Well, let me try. Imagine you're chasing a car (the Gameboy Color), and you want to know what's in the trunk (the ROM). The driver (the CPU) isn't talking. However, you've got a remote control button that can jam on the brakes (stop the clock). This by itself doesn't let you see what's in the trunk. But you find that by hitting the driver with a large blunt object (shorting the 3.3V to ground), you can daze him (randomize the registers) and eventually get him to do what you want, like listen to your request to open the trunk (run your code to dump the ROM).
It's a pretty old technique, but sometimes it permanently disables the driver and it often doesn't give you what you want. The devil is in the details.
Instead of figuring out better ways to cool hot datacenters, I think a better investment would be to figure out how to make the data centers run cooler in the first place.
Each new generation of semiconductor technology typically allows 2 ways to improve: faster computing for same power, or less power for same computing. You'd think that at some point, the second choice makes more sense than the first.
Of course, there is the problem of who can make these improvements. This is limited to the chip companies, whereas just about anyone who knows how to use a screwdriver can work on the cooling problem.
I solved my own version of the problem by switching to notebook computers instead of desktops. I get disgusted every time I look at the power requirements of modern desktops: they burn over 100 watts when sitting idle. That's just stupid. Most notebooks don't use that much even running flat-out.
I wonder how many data center boxes are burning 100 watts just sitting idle?
Sorry, but you're mistaken. It's "politician = criminal", not "politician = smart criminal". Politicians, just like criminals, can be smart and dumb. You only hear about the dumb ones. I'm not sure if the smart ones exist. (We never hear about them.)
I'd like to agree that high school is too late. This should be taught early on in elementary school.
I think the right approach is to introduce kids to the "right" way to type, and provide them the opportunity to learn it, rather than "forcing" it upon them.
Similarly, young kids should also learn what a computer is, the difference between hardware and software, what the internet is, and why you need to be very careful about how you use both the computer and the internet.
These are very basic subjects that are relevant and important for getting along in society today. Teaching kids fundamentals about the tools of modern life is not the same as forcing them all to become touch typists and computer scientists.
Come on; the publishers think they've found a new way to print money, and you're asking them not to do it? Might as well ask them to donate their profits to the Free Software Foundation.
That's what I'm thinking. Perhaps it's the level of _excitement_ that helps recall, and in this particular instance, the excitement level was increased by requiring violence. But that doesn't mean there aren't other, better ways to add excitement. Violence just happens to be the easiest (and basest).
And you live in an area of North America with the cheapest electricity (ie, you are an outlier).
Seattle charges $0.04 - $0.08 per kWh (cheaper rate for first 10-16 kWh per day, depending on season). In New York City, electricity costs around $0.20 per kWh (they say it's $0.10 for the actual juice plus $0.10 for delivery of said juice). That's a 5x difference.
That was a fantastic story ("Reason", by Isaac Asimov). I loved that the robot developed its own view of the universe that was perfectly valid and consistent with all its observations, and yet completely different from what the humans were telling it. It says a lot about beliefs, truth, and religion. I recommend it for everyone.
Every (?) smart phone has (at least) 2 CPUs: the Applications Processor and the BaseBand unit. The AP runs all your fancy UI + user software, while the BBU just deals with phone calls (well, interfacing with the cell phone network, at any rate).
I presume that BBUs are still very proprietary and locked down, or else cell phone companies would (rightly so) freak out over the possibility of network hacking?
As near as I can tell, those specs tell me that the device _won't_ work on AT&T's 3G network. T-Mobile uses 1700/2100 Mhz (paired together) for its 3G/WCDMA network. AT&T uses 850/1900 Mhz (independently) for its 3G/WCDMA network.
I don't understand exactly how AT&T uses 850/1900 Mhz for both 2G/GSM and 3G/WCDMA. Perhaps they use distinct subblocks of each band for 2G vs. 3G?
The cornerstone of any good relationship is communication. This has lots of implications:
1. Know when to say something. Don't let stuff that bugs you build up until you say something nasty. Set aside time to talk that works for both of you. 2. Know when not to say something. When your spouse offers criticism, this is not the time to bring up your own. Deal with your spouse's first. 3. Know how to say something. It's all in the tone of voice. Understand that your "normal" tone of voice might sound cutting sometimes, and try to soften it as the situation demands. 4. Know how to listen. It's more than just nodding your head and saying "uh-huh". Ask relevant questions. 5. Know when to really listen. Sometimes you can keep doing what you're doing; other times your spouse needs your full attention. 6. Remember what you talked about. Make notes. Make a to-do list. Put it in your calendar. Forgetting is bad. 7. Express yourself. Say what you want (otherwise you'll never get it). Don't be a stranger. This can be a challenge for introverted people, but the more you do, the easier it gets.
Some of these will be hard to do, simply because you're habituated to do otherwise. Habits are really hard to change. You'll need to find a way to constantly remind yourself of the what & why to change. Perhaps make your spouse a partner, wherein you both help each other change certain habits. It ought to be easier that way.
The "average" AC adapter that you've described is old-school (linear power supply).
Just about every AC adapter I've seen recently is a switching power supply. The steps are somewhat different: -First rectify the AC into (high voltage) DC. -Feed it through a "chopper" that outputs a high-frequency PWM squarish wave. -Feed that into a (small) transformer. -Rectify the transformer output. -Monitor the output and provide feedback for the chopper PWM duty cycle.
Since all the hard parts are integrated into an IC or two, these are actually cheaper to build than the linear power supply. The high-frequency switching allows a smaller transformer to be used, and the switching regulation is more efficient than using a linear regulator.
But what happens when the power is off? Ahh, that's where it gets fuzzy for me. One of the complications for switching power supplies is how you get them started running. That seems to be a whole art in itself. Perhaps another Slashdotter can enlighten us.
You know those ties that most all cables come with already to keep them bundled up?
I never throw them away. Rather, I use them to keep each cable at its proper length.
Lots of cable clutter comes from having excess cable. Simply keeping the excess tied up neatly helps a lot. If you don't know how much cable you'll need at first, just wrap the wire tie around the end of the cable. It'll be there once you figure out how long you need.
The next idea is to actually route the cables to avoid weaving & tangling. Sure, it's easier just to plug in one end and then just throw the cable over everything and plug in the other end, but you'll save yourself time later if you actually think this out and run the cable over things that move less frequently and under things that move more frequently and generally along with other cables going to the same place.
When you have several cables going to the same place, and you're not likely to move them in a while, then you should also bundle them together, again using wire ties (or, if you prefer velcro ties, plastic cable ties, wire loom, shoelaces, etc.).
Something else that helps is cable hold downs attached to the furniture. You can get sticky-back plastic cable-tie anchors (or use one of the 3M products you find everywhere). Stick them behind & under furniture so that you can hold cables off the floor and near the appliances they are going to. Again, you can use wire ties to attach the bundles to the anchors. (Of course if you have wire rack furniture, you can tie bundles directly to the rack).
Cable management takes time & thought, but done well it looks nice and saves time later.
Satellites that broadcast to the US don't also broadcast to the other side of the planet, so they'd have little incentive to work on hacking those signals. (Not to mention issues regarding western porn.)
Also, China is dependent on foreign trade, and those darn foreigners keep bringing those IP issues back to the trade table.
And even if they pirated everything piratable, that's still only a small dent in western economies.
It has a couple of benefits over the EVM: 1) it has a base that your hand can rest on. 2) the main buttons are thumb-activated, which is good if you were getting pain from finger movement.
The main disadvantage is the lack of a scroll wheel, but I've gotten so used to button scrolling that I don't find this to be a problem.
A mouse is a very personal thing, and a new one typically takes some getting used to. Some people will like it, others will not.
p.s. They have a wireless version too now, apparently.
Another way to abuse: mount your own GPS "satellites" on the roof of your car. They'll transmit a stronger signal than the ones in space. And since they're fixed to your car, you aren't going anywhere, relatively speaking.
I think such things exist already. They are called pseudolites.
It sounds like in both cases, the engineers designed multiple redundant systems that all failed simultaneously due to unexpected sensor input. In both cases, the result was catastrophic failure.
Yes, as you point out, the case of the airplane may have come out differently if pilots could have overridden the automatic systems.
Either way, it points out how "redundancy" can be pointless if a single (small, non-inherently fatal) failure condition takes everything out.
Slashdot Mirror
Well, let me try. Imagine you're chasing a car (the Gameboy Color), and you want to know what's in the trunk (the ROM). The driver (the CPU) isn't talking. However, you've got a remote control button that can jam on the brakes (stop the clock). This by itself doesn't let you see what's in the trunk. But you find that by hitting the driver with a large blunt object (shorting the 3.3V to ground), you can daze him (randomize the registers) and eventually get him to do what you want, like listen to your request to open the trunk (run your code to dump the ROM).
It's a pretty old technique, but sometimes it permanently disables the driver and it often doesn't give you what you want. The devil is in the details.
It is clickable. Haven't you installed the Linkification plugin?
Here: https://addons.mozilla.org/en-US/firefox/addon/190
Instead of figuring out better ways to cool hot datacenters, I think a better investment would be to figure out how to make the data centers run cooler in the first place.
Each new generation of semiconductor technology typically allows 2 ways to improve: faster computing for same power, or less power for same computing.
You'd think that at some point, the second choice makes more sense than the first.
Of course, there is the problem of who can make these improvements. This is limited to the chip companies, whereas just about anyone who knows how to use a screwdriver can work on the cooling problem.
I solved my own version of the problem by switching to notebook computers instead of desktops. I get disgusted every time I look at the power requirements of modern desktops: they burn over 100 watts when sitting idle. That's just stupid. Most notebooks don't use that much even running flat-out.
I wonder how many data center boxes are burning 100 watts just sitting idle?
> Get rid of kids and pets for a couple days
At first I thought you were talking about another useful side-effect of the boric acid.
Sorry, but you're mistaken. It's "politician = criminal", not "politician = smart criminal".
Politicians, just like criminals, can be smart and dumb.
You only hear about the dumb ones. I'm not sure if the smart ones exist.
(We never hear about them.)
That's exactly what I thought. Only problem is that most of that crap is advertising, which is presumably what brings in the money.
I can hear the complaints already: Google is providing yet another way to cut off our revenue stream!
I just tend to avoid news sites that don't present me with a list of summaries I can view before deciding to hit the article itself.
I'd like to agree that high school is too late. This should be taught early on in elementary school.
I think the right approach is to introduce kids to the "right" way to type, and provide them the opportunity to learn it, rather than "forcing" it upon them.
Similarly, young kids should also learn what a computer is, the difference between hardware and software, what the internet is, and why you need to be very careful about how you use both the computer and the internet.
These are very basic subjects that are relevant and important for getting along in society today. Teaching kids fundamentals about the tools of modern life is not the same as forcing them all to become touch typists and computer scientists.
Come on; the publishers think they've found a new way to print money, and you're asking them not to do it? Might as well ask them to donate their profits to the Free Software Foundation.
That's what I'm thinking. Perhaps it's the level of _excitement_ that helps recall, and in this particular instance, the excitement level was increased by requiring violence. But that doesn't mean there aren't other, better ways to add excitement. Violence just happens to be the easiest (and basest).
Perhaps they just haven't had time to rework their Linux HAL for the new hardware?
Have they said "No, Linux will never come to the PS3 Slim"?
You don't need space capability to hack satellites. You just need a dish and a computer.
I hope the Japanese invest in some super-heavy duty multi-layered security.
And you live in an area of North America with the cheapest electricity (ie, you are an outlier).
Seattle charges $0.04 - $0.08 per kWh (cheaper rate for first 10-16 kWh per day, depending on season).
In New York City, electricity costs around $0.20 per kWh (they say it's $0.10 for the actual juice plus $0.10 for delivery of said juice).
That's a 5x difference.
Near as I can tell, the rate in Tokyo is around $0.15-$0.20 per kWh (from a quick glance at http://www.tepco.co.jp/en/service/custom/guide/guide04-e.html ).
That was a fantastic story ("Reason", by Isaac Asimov). I loved that the robot developed its own view of the universe that was perfectly valid and consistent with all its observations, and yet completely different from what the humans were telling it. It says a lot about beliefs, truth, and religion. I recommend it for everyone.
Every (?) smart phone has (at least) 2 CPUs: the Applications Processor and the BaseBand unit.
The AP runs all your fancy UI + user software, while the BBU just deals with phone calls
(well, interfacing with the cell phone network, at any rate).
I presume that BBUs are still very proprietary and locked down, or else cell phone companies
would (rightly so) freak out over the possibility of network hacking?
As near as I can tell, those specs tell me that the device _won't_ work on AT&T's 3G network.
T-Mobile uses 1700/2100 Mhz (paired together) for its 3G/WCDMA network.
AT&T uses 850/1900 Mhz (independently) for its 3G/WCDMA network.
I don't understand exactly how AT&T uses 850/1900 Mhz for both 2G/GSM and 3G/WCDMA.
Perhaps they use distinct subblocks of each band for 2G vs. 3G?
Well, to be complete, you need to run Windows XP and see how the battery life compares. Otherwise you've only got 1/2 a data point.
Primesense... Isn't that the company MS bought for the Natal tech?
The cornerstone of any good relationship is communication. This has lots of implications:
1. Know when to say something. Don't let stuff that bugs you build up until you say something nasty. Set aside time to talk that works for both of you.
2. Know when not to say something. When your spouse offers criticism, this is not the time to bring up your own. Deal with your spouse's first.
3. Know how to say something. It's all in the tone of voice. Understand that your "normal" tone of voice might sound cutting sometimes, and try to soften it as the situation demands.
4. Know how to listen. It's more than just nodding your head and saying "uh-huh". Ask relevant questions.
5. Know when to really listen. Sometimes you can keep doing what you're doing; other times your spouse needs your full attention.
6. Remember what you talked about. Make notes. Make a to-do list. Put it in your calendar. Forgetting is bad.
7. Express yourself. Say what you want (otherwise you'll never get it). Don't be a stranger. This can be a challenge for introverted people, but the more you do, the easier it gets.
Some of these will be hard to do, simply because you're habituated to do otherwise. Habits are really hard to change. You'll need to find a way to constantly remind yourself of the what & why to change. Perhaps make your spouse a partner, wherein you both help each other change certain habits. It ought to be easier that way.
The "average" AC adapter that you've described is old-school (linear power supply).
Just about every AC adapter I've seen recently is a switching power supply. The steps are somewhat different:
-First rectify the AC into (high voltage) DC.
-Feed it through a "chopper" that outputs a high-frequency PWM squarish wave.
-Feed that into a (small) transformer.
-Rectify the transformer output.
-Monitor the output and provide feedback for the chopper PWM duty cycle.
Since all the hard parts are integrated into an IC or two, these are actually cheaper to build than the linear power supply. The high-frequency switching allows a smaller transformer to be used, and the switching regulation is more efficient than using a linear regulator.
But what happens when the power is off? Ahh, that's where it gets fuzzy for me. One of the complications for switching power supplies is how you get them started running. That seems to be a whole art in itself. Perhaps another Slashdotter can enlighten us.
You know those ties that most all cables come with already to keep them bundled up?
I never throw them away. Rather, I use them to keep each cable at its proper length.
Lots of cable clutter comes from having excess cable. Simply keeping the excess tied up neatly helps a lot.
If you don't know how much cable you'll need at first, just wrap the wire tie around the end of the cable.
It'll be there once you figure out how long you need.
The next idea is to actually route the cables to avoid weaving & tangling. Sure, it's easier just to plug in one end and then just throw the cable over everything and plug in the other end, but you'll save yourself time later if you actually think this out and run the cable over things that move less frequently and under things that move more frequently and generally along with other cables going to the same place.
When you have several cables going to the same place, and you're not likely to move them in a while, then you should also bundle them together, again using wire ties (or, if you prefer velcro ties, plastic cable ties, wire loom, shoelaces, etc.).
Something else that helps is cable hold downs attached to the furniture. You can get sticky-back plastic cable-tie anchors (or use one of the 3M products you find everywhere). Stick them behind & under furniture so that you can hold cables off the floor and near the appliances they are going to. Again, you can use wire ties to attach the bundles to the anchors. (Of course if you have wire rack furniture, you can tie bundles directly to the rack).
Cable management takes time & thought, but done well it looks nice and saves time later.
Satellites that broadcast to the US don't also broadcast to the other side of the planet, so they'd have little incentive to work on hacking those signals. (Not to mention issues regarding western porn.)
Also, China is dependent on foreign trade, and those darn foreigners keep bringing those IP issues back to the trade table.
And even if they pirated everything piratable, that's still only a small dent in western economies.
I'm rather fond of this one:
http://www.amazon.com/3M-EM500GPL-Ergonomic-Mouse-EM500GPLLARGE/dp/B00008KWWF
It has a couple of benefits over the EVM:
1) it has a base that your hand can rest on.
2) the main buttons are thumb-activated, which is good if you were getting pain from finger movement.
The main disadvantage is the lack of a scroll wheel, but I've gotten so used to button scrolling that I don't find this to be a problem.
A mouse is a very personal thing, and a new one typically takes some getting used to. Some people will like it, others will not.
p.s. They have a wireless version too now, apparently.
Another way to abuse: mount your own GPS "satellites" on the roof of your car. They'll transmit a stronger signal than the ones in space. And since they're fixed to your car, you aren't going anywhere, relatively speaking.
I think such things exist already. They are called pseudolites.
It sounds like in both cases, the engineers designed multiple redundant systems that all failed simultaneously due to unexpected sensor input. In both cases, the result was catastrophic failure.
Yes, as you point out, the case of the airplane may have come out differently if pilots could have overridden the automatic systems.
Either way, it points out how "redundancy" can be pointless if a single (small, non-inherently fatal) failure condition takes everything out.
http://en.wikipedia.org/wiki/Ariane_5_Flight_501