Basically the underground temperature will roughly be the mean year-round overground temperature, so long as you're not in a geothermally active area. That's where your 10-13C number comes from. But that only applies if you don't inject more heat than the ground can conduct away. The year-round mean temperature in London is probably around 10C. But the temperature in the London Underground in mid summer on the deep lines is more like 30C. Too much heat generated by the trains and people for the rock/clay to conduct away. A data centre generates for more heat in a small space than an underground railway does.
OK, so I installed the Silverlight plugin for MacOS, just to see. Oh dear. My two-year old Mac Mini which can play DVDs, H.264, iPlayer, or Flash video, full screen, no problem, cannot play these videos without stuttering. 3 seconds of video, 200ms pause, 3 seconds of video, 200ms pause. Unwatchable. I'd have advised people against Silverlight anyway, but issues to do with licensing don't work well with many people. Now I can say with certainty to my less enlightened friends and colleagues: avoid it - it doesn't work well enough for production use.
Guess I'll have to watch the YouTube versions now.
If the materials were published in the US just how can an English court have jurisdiction? Are they claiming that because English addresses received the materials that English laws are in play?
I think the reasoning was probably that they were in England when they published for a British audience. The act of publishing was done from England (a server itself isn't capable of independent action, at least not in the eyes of the law).
If they'd flown to the US, written the materials there, published them from there, then returned home, I think they would probably have not have been able to be convicted, even if the intended audience was British.
But I'm not a lawyer, so don't take this as legal advice.
If I recall correctly, Native Client is x86 specific because it uses segmentation to go fast. I believe you should be able to do the same thing with paging on ARM, but I'd expect there to be some performance penalty.
Using your numbers, the power station is 60% efficient, the grid is 95% efficient (a guess) and the heat pump is 300% efficient. 0.6 * 0.95 * 3 = 1.7.
You get 70% more heat for your gas burning it in a power station and using a heat pump than you would burning it at home (and that assumes the home furnace is 100% efficient, so in practice you do even better than this).
No, electricity using a heat pump is the best way to heat, short of geothermal or solar thermal.
OK, but where does that heat go from the ceiling your incandescent lightbulbs warmed? Some of it radiates back to the rest of the room, and some warms the room upstairs. The heat is only wasted if it is on the top floor and the loft is poorly insulated. And then not all light bulbs are ceiling mounted - we've plenty of table or standard lamps too.
There's a much better reason why you don't want to use lightbulbs to generate heat, which is that resistive heating makes very little sense. Either use gas, where you get a good efficiency by burning it at the point you want heat, or even better, use a heat pump. Even an air-source heat pump can have an efficiency rating of around 250-300%, meaning you get three times the heat out that you would with resistive heating. Ground-source pumps can have efficiencies of close to 400%, but they're not usually suitable for cities. With those kind of efficiency ratings, it is actually more efficient to burn gas in a gas-fired power station to generate electricity to power the heat pump than it is to burn the gas for heat at home.
The government will develop information systems to allow it to launch denial-of-service attacks and to spy on chosen targets, said the official. "We will have a whole range of offensive capabilities, including distributed denial-of-service," said the official. "DDoS is not a first response -- we definitely need graduated responses."
You might have thought it would be better to fund development of mechanisms to prevent or mitigate DDoS attacks, rather than rely on using them. The bad guys will always be able to command more bots than any legal response could.
It just occurred to me that the 10kW number may be for their multi-plate ramp, which has nine ramps close together. But those look like they're only an inch or so high. So multiply by nine, but divide by three. This would still require two crossings per second, so even that doesn't work.
The company that makes this system has some videos that explain how it works.
From their FAQ:
Q1. Doesn't the ramp just steal pennies from our petrol tanks?
A1. The ramp is designed to be situated in parts of the roadway where vehicles are having to slow down, for example on downhill gradients, when approaching traffic lights or roundabouts as well as replacing sleeping policemen and traditional traffic calming measures.
In the these situations, the kinetic energy of the car is being dissipated into heat (i.e. through the braking system) anyway; the ramp at this point scavenges a degree of kinetic energy as the car passes over it, but this is far less than is lost through other mechanisms.
Seems to me like it probably works if it's deployed in the right place. So the idea seems OK.
But what about the numbers? The website claims it can generate 5-10kW. Looks like at least one of the plates moves about three inches (7.5cm). So, lets use their numbers:
10kWh = 36MJ. Taking your 18.1kN force from your 2 ton car, that requires a distance of about 2km. 2km / 7.5cm = 26700 crossings in that hour. Thats 7 per second. No, still doesn't add up.
Best you could reasonably hope for is a car every two seconds. That would give a distance of 7.5cm * 1800 = 135m in an hour. Your 2-tonne car falling 135m would generate 2.4MJ in an hour, so that's about 670W average. And that's assuming 100% efficiency. Likely this thing can power a streetlight or two.
But is it cost effective? Lets say it operates at that rate for 10 hours a day (pretty optimistic for a car park, but maybe on a busy road). 670W gives 6.7kWh per day, or 2400kWh per year. Electricity costs maybe 7p/kWh, so that's GBP171 (or $270). No, this doesn't seem cost effective anywhere where you can get mains electricity.
If the plates are positioned at the bottom of a downhill exit ramp, they will aid drivers braking, prividing kinetic energy without "stealing" drivers fuel. Somehow, I doubt this is where they will be positioned though:-)
Agreed. Sainburys seems to care about looking green, rather than being green. At their Kingston store the large Sainsburys sign has a smallish wind turbine and a solar panel attached to it. Trouble is the wind turbine is positioned between buildings, so it never gets a clear airflow, and the solar panel is positioned facing East. East? What were they thinking?
I had the same problem with the extremely bright LED on my Toshiba 1TB external disk. At least this problem is easily cured with black electrical tape.
Well, a traditional operating system is essentially an abstraction layer between user applications and the hardware. It manages storage, performs task scheduling, handles I/O in a standard way for applications, maintainss user sessions, provides isolation between applications for security, and provides programming APIs that expose all this functionality in a unified way.
A browser is becoming similar, what with browser plugins (PDF, Flash) and Javascript. It does limited storage management (mostly cookies, cache and bookmarks), performs internal scheduling, handles some I/O (display, mouse, etc) in a standard way for "browser applications", maintains separate sessions (tabs), provides (limited) isolation between tabs (could be better though), and provides programming APIs that expose much of this functionality in a unified way (especially for javascript).
It's not completely there as an OS, but it's certainly got many of the properties of an OS.
Alpha is confused by my hometown of Kingston upon Thames. One of its suggestions is to enter a town, so I did. It got confused, presumably because it was trying to read some other meaning into "upon". It then suggested searching for Kingston, or for Kingston Thames. The former found plenty of Kingstons, but not mine (and it's not a small place). The latter told me how far it was from Kington, Jamaica to Thames, New Zealand.
Hope they're better than my current dimmable CFLs from DigiFlux. They don't dim all that low before switching off, but most annoying is that when you switch them on, they take a second or two to start. And I don't mean that like most CFLs they come on dimly at first - they just produce no light at all then suddently switch on properly. In a dimmer this is really annoying, because when you press the dimmer and it doesn't come on, you don't know if before you pressed it it was dimmed down too far (and now you've turned it off) or if it's just not come on yet (and you need to wait). You tend to press the dimmer again right at the moment when it finally turns on, and switch it off again, curse, and repeat the process. Mine are now sitting in a drawer, unused, and I've gone back to conventional bulbs for dimming.
Running a web spider to add noise to your traffic signature.
Anonymous remailers.
Most of these have been tools for privacy freaks and people with something to hide. Running them is enough to raise suspicion. But these kind of data retension measures are much more likely to force such tools to become mainstream. This could backfire on law enforcement and security forces in ways they really don't want.
I think you're probably right that there are failure modes for Ares that are not survivable. SRBs don't tend to undergo catastrophic explosions, but I could certainly guess that a failure of the nozzle gimballing mechanism could be really bad as you can't shut it off while you're cartwheeling. If you're near to the ground, you'd have to be pretty quick on the abort to release the Orion before the cartwheeling booster caused the abort to be downwards. And a gimbal locked hard over at high supersonic speeds would certainly make for an exciting ride. These problems affect any launch vehicle though; when it comes down to it, this is rocket science. And it's never going to be 100% safe. But riding a booster you can't shut off seems to be asking for trouble.
Personally, I think the Ares SRB booster is a mistake. Probably not as big a mistake as Shuttle was though. Don't get me wrong - the shuttle is an amazing piece of technology. But it is fundamentally flawed.
Yes, I know about the theory of these. But the reality is they're really only useful for a non-catastrophic main engine failure. An SRB failure is fatal, as we unfortunately found out. And an uncontained main engine failure is fatal. Main engine pogo is probably fatal (you can't shut down all the main engines while the SRBs are running. If you have a single main engine shut down safely, then one of these might be useful. Or if more than one main engine shuts down safely after SRB separation, then one of these might be possible. But in the first two minutes of flight, when everything is most stressed, you're really out of luck. So realistically, if you compare the failure modes, the shuttle just has so few options for in-flight abort.
I agree that overall safety can only be assessed based on a large enough statistical sample, and we don't have that. But there are several known failure modes of the Shuttle that Apollo and Orion either don't have, or have backup safety systems that the Shuttle doesn't have:
A launchpad (and post launch) escape system that can pull the re-entry vehicle clear of an exploding launch vehicle.
The potential to abort a mission after launch before reaching orbit.
Re-entry heat shield is protected from impacts from ice/foam during launch.
Re-entry vehicle is statically stable during re-entry.
Propellant tanks and fuel for fuel cells stored outside the re-entry vehicle.
All of these seem to argue in favor of Orion being safer than Shuttle.
There are two obvious downsides:
Parachutes have potential failure modes shuttle does not have.
Re-use has the potential to reduce risks (most of the parts have already been test-flown). There's no way to test-fly a non-reusable vehicle.
On balance, I tent to like the KISS approach, so favor the capsule. But you're correct; actual safety comes down to how well all the systems are actually designed and implemented. A simpler approach, poorly implemented, is no safer than a complicated approach implemented well.
I use one of these Cherry mechanical switch compact keyboards. It's not really pretty, but it's done wonders for my typing accuracy. Wish my laptop keyboard was half as good.
It is, and should be, a GPL violation to secretly get a patent license. If TomTom were to cave to MS without getting slapped with a GPL violation, then anyone who uses tomtom's work would be opening themselves up to patent infringement suits.
Actually, at this point, anyone distributing the Linux kernel (whether they are TomTom, RedHat or Linus) has already opened themselves up to patent infringement suits if they are located in the US. Whether or not TomTom license the patents does not change this. The only thing that changes this is if TomTom fight it and the patents are found invalid.
Whether TomTom licenses the patents or not, right now no-one in the US can legally distribute the Linux kernel. There really isn't any doubt that it violates the FAT patents. As we all know about these patents now, none of you located in the US can really agree to the terms of the GPL when you redistribute Linux.
It would be amusing if it came to court, Amazon claimed the ADA required them to do this, and the end result was that publishers were banned from printing paper books are inaccessible to the blind. Stranger things have happened in court.
The Authors Guild today filed suit under the DMCA against the New York Public Library for allowing readers to shine light onto the pages of books. "The electric lights in the public reading room permit the words printed on the page to be copied onto the retina of the library's readers", said Paul Aiken, executive director of the Authors Guild. "We equipped all our books with covers as a way to prevent just this sort of illegal copying. The electric lights are a way to circumvent our copy protection mechanism and therefore are illegal under the DMCA."
Rumor has it that if they are successful, the Authors Guild will next file suit against God for providing a source of light outside in daytime.
Maybe the best way would be to keep using the machine? Not every day, but boot it up once a year for the class history project. Provide a list of things that need to be done each year as preventative maintenance: copy from one set of memory sticks to another, or whatever. And what to do when a read fails - you have redundant copies, but you need to make sure to regain that redundancy once the first failure occurs. And you need instructions to take out the motherboard battery afterwards and to replace it every 5th year. Provide a logbook so people can tick off these things. This way you not only preserve the hardware, but preserve a little of the knowledge about the machine itself and how to use it.
I doubt you can get disks or CDs to last 50 years, but your best bet might well be to ensure they're spun up occasionally. And for disks, the surface will need to be read and re-written.
Any hardware failures will be easier to repair the earlier they are detected. If something fails 15 years from now, you have some hope of finding a spare on Ebay or whatever the equivalent is then. If you only detect it in 50 years, the task is much harder.
I might also think about replacing the stock power supply with something simpler and containing no electrolytic capacitors (no soft-power settings - just on/off 5V and 12V). With some built-in self-test before you throw the switch to send power to the PC itself. Finding a simple replacement power supply 50 years from now should be simple if you've got the specs written down. Finding a replacement motherboard because you just sent overvoltage to it would be a lot harder.
Slashdot Mirror
Basically the underground temperature will roughly be the mean year-round overground temperature, so long as you're not in a geothermally active area. That's where your 10-13C number comes from. But that only applies if you don't inject more heat than the ground can conduct away. The year-round mean temperature in London is probably around 10C. But the temperature in the London Underground in mid summer on the deep lines is more like 30C. Too much heat generated by the trains and people for the rock/clay to conduct away. A data centre generates for more heat in a small space than an underground railway does.
Guess I'll have to watch the YouTube versions now.
I think the reasoning was probably that they were in England when they published for a British audience. The act of publishing was done from England (a server itself isn't capable of independent action, at least not in the eyes of the law).
If they'd flown to the US, written the materials there, published them from there, then returned home, I think they would probably have not have been able to be convicted, even if the intended audience was British.
But I'm not a lawyer, so don't take this as legal advice.
If I recall correctly, Native Client is x86 specific because it uses segmentation to go fast. I believe you should be able to do the same thing with paging on ARM, but I'd expect there to be some performance penalty.
I wonder if they have Google Native Client in mind when they say they're going to re-engineer security from the ground up? Very cool technology.
You get 70% more heat for your gas burning it in a power station and using a heat pump than you would burning it at home (and that assumes the home furnace is 100% efficient, so in practice you do even better than this).
No, electricity using a heat pump is the best way to heat, short of geothermal or solar thermal.
There's a much better reason why you don't want to use lightbulbs to generate heat, which is that resistive heating makes very little sense. Either use gas, where you get a good efficiency by burning it at the point you want heat, or even better, use a heat pump. Even an air-source heat pump can have an efficiency rating of around 250-300%, meaning you get three times the heat out that you would with resistive heating. Ground-source pumps can have efficiencies of close to 400%, but they're not usually suitable for cities. With those kind of efficiency ratings, it is actually more efficient to burn gas in a gas-fired power station to generate electricity to power the heat pump than it is to burn the gas for heat at home.
You might have thought it would be better to fund development of mechanisms to prevent or mitigate DDoS attacks, rather than rely on using them. The bad guys will always be able to command more bots than any legal response could.
It just occurred to me that the 10kW number may be for their multi-plate ramp, which has nine ramps close together. But those look like they're only an inch or so high. So multiply by nine, but divide by three. This would still require two crossings per second, so even that doesn't work.
From their FAQ:
Q1. Doesn't the ramp just steal pennies from our petrol tanks?
A1. The ramp is designed to be situated in parts of the roadway where vehicles are having to slow down, for example on downhill gradients, when approaching traffic lights or roundabouts as well as replacing sleeping policemen and traditional traffic calming measures. In the these situations, the kinetic energy of the car is being dissipated into heat (i.e. through the braking system) anyway; the ramp at this point scavenges a degree of kinetic energy as the car passes over it, but this is far less than is lost through other mechanisms.
Seems to me like it probably works if it's deployed in the right place. So the idea seems OK.
But what about the numbers? The website claims it can generate 5-10kW. Looks like at least one of the plates moves about three inches (7.5cm). So, lets use their numbers:
10kWh = 36MJ. Taking your 18.1kN force from your 2 ton car, that requires a distance of about 2km. 2km / 7.5cm = 26700 crossings in that hour. Thats 7 per second. No, still doesn't add up.
Best you could reasonably hope for is a car every two seconds. That would give a distance of 7.5cm * 1800 = 135m in an hour. Your 2-tonne car falling 135m would generate 2.4MJ in an hour, so that's about 670W average. And that's assuming 100% efficiency. Likely this thing can power a streetlight or two.
But is it cost effective? Lets say it operates at that rate for 10 hours a day (pretty optimistic for a car park, but maybe on a busy road). 670W gives 6.7kWh per day, or 2400kWh per year. Electricity costs maybe 7p/kWh, so that's GBP171 (or $270). No, this doesn't seem cost effective anywhere where you can get mains electricity.
Agreed. Sainburys seems to care about looking green, rather than being green. At their Kingston store the large Sainsburys sign has a smallish wind turbine and a solar panel attached to it. Trouble is the wind turbine is positioned between buildings, so it never gets a clear airflow, and the solar panel is positioned facing East. East? What were they thinking?
I had the same problem with the extremely bright LED on my Toshiba 1TB external disk. At least this problem is easily cured with black electrical tape.
A browser is becoming similar, what with browser plugins (PDF, Flash) and Javascript. It does limited storage management (mostly cookies, cache and bookmarks), performs internal scheduling, handles some I/O (display, mouse, etc) in a standard way for "browser applications", maintains separate sessions (tabs), provides (limited) isolation between tabs (could be better though), and provides programming APIs that expose much of this functionality in a unified way (especially for javascript).
It's not completely there as an OS, but it's certainly got many of the properties of an OS.
Epic fail!
Hope they're better than my current dimmable CFLs from DigiFlux. They don't dim all that low before switching off, but most annoying is that when you switch them on, they take a second or two to start. And I don't mean that like most CFLs they come on dimly at first - they just produce no light at all then suddently switch on properly. In a dimmer this is really annoying, because when you press the dimmer and it doesn't come on, you don't know if before you pressed it it was dimmed down too far (and now you've turned it off) or if it's just not come on yet (and you need to wait). You tend to press the dimmer again right at the moment when it finally turns on, and switch it off again, curse, and repeat the process. Mine are now sitting in a drawer, unused, and I've gone back to conventional bulbs for dimming.
Most of these have been tools for privacy freaks and people with something to hide. Running them is enough to raise suspicion. But these kind of data retension measures are much more likely to force such tools to become mainstream. This could backfire on law enforcement and security forces in ways they really don't want.
Personally, I think the Ares SRB booster is a mistake. Probably not as big a mistake as Shuttle was though. Don't get me wrong - the shuttle is an amazing piece of technology. But it is fundamentally flawed.
Yes, I know about the theory of these. But the reality is they're really only useful for a non-catastrophic main engine failure. An SRB failure is fatal, as we unfortunately found out. And an uncontained main engine failure is fatal. Main engine pogo is probably fatal (you can't shut down all the main engines while the SRBs are running. If you have a single main engine shut down safely, then one of these might be useful. Or if more than one main engine shuts down safely after SRB separation, then one of these might be possible. But in the first two minutes of flight, when everything is most stressed, you're really out of luck. So realistically, if you compare the failure modes, the shuttle just has so few options for in-flight abort.
All of these seem to argue in favor of Orion being safer than Shuttle. There are two obvious downsides:
On balance, I tent to like the KISS approach, so favor the capsule. But you're correct; actual safety comes down to how well all the systems are actually designed and implemented. A simpler approach, poorly implemented, is no safer than a complicated approach implemented well.
I use one of these Cherry mechanical switch compact keyboards. It's not really pretty, but it's done wonders for my typing accuracy. Wish my laptop keyboard was half as good.
Actually, at this point, anyone distributing the Linux kernel (whether they are TomTom, RedHat or Linus) has already opened themselves up to patent infringement suits if they are located in the US. Whether or not TomTom license the patents does not change this. The only thing that changes this is if TomTom fight it and the patents are found invalid.
Whether TomTom licenses the patents or not, right now no-one in the US can legally distribute the Linux kernel. There really isn't any doubt that it violates the FAT patents. As we all know about these patents now, none of you located in the US can really agree to the terms of the GPL when you redistribute Linux.
It would be amusing if it came to court, Amazon claimed the ADA required them to do this, and the end result was that publishers were banned from printing paper books are inaccessible to the blind. Stranger things have happened in court.
Rumor has it that if they are successful, the Authors Guild will next file suit against God for providing a source of light outside in daytime.
Well, for what it's worth, this crash is listed here:
http://aviation-safety.net/database/record.php?id=19880626-0
Maybe the best way would be to keep using the machine? Not every day, but boot it up once a year for the class history project. Provide a list of things that need to be done each year as preventative maintenance: copy from one set of memory sticks to another, or whatever. And what to do when a read fails - you have redundant copies, but you need to make sure to regain that redundancy once the first failure occurs. And you need instructions to take out the motherboard battery afterwards and to replace it every 5th year. Provide a logbook so people can tick off these things. This way you not only preserve the hardware, but preserve a little of the knowledge about the machine itself and how to use it.
I doubt you can get disks or CDs to last 50 years, but your best bet might well be to ensure they're spun up occasionally. And for disks, the surface will need to be read and re-written.
Any hardware failures will be easier to repair the earlier they are detected. If something fails 15 years from now, you have some hope of finding a spare on Ebay or whatever the equivalent is then. If you only detect it in 50 years, the task is much harder.
I might also think about replacing the stock power supply with something simpler and containing no electrolytic capacitors (no soft-power settings - just on/off 5V and 12V). With some built-in self-test before you throw the switch to send power to the PC itself. Finding a simple replacement power supply 50 years from now should be simple if you've got the specs written down. Finding a replacement motherboard because you just sent overvoltage to it would be a lot harder.