can't forget about IBM's 75GXP failure fun.
Not always.
Remember, the AT&T vs. BSD agreement was never completely disclosed. Just because it's a public company doesn't mean that every piece of relevant information about the operations of the company needs to be disclosed to the shareholders.
They can't afford to take the risk of a $1b judgement. That's why they settled. And the details of the settlement are probably not available for disclosure because it's in neither Sun nor Kodak's best interest to disclose them.
The shuttle costs too much because it takes a lot of effort to recondition everything, and if anything breaks, you've just lost a crew, so you'd better make sure it's right.
The big thing that it's showing is that space startups have finally learned their lesson.
There have been startups since the 70s trying to get to orbit for cheap, but they've either been squeezed out by the existing players or run out of money before they actually prove their point, never even finishing the prototypes.
The 2000s startups are much smarter and have managed to get flying and actually produce real hardware.
The biggest contribution to progress, I think, is showing that if you don't run out of money, you *can* deliver on your promises. Which really tends to help other, newer, startups find investors so that they can have a little better assurance that they won't be pouring money down the drain.
Armadillo still has the option to go either way. They have a mixed-monopropellant engine that's good for cheap suborbital flights -- a little bit of methanol mixed in with hydrogen peroxide gives you a reasonable amount of boost. And they are also working on a conventional bipropellant engine. They don't need to decide on where to put their effort for another year, according to the latest updates.
Also don't forget SpaceX, who is completely not an X-prize contender but is simply concentrating on reducing the cost of payload to orbit, but by a large margin. Or XCOR, who has been spending a lot of time on economical and safe liquid-fueled engines.
Rutan did next to nothing as far as propellant development goes. Two hybrid-rocket startups competed for the job. So he doesn't care about weather there's a HTPB+NOS engine or if there's an XCOR liquid rocket, as long as it fits in the bolt-on propulsion section.
There's a number of interesting ways to get down from space that haven't been adequately explored. The denser your vehicle, the hotter it is on reentry, so simply carying more fuel in the upper stage, or having an SSTO, means that you can use lighter weight heat shielding. Or you can just have a carbon-carbon base on your craft and Burt's orange ablative covering on the rest and fly a feathered approach like SpaceShipOne.
He's not worth that.
Especially if he breaks it. SS1 belongs in a museum so when I have kids, I can take them to see it.
No, I think that all Darl deserves is to go on a parachute flight with an empty backpack labeled "PARACHUTE" strapped to his back. Much more efficent.
SS1 isn't going to be the commercial vehicle.
Remember, the VSS Enterprise is a new design. Same aerodynamic shapings, but they are going to have to redesign it to be larger, so they will probably be fixing the roll issue at that point.
This is normal. A prototype will have various little problems here and there that need to be fixed. They end up adding a little fin or adjust the controls or things like that to fix the quirk and move on. Most aircraft require some tweaks after the first few flights, even when they are designed by genius designers.
Remember, the roll was happening when there wasn't very much atmosphere, so they were well within the stress limits of the craft. At no point was it in danger; once SS1 was feathered, it didn't matter what the orientation was, it could have been upside-down, backwards, and rolling and it would still get down OK. They intentionally re-entered upside-down on a test flight to prove that it was, in fact, tollerant of trouble.
No, there was no vehicle changes. They just knew where it would have a propensity to roll and flew to avoid it. You have to remember that it was nowhere near the point of structural damage, so as long as Mike didn't black out, the rolling motion would go away when reentry started.
The problem is that the amount of energy dissapated is much much higher.
So while you can probably use at least some of that technique to make a orbital vehicle a little more "trouble free", you still need to dissipate more energy, either by having an orbital shuttle be not especially dense and getting a little hotter, or being prepared to have your vehicle a *lot* hotter and build it out of high-temperature materials like the X-33 or space shuttle.
I'd disagree.
It's not the computers that's the problem.
It's that Burt and his famous predecessors in the same veign (Lockheed's Skunk Works is one of these) are content to have a design that's 80% of the best it can be yet can be made quickly and efficently, instead of trying to be 100% perfect.
Sometimes, that means that you stick-and-rudder fly it, but sometimes it doesn't. Remember that Black Armadillo is equally pragmatic, yet they are 100% computerized.
I'd disagree.
Rutan's *always* been running his mouth and complaining about stuff. He first worked for the government. Didn't like the bureacracy. Then he started making tiny homebuilt aircraft. Got sued by the family of an "escort" who had the poor choice to go flying in the back of a Rutan-designed homebuilt that was piloted by some drunk rich dude. Of course he's not going to be designing an aircraft that would be safe in the hands of a drunk pilot, but he got sued anyway. He built a prototype and helped with the design of a Raytheon/Beech aircraft that is currently being purchased up to be scrapped because Beech wants to bury the memory. Now he does experimental, revolutionary craft at a tiny fraction of the cost it would take a "normal" group of people. He, like every homebuilder, knows that it's expensive FAA requirements and a depressed general aviation market in general, that prevents him from selling fully complete LongEZs from a dealership right by the local landing strip.
The thing is, we came no closer to being a truly spacefaring people in the decades of 1980-2000. So he's not the only one who's annoyed at the promise of our "other" space program being squandered, sometimes because of the fault of the program itself, sometimes just because of Congress.
The thing to remember about being privately funded is that they have the same sort of usually self-preservation-related tendnacy to check, recheck, and document every tightened bolt of importance. I mean, every Helecopter has the "Jesus" bolt where if it comes loose, the rotors fall off the aircraft. Naturally, they are routinely inspected, carefully torqued down, monitored, etc. The trick is just to have fewer bolts to tighten, not to just wing it and hope nothing comes undone.
The thing to remember is that the shuttle is far from perfect. It's cheaper, for the same amount of payload, to put it atop an expendable booster and launch it that way. It was supposed to be the other way around. It's an accomplishment, but only in the same was as the NS Savannah was -- A technologically advanced form of transportation that just didn't make fiscal sense.
True, but, on the other hand, any not-currently-patented algorythms used in open-source will be patent-free in the future. Remember, as long as the source code is present, it can be used as prior art in any patent dispute.
So, it sucks to be you, but is good for everybody else.
See, that occured to me.
However, in either case, you are dealing with a lot of modifications. They managed to convince the F-16's FBW system to work in the F-117 prototype and the F-15's to make the DC-X work. I don't know if anybody's tried that with Airbus avionics in the same way, but I imagine that a fighter's control computer would be much more able to handle the changes.
And these modifications would have to be tested.
It just adds time and money to the program. I think it's pretty cool that they've managed to do it with all simple controls. Remember, the Boeing "Bird of Prey" stealth aircraft prototype also doesn't require any fly-by-wire.
As an experimental aircraft, any avionics aren't required to be certified in the same way as a production aircraft. However, if it, or a derivative, makes it to the production line, it will require certification.
Well, because it's harder than you think.
Remember, he's got a stick for subsonic flight. He's got trim for supersonic flight. And then he's got thrusters for space usage. Plus backup systems, which you have to know when they should be activated. So it can't just be an off-the-shelf system.
The thing is, if you *needed* the autopilot, you'd need to have redundancy and reliability and whatnot. If you don't *need* the autopilot, it's an added expense, a waste of time, and it takes up weight that can be used for something else. So, for an experimental aircraft that's going to be flown by Scaled's best pilots, why not?
The other problem is that the main folks who have an off-the-shelf flight computer that would be suitable is the Air Force. Who obviously isn't going to sell one to "just anyone", which means that an X-prize contender can't have it.
Yeah, it was nerve-wracking as all hell live. It looked, based on inadequate resolution, like not only was it rolling, it was pitching and yawing.
Thankfully it was "just" rolling.
The other part was that a bunch of folks were on IRC gabbing about it, which meant that some folks were 30 seconds less lagged, which meant that I had 30 extra seconds of tension.
Well, the next one to get a launch vehicle working will be able to compete with SpaceShipOne for the real payoff -- commercial spaceline companies. Especially if they can do it for less money, safer, or better.
There's other problems at play, too.
You don't want to recreate Apollo 1, for one.
Also, all of your equipment is generally certified for 1 ATM and the cooling available from that amount of atmosphere. Lower pressures require different components.
On the other hand, it does mean that you don't need to spend as much time pre-breathing before you go out in your low-pressure suit.
True, but is the ice going to be structurally sound enough and not evaporate? The advantage of an inflatable module is that it's compact and lightweight, but I'd say that if you want ease of repair for Things That Can't Be Replaced, you'd probably stick with a hard-shelled Aluminum case.
So a mars probe with an inflatable habitat would end up like the characters in Dark Star -- Sleeping in the meat locker.
They just handle space debris like all of the other modules. They use a series of hard-and-soft layers to deaccelerate and absorb the energy of impacts. In order to even be considered for use, it *must* be at least as good as the hard-sided modules. If you RTFA, you'd read that they already fired projectiles at orbital velocity at the module. They found out that it works just fine, except that it's cheaper.
You also have to remember that astronauts have never actually repaired any sort of hull puncture in space.
It's not plastic.
It's Kevlar fiber, generally, along with a variety of other materials, all of which have been tested in space or are currently on the space station. Just because a cheap injection-molded plastic toy breaks easily doesn't mean that all non-metallic materials are easily broken.
They've already worked the water supply angle out there. It depends on the project. The Transhab had the water supply going down the center, so you could stay in the shadow of it during radiation events.
Leakproofing is one of the problems they solve to make it all work. Just because a cheap rubber balloon can't hold pressure for years on end doesn't mean that all non-metallic materials have leaks.
Just like any other object, in both cases.
A ballon is a few PSI of pressure higher than the surrounding atmosphere. A inflatable space module is 14.7 PSI higher than the surrounding atmosphere.
All space modules need to deal with being inflated at 14.7 PSI. It's just that the whole point of the inflatable space module, like the older Atlas boosters, are designed such that they work *with* the overpresure instead of working against it.
The main problem is if you leak out all of your atmosphere, it won't stay inflated. But generally, it not staying inflated is the least of your problems.
You are thinking that the walls will be like a rubber ballon.
The walls will be thick enough to provide the same level of orbital debris protection as existing space station modules. Remember, the Atlas booster had aluminum ballon fuel tanks -- it would actually collapse if the tanks were empty. When other engineers were suggesting it would never fly, they presurized it and gave them a sledgehammer to try to break the booster.
It rebounded and almost hurt the engineer swinging the hammer.
The instant-fill holes are harder than you'd think. Instant-fill tire stuff is designed to work in an atmosphere that you aren't breathing.
So you don't fly airliners because they experience turbulence and therefore aren't 100% reliable.
I'd love to see the ticket counter agents make fun of you after you leave if you tried to argue that you should be refunded your ticket on the grounds of too much turbulence.