Nobody says "it's too expensive to build hydro plants" here, because all of our power is from hydro, and it's quite profitable for the government (who owns the power company)...
If the Hoover Dam would have cost $10 billion today, that would only serve to bump up my cost estimate by an order of magnitude. I don't think that there's much of a difference between a $15 trillion and $150 trillion public works project, both are effectively "infinity dollars".
One mile of this wall would seem to me to be like roughly five hoover dams. The hoover dam cost $750 million in today's dollars. So wouldn't one mile of this super-wall cost $3.75 billion, not $160 million?
On the other hand, building a concrete *anything* that is a thousand feet tall and 165 feet thick isn't easy. They're claiming that a one-mile stretch of the wall would cost $160 million, which comes out to 871.2 million cubic feet of concrete, or a cost per cubic foot (including labour and materials) of about $0.18. That sounds really unlikely to me.
Let me put it this way, the hoover dam is actually relatively similar to what we're talking about here. It's roughly 700 feet tall, varies from 45 to 600 feet thick, and is about a fifth of a mile wide... So let's say that the cross section of the hoover dam has about the same area as this proposed wall.
OK, so now we just need the length of the wall. Well, the circumference of the American midwest is roughly 3900 miles (cutting through the great lakes, because what the hell). So basically, what we need to do, is build the equivalent of roughly 20,000 hoover dams.
The hoover dam cost the equivalent of about $750 million to build. I suspect it would cost a lot more today than pure inflation would account for (unions, health and safety standards, etc), but let's say that technological progress would counteract all that...
So, $750 million, times 20,000... and we come up with $15 trillion.
It's actually a team of just 4 working on No Man's Sky. The other 6 people are working on other things (not everybody at the company is working on this one game).
My ISP is an indie ISP. They used to charge $0.25 per gig, but then the incumbent carriers got the federal regulator to increase their tariffed costs to smaller ISPs by an order of magnitude (incumbents now get to charge indies up to $20/megabit/month for connecting to their networks), so it had to go up.
My ISP does still have overage caps (of $50) on their slower tiers (as in it's unlimited after that much overage), and they are trying various strategies to ease the pain on customers (I have a 300GB cap, but they don't count upload, and they don't count usage during off-peak, which they've defined as 2AM to 8AM).
But if I open up my wifi to the general public, I won't have the ability to manage my bandwidth, to shift my downloads off-peak to reduce my billable usage.
I've got a reasonably fast connection (50 megs down, 10 megs up), but I have a cap. My ISP charges $0.50 per gig overage, who is going to pay for that when strangers pump my monthly bill up?
I don't get it, why are they auctioning money? Why don't they just exchange them for USD? They will necessarily get less than the market value for them, because nobody would buy money for more than it's worth...
Parachutes won't slow down something that big slow enough for it to survive landing on dry land. They put parachutes on the early Falcon 9, it didn't survive even a water landing.
The shuttle's SRBs used parachutes and survived, but they also hit the ocean, and you can't land in the ocean and be rapidly reusable (need to refurbish after the saltwater damage).
You also have little to no control over a parachute landing (if you also want to land at a sufficiently slow speed), so instead of being able to land rather precisely on a small pad, you'd need an absolutely enormous potential landing area. So you'd pretty much have to land at sea, which as I said, makes rapid reusability impossible.
The problem with stacking is the thermal/power situation. Specifically, how much power can a processor use before it's impractical to power and cool it? And when you have two or more processor dies stacked on top of eachother, the heatsink is only going to contact the topmost one. How do you remove that heat from the bottom one?
I suspect the answers to those questions are, it's not practical to use that much more power that we use in high-end desktop chips today (150-200W is probably the limit of practicality), and I recall some interesting stuff from IBM years ago where they were building vertical cooling channels into CPU dies to handle stacking, so that the heat could be moved from lower dies up to where it could be removed.
Perhaps the approach could be going with CPU designs that optimize for power consumption rather than performance (but still more efficient, consuming less power per unit of work), and then stack a bunch of them.
It'd be impossible to land something the shape of a Falcon 9 first stage precisely enough (and on its landing legs) when using a parachute, and it's easier to simply refuel a rocket than to refuel a rocket and replace the parachutes (which tends to be a somewhat destructive process, if you've seen pictures of the Dragon after parachute deployment, where the parachute cords are stored beneath ablative insulation that they rip out).
The fins have greater surface area, and work better at high speeds than regular fins. It has nothing to do with Mars, particularly because the Falcon 9 first stage will never leave Earth's atmosphere.
You wouldn't have to fly it back to the launch site; the Falcon 9 travels by road, and the diameter (3.66 metres) was specifically chosen to be the largest diameter that can be transported by truck on regular roads. It's a cost-saving measure.
Submarines, and other naval use? Perhaps. But the aircraft flying to Afghanistan are already going to be flying there for all their other stuff, it's not like you're dedicating an entire aircraft for one item.
It's not like you're dispatching the part on a plane all by itself. The military has their own transportation network, moving all sorts of stuff, and unless you're on the front somewhere, shipping companies can do it too.
Naval use, though, that's a pretty good counter-example, particularly on submarines where resupply is less frequent. It's much easier to ship something to a base somewhere than it is to ship something to a moving target.
I'm not necessarily debating the utility of using a 3D printer for small orders, I'm arguing that sending them out into the field for military use doesn't really make sense. The parts are probably already manufactured, so the time/cost difference is between simply shipping an already existing part, versus shipping a very large and heavy metal printer out to the front somewhere. Shipping doesn't take very long, so the low speed of 3D printing means you'd probably get the part faster by shipping it rather than printing it in the field. And you'd only save costs if you print a large number of things in the field to justify having shipped a huge printer.
And if the part doesn't exist, does it make sense to ship the printer to the field (again, those EOS printers are huge), or have the printer back home, where it can be printing other stuff than just what people nearby the deployed-to-field one would need?
While 3d printers that use sintering have their uses, the huge cost/weight and low speed of the 3D printer mean that you can ship a traditionally manufactured cost to wherever it needs to go faster and more cheaply than printing it in the field.
Why would there be any size limitations to laser sintering? I don't see any reason why it can't be scaled to any size required. SpaceX is building rocket engines using the process, for example. The rocket engines in question aren't exactly huge, but they still put out more than 16 thousand pounds of thrust.
I don't know why all your posts are so focused on the drive's internal checksums not detecting errors. As you say, that's very rare. A far more common occurrence, and one that I've seen many times, is a drive detecting corrupt data and being unable to correct it. At that point, it's up to the filesystem to use whatever redundancy you've provided (be it duplication or parity) to recover the lost data. The error correction on a drive can't do squat if a block is sufficiently corrupt.
You act as if the drive missing corruption is the problem. It's not.
Yes, there are. There are filesystems that do per-block checksums. If data corruption occurs, it knows about it as soon as it tries to read the block. If it has no redundancy, ZFS will tell you which file is corrupt and suggest restoring it from backup.
Slashdot Mirror
Nobody says "it's too expensive to build hydro plants" here, because all of our power is from hydro, and it's quite profitable for the government (who owns the power company)...
If the Hoover Dam would have cost $10 billion today, that would only serve to bump up my cost estimate by an order of magnitude. I don't think that there's much of a difference between a $15 trillion and $150 trillion public works project, both are effectively "infinity dollars".
One mile of this wall would seem to me to be like roughly five hoover dams. The hoover dam cost $750 million in today's dollars. So wouldn't one mile of this super-wall cost $3.75 billion, not $160 million?
On the other hand, building a concrete *anything* that is a thousand feet tall and 165 feet thick isn't easy. They're claiming that a one-mile stretch of the wall would cost $160 million, which comes out to 871.2 million cubic feet of concrete, or a cost per cubic foot (including labour and materials) of about $0.18. That sounds really unlikely to me.
Let me put it this way, the hoover dam is actually relatively similar to what we're talking about here. It's roughly 700 feet tall, varies from 45 to 600 feet thick, and is about a fifth of a mile wide... So let's say that the cross section of the hoover dam has about the same area as this proposed wall.
OK, so now we just need the length of the wall. Well, the circumference of the American midwest is roughly 3900 miles (cutting through the great lakes, because what the hell). So basically, what we need to do, is build the equivalent of roughly 20,000 hoover dams.
The hoover dam cost the equivalent of about $750 million to build. I suspect it would cost a lot more today than pure inflation would account for (unions, health and safety standards, etc), but let's say that technological progress would counteract all that...
So, $750 million, times 20,000... and we come up with $15 trillion.
It's actually a team of just 4 working on No Man's Sky. The other 6 people are working on other things (not everybody at the company is working on this one game).
My ISP is an indie ISP. They used to charge $0.25 per gig, but then the incumbent carriers got the federal regulator to increase their tariffed costs to smaller ISPs by an order of magnitude (incumbents now get to charge indies up to $20/megabit/month for connecting to their networks), so it had to go up.
My ISP does still have overage caps (of $50) on their slower tiers (as in it's unlimited after that much overage), and they are trying various strategies to ease the pain on customers (I have a 300GB cap, but they don't count upload, and they don't count usage during off-peak, which they've defined as 2AM to 8AM).
But if I open up my wifi to the general public, I won't have the ability to manage my bandwidth, to shift my downloads off-peak to reduce my billable usage.
I've got a reasonably fast connection (50 megs down, 10 megs up), but I have a cap. My ISP charges $0.50 per gig overage, who is going to pay for that when strangers pump my monthly bill up?
Fine, but that doesn't change my basic point. Why bother with an auction that will necessarily get less than an open market?
I don't get it, why are they auctioning money? Why don't they just exchange them for USD? They will necessarily get less than the market value for them, because nobody would buy money for more than it's worth...
Parachutes won't slow down something that big slow enough for it to survive landing on dry land. They put parachutes on the early Falcon 9, it didn't survive even a water landing.
The shuttle's SRBs used parachutes and survived, but they also hit the ocean, and you can't land in the ocean and be rapidly reusable (need to refurbish after the saltwater damage).
You also have little to no control over a parachute landing (if you also want to land at a sufficiently slow speed), so instead of being able to land rather precisely on a small pad, you'd need an absolutely enormous potential landing area. So you'd pretty much have to land at sea, which as I said, makes rapid reusability impossible.
Either that or the die double sided with a heat sink on both sides, that could let you stack three cpu's together.
So you're telling me you want to go back to the days of slot-loading CPUs :P
The problem with stacking is the thermal/power situation. Specifically, how much power can a processor use before it's impractical to power and cool it? And when you have two or more processor dies stacked on top of eachother, the heatsink is only going to contact the topmost one. How do you remove that heat from the bottom one?
I suspect the answers to those questions are, it's not practical to use that much more power that we use in high-end desktop chips today (150-200W is probably the limit of practicality), and I recall some interesting stuff from IBM years ago where they were building vertical cooling channels into CPU dies to handle stacking, so that the heat could be moved from lower dies up to where it could be removed.
Perhaps the approach could be going with CPU designs that optimize for power consumption rather than performance (but still more efficient, consuming less power per unit of work), and then stack a bunch of them.
It'd be impossible to land something the shape of a Falcon 9 first stage precisely enough (and on its landing legs) when using a parachute, and it's easier to simply refuel a rocket than to refuel a rocket and replace the parachutes (which tends to be a somewhat destructive process, if you've seen pictures of the Dragon after parachute deployment, where the parachute cords are stored beneath ablative insulation that they rip out).
The fins have greater surface area, and work better at high speeds than regular fins. It has nothing to do with Mars, particularly because the Falcon 9 first stage will never leave Earth's atmosphere.
You wouldn't have to fly it back to the launch site; the Falcon 9 travels by road, and the diameter (3.66 metres) was specifically chosen to be the largest diameter that can be transported by truck on regular roads. It's a cost-saving measure.
Submarines, and other naval use? Perhaps. But the aircraft flying to Afghanistan are already going to be flying there for all their other stuff, it's not like you're dedicating an entire aircraft for one item.
It's not like you're dispatching the part on a plane all by itself. The military has their own transportation network, moving all sorts of stuff, and unless you're on the front somewhere, shipping companies can do it too.
Naval use, though, that's a pretty good counter-example, particularly on submarines where resupply is less frequent. It's much easier to ship something to a base somewhere than it is to ship something to a moving target.
I'm not necessarily debating the utility of using a 3D printer for small orders, I'm arguing that sending them out into the field for military use doesn't really make sense. The parts are probably already manufactured, so the time/cost difference is between simply shipping an already existing part, versus shipping a very large and heavy metal printer out to the front somewhere. Shipping doesn't take very long, so the low speed of 3D printing means you'd probably get the part faster by shipping it rather than printing it in the field. And you'd only save costs if you print a large number of things in the field to justify having shipped a huge printer.
And if the part doesn't exist, does it make sense to ship the printer to the field (again, those EOS printers are huge), or have the printer back home, where it can be printing other stuff than just what people nearby the deployed-to-field one would need?
While 3d printers that use sintering have their uses, the huge cost/weight and low speed of the 3D printer mean that you can ship a traditionally manufactured cost to wherever it needs to go faster and more cheaply than printing it in the field.
Why would there be any size limitations to laser sintering? I don't see any reason why it can't be scaled to any size required. SpaceX is building rocket engines using the process, for example. The rocket engines in question aren't exactly huge, but they still put out more than 16 thousand pounds of thrust.
Well, the first time I saw IkeaHackers, I really did think it was an official ikea site. So I've got to side with IKEA here.
I don't know why all your posts are so focused on the drive's internal checksums not detecting errors. As you say, that's very rare. A far more common occurrence, and one that I've seen many times, is a drive detecting corrupt data and being unable to correct it. At that point, it's up to the filesystem to use whatever redundancy you've provided (be it duplication or parity) to recover the lost data. The error correction on a drive can't do squat if a block is sufficiently corrupt.
You act as if the drive missing corruption is the problem. It's not.
Uh huh. As somebody who uses raidz, and has a decent high-level idea of how it works, I'm going to say you're full of shit.
Where did I claim otherwise? You must have failed to read the part where I said "If ZFS or Btrfs has redundancy".
Nope, wagnerrp is correct. raidz does exactly what he describes, and your claim that raidz was "scrapped due to various problems" is incorrect.
Yes, there are. There are filesystems that do per-block checksums. If data corruption occurs, it knows about it as soon as it tries to read the block. If it has no redundancy, ZFS will tell you which file is corrupt and suggest restoring it from backup.
So? If ZFS or Btrfs has redundancy, an unreadable sector gets corrected just the same as a garbage sector.