what are you supposed to do for your refrigerator [...] or your oven
LEDs are absolutely perfect for the job. Much more energy efficient than incandescent, and handle various temperatures, surges, and power cycling far better.
I'll put more faith in this alleged consumer demand when Linux boxes start outselling all other systems by a 2-to-1 margin.
The 2:1 margin was for REQUESTS. Obviously, since they're selling Windows, nobody is going to REQUEST it.
So, no, it's not going to sell 2:1, or even 1:2, and it's stupid to think it should, based on this story.
HOWEVER, this story is misleading, as the "No Extra Software" (on Windows) is divided into two different options, and if vote were combined, would be about 80% of the Linux option. So, it's decidedly not 2:1.
If anyone can generate enough concentrated energy to boost over interstellar distances, there's going to be some way to use it destructively.
There's always a way to use something destructively. Most power sources don't inherently lend themselves to global destructiveness (like nuclear power does) though. The destructive uses could be vastly impractical on a large scale.
You could always pull a rock out of orbit and toss it at a planet, but that's difficult (and slow), and could potentially be intercepted.
Our ease and practical level of (global) destructiveness could well be unique, or at least uncommon.
any civilization that develops energy sources big enough to power interstellar travel also has the ability to blow itself up.
I don't buy it. Certainly that's the way it worked here, but that could just be a fluke.
Certainly, solar offers far more power than fission, but would be far more difficult to effectively weaponize.
Re:Humans can handle more than 1 G
on
Interstellar Ark
·
· Score: 1
The nine G figure is unrealistically high, but there are no reasons to assume you can't have a realitivistic rocket that starts out with six G for a short while and then drops its acceleration off to about two G.
Traveling at a gradually increasing G-forces would be extremely interesting...
How versatile is the human body? If G-Forces only gradually increase about 5% per year, will humans grow to handle 3Gs? Sure, they'll probably be quite short and bulky, while possibly requiring many more calories than normal. But, if that is possible, then the range of inhabitable planets increases tremendously. Giant planets may be human habitable.
DRM-protected legitimate content as easily accessible and convenient as unprotected illegitimate content is to consumers.
An online store can be much easier and more convenient than tracking down music on the current P2P networks. More than enough to make up for the inconvenience of having to enter credit card details, and paying a few cents per song (or per-month).
OK, I'm gonna say that back, but a little more directly. I don't think you know WTF you're talking about either.:)
SInce you seem obsessed with Wikipedia, allow me to point out a small section you seem to be ignoring:
Congestion in the Internet backbone is very difficult to deal with. Fortunately, cheap fiber-optic lines have reduced costs in the Internet backbone. The backbone can thus be provisioned with enough bandwidth to (usually) keep congestion at the periphery.
Frankly, wikipedia is just a terrible source of information. Their portrayal of TCP flow control as perfect, and everything working out just fine, has no basis in reality.
Except the telcos don't want to spend money implementing QOS until they can find a way to recover the cost.
The cost in nominal. Many already have it in-place to throttle-back P2P. Their way to recover costs is their monthly fee. If there's a bandwidth crunch, they'll have a lot of motivation...
But do all the customers of a given ISP look at the same content at the same time?
Yes. The majority of traffic is redundant, within a close time-frame. You wouldn't see caching proxies like squid if that wasn't the case.
How much would it cost to cache all of Flickr and all of YouTube and all of Newgrounds?
Caching the most popular videos and images would make a huge difference. Storage is quite cheap, and a $10,000 box at each ISP would greatly reduce the internet bandwidth used.
But how does the protocol learn that parts of the file are available locally?
Bittorrent uses a single tracker to keep track of all peers with any amount of the file, and will report them all. Gnutella uses both a network search, and each node you download a file from will report all peers it knows of with the same file.
Several features in TCP, such as slow start and exponential backoff, were carefully tuned to make sure this doesn't happen.
You have a very interesting point of view, there.
These things weren't originally designed to be the method of congestion control to begin with. That came later, as a response to implementations of the stack.
The exponential backoff of TCP would only help if a large number of hosts were doing it at the same time. That doesn't happen. Slow start may help slightly, but the way in which TCP goes ever faster, until it fails again, is really the cause of the problem, not the solution to it. It's that sawtooth pattern which causes the most serious, long-term problem to begin with. With something like UDP, you'll still be able to get some traffic through, while TCP will backoff almost into a standstill, but continue to send, and send.
It can still happen if you have a very sudden, severe spike in traffic, but even in that case it recovers in a few minutes.
I get the impression you're thinking of a different scenario than I am. There's nothing sudden here. This is what will happen if the level of traffic (gradually) increases over time, while the backbone does not (and that's not just IMHO). No sudden, surge route failures, etc. are needed to set this off.
When we all used 300 baud modems, was there a "bandwidth shortage"?
Just the opposite. The slower the end-user links, the less chance of backbone congestion.
Indeed, that would be the easy solution, should this happen, but I don't count on any ISPs to volunteer to be the first to violate their contract, and piss off their paying customers, in hopes that everyone else will do the same shortly thereafter, saving the internet for all...
Bit of a prisoners dilemma there, and self-interested corporations don't fare well in such scenarios. After a few months, the government might wake up and step in with such an option.
Except fans of network neutrality tend to rail against how they imagine that telcos would implement QoS.
Net neutrality has always been about giving similar protocols the same priority, no matter the company in question. It has nothing to do with QoS in general.
Pages sent to logged-in users are generated dynamically based on the content of the HTTP session cookie. How can these be cached?
They can't really. However, that's a real minority of even the WWW, and the WWW is a real minority of internet traffic. Dynamically generated web pages are only about a KB or two, and can be far smaller still with compression.
The non-dynamic content is still cached. External JS/CSS, images, flash, videos, etc. will benefit greatly.
And potentially invite the same kinds of legal action leveled against universities for "inducing" copyright infringement by tolerating local SMB or Hotline or DC++ file sharing hubs.
Nope. The MPAA/RIAA can push around universities, but they know not to mess with the large communications companies.
The ISPs face no added liability for MORE of the traffic going over their own lines. Since they are common carriers, that should be no problem at all (schools aren't common carriers).
do there now exist decentralized P2P file sharing protocols that include network topology in the priority calculations?
Well, yes, but let's pretend that there aren't...
If you get 1Mbps over the internet, and 10Mbps locally, the problem just works itself out naturally. If anyone has part of the file locally, you'll download 90% of the file from them, in the time it takes to download the other 10% of the file from any number of other nodes over the internet. And that's in the worst case, with braindead simple protocols. Anti-leech protocols like bittorrent (particularly bittyrant) will exaggerate this effect even more.
all those "high-bandwidth" things they complain about (YouTube) will be too slow to get at properly, and people will give up and go watch TV or something instead.
YouTube, maybe, but certainly not P2P, large file downloads, etc. Once anyone starts feeling the effects of internet congestion, the cascading failure has already started, and can only get worse.
An overwhelmed website isn't the same as an overwhelmed backbone. It doesn't downgrade cleanly, things get very messy.
if it is true that the internet won't scale in the scenarios outlined above, it won't scale only in a specific context: the context of bps hungry applications
ok: so you won't be able to watch the latest youtube laugh video. whoop de friggin doo
you'll still be able to communicate, plain text emails, simple html pages, etc.
Actually, it doesn't work that way. Once ANY network starts nearing capacity, widespread congestion takes hold, and nobody can do anything. It's especially true for the internet, as TCP basically explodes in high packet-loss senarios. What's more, it basically goes exponentially.
When congestion causes 1% higher packet losses, everything starts sending 1% more traffic, that 1% more traffic causes 2% more packet loss, everything starts sending 2% more traffic, that causes 4% more packet loss... TCP congestion is the straw the breaks the camels back. It's perfectly fine, right up until you reach the bandwidth/congestion limit, then TCP explodes, and everything comes suddenly tumbling down to a standstill.
It's really not a difficult problem. It's just that bandwidth has been so cheap, links so fast, and ever expanding, that there's been no motivation until recently.
First: QoS. Edge routers can do it all. Make sure each group, sub group, sub-sub group (etc) gets only an even share of the available bandwidth, then downgrade speeds as needed.
Second: Caching proxies can make a huge difference as well. In this day and age, with incredibly high-capacity hard drives being dirt cheap, it's unbelievable that every ISP doesn't already have caching proxies with dozens of terabytes of data stored.
Third: Multiple speeds. Right now, it's just as fast for me to download an ISO from the other side of the planet, as it is to download it from the data center a few blocks away. ISPs need to change their bandwidth rules, so that traffic over the backbone is limited to whatever speed (eg. 756k) but traffic that doesn't exit their own network is unlimited. It certainly gives huge incentives for people to actively try to use their ISP's caching proxy listed above. It will keep the vast majority of P2P traffic in-network, as you can download 90% of the file from your neighbor more quickly than 10% of the file over the internet. This may just happen naturally, as 802.11 becomes more popular, and it's faster to download most large files from your neighbors than from your ISP's wireless router (ie. the internet).
And finally, but perhaps more importantly, we should finally fix TCP. Using dropped packets to regulate traffic flow is pure nonsense, caused by a historical fluke. Purpose-made congestion notification (like source quench) would result in much higher utilization of existing links, nearly to their maximum. Unlike the current system, communications wouldn't go from perfectly fine, to utterly impossible, when line congestion reaches a certain threshold.
Hell, I like the idea. When throwing bandwidth at these problems no longer works, most people on/. will be in huge demand to work out solutions.
Eventually, yes, computer skills become important, fundamental even. I just worry how they're to be used in class, that's all. I sure hope they aren't going to be expected to replace teachers, and I hope budget-strapped schools favor good staff over 100 dollar laptops.
It's hard to learn how to spell when there's one dictionary every 20 miles...
It's hard to learn to learn arithmetic when your school can't afford to buy more than a handful of math books.
It's hard to learn about a lot of things, when you don't have the tools to do any of it. This computer can take the place of a great many tools that simply non-existent in the 3rd world, not the least of which is books and calculators.
Unlike the US, where "computers" means learning to format Microsoft Word documents, and being otherwise restricted, these computers can perform an infinitely valuable service, both for education in the class, and perhaps more importantly, education outside the class as well.
Nothing I don't already know. There's really no point in explaining it, in detail, when the basic premise is good enough for the topic at hand.
In other words, there may not be much heat energy in space, but there's plenty of light (when near the sun) and there's nothing to freeze you.
They rarely allow space walks in sunlight, for good reason, so you can expect to always be freezing. For this reason, space suits have electric heaters to keep the occupants warm.
A vacuum prevents convection, but the fact of the matter is that everything loses heat faster than the human body can generate it. For an example, take a look at Apollo 13, when they were forced to operate without electrical heating systems for days on end. Compared to the environments humans are used-to, space is certainly freezing cold. Yes, it can also be boiling hot, but very little of it is.
As opposed to the commercial software industry, which you can hardly STOP from innovating...
Give me a break. 99% of software out there was copied from something else, with trivial improvements, be it commercial, or Open Source. Or did you think Microsoft invented the Word Processor, and Spreadsheet?
IMHO, at the end of the day, OSS is innovating much more than the commercial software industry. Copying of ideas goes both ways.
Well, if I only paid $0.00 for it in the first place, I might expect to be asked to return the car or pay a fair price later.
Yes, well a price of $0 for a $12,000 (used) car is overwhelmingly obviously wrong. A price of $2,000 might not be...
The same goes for DVDs. It's not hard to believe Amazon was giving out a $20 loss-leader to try and generate interest, to increase sales.
Even if the cashier says "just go ahead and take it", that doesn't make it right. [...] Taking advantage of a broken automated system isn't any more moral than stealing if you know the price isn't appropriate.
There have been many times I've taken something to check-out, only to have the cashier tell me there's an promotional $20-off coupon (which I didn't know about) and applying it.
I don't think twice about paying 1/10th of what I expected to, and walking out. Chock it up to luck, being in the right place, at the right time.
Neither situation is remotely comparable to a situation where you know exactly how much money (or change) you are supposed to get, but can see you've been given a different amount.
You have to make the space suits out of something... it may as well be something that can recapture energy normally wasted in motion.
There is NO SUCH THING as "wasted energy". It always does SOMETHING. In the freezing vacuum of space, in particular, that heat is very useful in keeping the astronaut warm, meaning they need to use that much less electricity to stay alive.
You have to make the space suits out of something... it may as well be something that can recapture energy normally wasted in motion.
Unless this material can function as well as the material currently in space-suits, while being easier to flex, using it in space suits will mean much more effort to move.
And if it WAS better than existing materials, they would use it, and just not take advantage of its ability to generate electricity, thereby requiring even less effort to move.
You really can't get energy for nothing.
I suspect the space-suit idea was something misunderstood, or taken vastly out of context by whoever reported on this. The OTHER ideas for use of this material are perfectly sensible.
Slashdot Mirror
If power usage in California goes down... They aren't going to decide to use less hydroelectric... Less solar... Less wind... etc.
LEDs are absolutely perfect for the job. Much more energy efficient than incandescent, and handle various temperatures, surges, and power cycling far better.
Well then it's a good thing I use FreeBSD. *ducks*
It says the same thing as it did when I read it the first time.
The 2:1 margin was for REQUESTS. Obviously, since they're selling Windows, nobody is going to REQUEST it.
So, no, it's not going to sell 2:1, or even 1:2, and it's stupid to think it should, based on this story.
HOWEVER, this story is misleading, as the "No Extra Software" (on Windows) is divided into two different options, and if vote were combined, would be about 80% of the Linux option. So, it's decidedly not 2:1.
There's always a way to use something destructively. Most power sources don't inherently lend themselves to global destructiveness (like nuclear power does) though. The destructive uses could be vastly impractical on a large scale.
You could always pull a rock out of orbit and toss it at a planet, but that's difficult (and slow), and could potentially be intercepted.
Our ease and practical level of (global) destructiveness could well be unique, or at least uncommon.
I don't buy it. Certainly that's the way it worked here, but that could just be a fluke.
Certainly, solar offers far more power than fission, but would be far more difficult to effectively weaponize.
Traveling at a gradually increasing G-forces would be extremely interesting...
How versatile is the human body? If G-Forces only gradually increase about 5% per year, will humans grow to handle 3Gs? Sure, they'll probably be quite short and bulky, while possibly requiring many more calories than normal. But, if that is possible, then the range of inhabitable planets increases tremendously. Giant planets may be human habitable.
An online store can be much easier and more convenient than tracking down music on the current P2P networks. More than enough to make up for the inconvenience of having to enter credit card details, and paying a few cents per song (or per-month).
SInce you seem obsessed with Wikipedia, allow me to point out a small section you seem to be ignoring:
Frankly, wikipedia is just a terrible source of information. Their portrayal of TCP flow control as perfect, and everything working out just fine, has no basis in reality.
And while the alarms are waiting a few seconds, the autodialer is immediately reconnecting to stop anyone from dialing out.
The cost in nominal. Many already have it in-place to throttle-back P2P. Their way to recover costs is their monthly fee. If there's a bandwidth crunch, they'll have a lot of motivation...
Yes. The majority of traffic is redundant, within a close time-frame. You wouldn't see caching proxies like squid if that wasn't the case.
Caching the most popular videos and images would make a huge difference. Storage is quite cheap, and a $10,000 box at each ISP would greatly reduce the internet bandwidth used.
Bittorrent uses a single tracker to keep track of all peers with any amount of the file, and will report them all. Gnutella uses both a network search, and each node you download a file from will report all peers it knows of with the same file.
You have a very interesting point of view, there.
These things weren't originally designed to be the method of congestion control to begin with. That came later, as a response to implementations of the stack.
The exponential backoff of TCP would only help if a large number of hosts were doing it at the same time. That doesn't happen. Slow start may help slightly, but the way in which TCP goes ever faster, until it fails again, is really the cause of the problem, not the solution to it. It's that sawtooth pattern which causes the most serious, long-term problem to begin with. With something like UDP, you'll still be able to get some traffic through, while TCP will backoff almost into a standstill, but continue to send, and send.
I get the impression you're thinking of a different scenario than I am. There's nothing sudden here. This is what will happen if the level of traffic (gradually) increases over time, while the backbone does not (and that's not just IMHO). No sudden, surge route failures, etc. are needed to set this off.
Just the opposite. The slower the end-user links, the less chance of backbone congestion.
Indeed, that would be the easy solution, should this happen, but I don't count on any ISPs to volunteer to be the first to violate their contract, and piss off their paying customers, in hopes that everyone else will do the same shortly thereafter, saving the internet for all...
Bit of a prisoners dilemma there, and self-interested corporations don't fare well in such scenarios. After a few months, the government might wake up and step in with such an option.
Net neutrality has always been about giving similar protocols the same priority, no matter the company in question. It has nothing to do with QoS in general.
They can't really. However, that's a real minority of even the WWW, and the WWW is a real minority of internet traffic. Dynamically generated web pages are only about a KB or two, and can be far smaller still with compression.
The non-dynamic content is still cached. External JS/CSS, images, flash, videos, etc. will benefit greatly.
Nope. The MPAA/RIAA can push around universities, but they know not to mess with the large communications companies.
The ISPs face no added liability for MORE of the traffic going over their own lines. Since they are common carriers, that should be no problem at all (schools aren't common carriers).
Well, yes, but let's pretend that there aren't...
If you get 1Mbps over the internet, and 10Mbps locally, the problem just works itself out naturally. If anyone has part of the file locally, you'll download 90% of the file from them, in the time it takes to download the other 10% of the file from any number of other nodes over the internet. And that's in the worst case, with braindead simple protocols. Anti-leech protocols like bittorrent (particularly bittyrant) will exaggerate this effect even more.
Then the terrorists will truly have won...
YouTube, maybe, but certainly not P2P, large file downloads, etc. Once anyone starts feeling the effects of internet congestion, the cascading failure has already started, and can only get worse.
An overwhelmed website isn't the same as an overwhelmed backbone. It doesn't downgrade cleanly, things get very messy.
Actually, it doesn't work that way. Once ANY network starts nearing capacity, widespread congestion takes hold, and nobody can do anything. It's especially true for the internet, as TCP basically explodes in high packet-loss senarios. What's more, it basically goes exponentially.
When congestion causes 1% higher packet losses, everything starts sending 1% more traffic, that 1% more traffic causes 2% more packet loss, everything starts sending 2% more traffic, that causes 4% more packet loss... TCP congestion is the straw the breaks the camels back. It's perfectly fine, right up until you reach the bandwidth/congestion limit, then TCP explodes, and everything comes suddenly tumbling down to a standstill.
It's really not a difficult problem. It's just that bandwidth has been so cheap, links so fast, and ever expanding, that there's been no motivation until recently.
/. will be in huge demand to work out solutions.
First: QoS. Edge routers can do it all. Make sure each group, sub group, sub-sub group (etc) gets only an even share of the available bandwidth, then downgrade speeds as needed.
Second: Caching proxies can make a huge difference as well. In this day and age, with incredibly high-capacity hard drives being dirt cheap, it's unbelievable that every ISP doesn't already have caching proxies with dozens of terabytes of data stored.
Third: Multiple speeds. Right now, it's just as fast for me to download an ISO from the other side of the planet, as it is to download it from the data center a few blocks away. ISPs need to change their bandwidth rules, so that traffic over the backbone is limited to whatever speed (eg. 756k) but traffic that doesn't exit their own network is unlimited. It certainly gives huge incentives for people to actively try to use their ISP's caching proxy listed above. It will keep the vast majority of P2P traffic in-network, as you can download 90% of the file from your neighbor more quickly than 10% of the file over the internet. This may just happen naturally, as 802.11 becomes more popular, and it's faster to download most large files from your neighbors than from your ISP's wireless router (ie. the internet).
And finally, but perhaps more importantly, we should finally fix TCP. Using dropped packets to regulate traffic flow is pure nonsense, caused by a historical fluke. Purpose-made congestion notification (like source quench) would result in much higher utilization of existing links, nearly to their maximum. Unlike the current system, communications wouldn't go from perfectly fine, to utterly impossible, when line congestion reaches a certain threshold.
Hell, I like the idea. When throwing bandwidth at these problems no longer works, most people on
It's hard to learn how to spell when there's one dictionary every 20 miles...
It's hard to learn to learn arithmetic when your school can't afford to buy more than a handful of math books.
It's hard to learn about a lot of things, when you don't have the tools to do any of it. This computer can take the place of a great many tools that simply non-existent in the 3rd world, not the least of which is books and calculators.
Unlike the US, where "computers" means learning to format Microsoft Word documents, and being otherwise restricted, these computers can perform an infinitely valuable service, both for education in the class, and perhaps more importantly, education outside the class as well.
They rarely allow space walks in sunlight, for good reason, so you can expect to always be freezing. For this reason, space suits have electric heaters to keep the occupants warm.
A vacuum prevents convection, but the fact of the matter is that everything loses heat faster than the human body can generate it. For an example, take a look at Apollo 13, when they were forced to operate without electrical heating systems for days on end. Compared to the environments humans are used-to, space is certainly freezing cold. Yes, it can also be boiling hot, but very little of it is.
As opposed to the commercial software industry, which you can hardly STOP from innovating...
Give me a break. 99% of software out there was copied from something else, with trivial improvements, be it commercial, or Open Source. Or did you think Microsoft invented the Word Processor, and Spreadsheet?
IMHO, at the end of the day, OSS is innovating much more than the commercial software industry. Copying of ideas goes both ways.
Yes, well a price of $0 for a $12,000 (used) car is overwhelmingly obviously wrong. A price of $2,000 might not be...
The same goes for DVDs. It's not hard to believe Amazon was giving out a $20 loss-leader to try and generate interest, to increase sales.
There have been many times I've taken something to check-out, only to have the cashier tell me there's an promotional $20-off coupon (which I didn't know about) and applying it.
I don't think twice about paying 1/10th of what I expected to, and walking out. Chock it up to luck, being in the right place, at the right time.
Neither situation is remotely comparable to a situation where you know exactly how much money (or change) you are supposed to get, but can see you've been given a different amount.
There is NO SUCH THING as "wasted energy". It always does SOMETHING. In the freezing vacuum of space, in particular, that heat is very useful in keeping the astronaut warm, meaning they need to use that much less electricity to stay alive.
Unless this material can function as well as the material currently in space-suits, while being easier to flex, using it in space suits will mean much more effort to move.
And if it WAS better than existing materials, they would use it, and just not take advantage of its ability to generate electricity, thereby requiring even less effort to move.
You really can't get energy for nothing.
I suspect the space-suit idea was something misunderstood, or taken vastly out of context by whoever reported on this. The OTHER ideas for use of this material are perfectly sensible.
Yes, but it has to do that faster than the dial-in system can reconnect.