Wrong. OP is correct. All things being equal you can fit the same amount of data in 700-720MHz as in 2.5-2.52GHz. As another poster mentioned the difference is not capacity, but instead reusability.
Maybe you should study how routing protocols work.
For the type of ubiquitous network described, geographic routing should work plenty well, and could easily be implemented. If not there is a large amount of other protocols (AODV, Fish-eye Routing, etc) that have been developed to deal with the scalability problems.
Of course my real concern with this approach is bandwidth. In my opinion there is no way to cram that much data into the available wireless spectrum. It's been proven that the per-user bandwidth-distance ratio of such a network is monotonically decreasing as the number of users increases [1]. The fact is we will always need land-lines, since they have the unique and desirable properties of (1) not interfering with each other, and (2) being able to cross really long distances with relatively small amounts of power.
References:
1. P. Gupta and P. R. Kumar, "The capacity of wireless networks," IEEE
Trans. Inform. Theory, vol. 46, pp. 388-404, Mar. 2000
I'm also an EE, and while this is not my area (it's not yours either!), I think you are simply trying to be over-pedantic.
First of all context is everything. If I say that this line is at 5V, then someone in the power field of EE would think there was really 7.07 V (peak) on it since they alway deal with RMS. Different fields can make different assumptions: in the digital field I can roughly assume that the voltage in my circuits is in 1 of two states (well mostly).
The leakage losses occur because silicon is an imperfect insulator. Even when transistors are 'off' (or switching) some current leaks through to the drain and bulk. This doesn't depend on the amount of current in the transistor doing useful work or even on the switching frequency, but only on the voltage. The actual power loss depends on the layout and operation of each transistor (with really complex interactions among them). There isn't really a simple resistor (in fact most models include 3 or more), but I was trying to give the layman's version. I may have goofed with the formula and I should have written V^2/R (though this is still far from accurate).
The power lost through switching is not reactive power. Reactive power is useful if you are planning a power distribution network, but not so much when you are calculating heat generation (since reactive power doesn't create heat). The switching losses occur because of the way CMOS logic works. When the pair of transistors changes from a 1 to a 0 the charge built up on the source of the NMOS transistor is dumped to ground (and from 0 to 1 with the PMOS dumped to Vcc). This charge is due to (among other things) the capacitance across the transistor, and when it is dumped through the NMOS transistor, all that energy is lost. The formula for energy in a capacitor is 0.5 C V^2, and since this amount of energy is lost every switching cycle, the power lost is 0.5 f C V^2. This is not the complete picture as there are other losses (and some devices shutoff portions of the chip not in use [clock gating, etc]).
Yes, the current will vary with the voltage, but there's really not need to over-specify. As long as you are using silicon processes the parameters are going to be roughly the same (though it would have been nice if they mentioned the fab process or the scale). If you can calculate the power with P=V^2/R and everybody knows R, why bother to provide I? Also since the second power of voltage is in all those equations, halving the voltage means you can more than quadruple the frequency with the same power consumption (okay not really, due to the transistor switching times), so small changes in frequency are insignificant compared to changes in voltage.
The difference is we are talking about semiconductor devices. Losses from these semiconductor devices are primarily due to leakage and switching. As long as we're still using silicon, leakage will be roughly 0.5 V^2/R, no matter how much current you pump through the transistors. Switching losses occur in when logic changes from 1's to 0's due to the capacitance of the transistors. The power lost here is roughly 0.5 f C V^2, where f is the switching frequency and C is the capacitance (material dependent). The V^2's means that reducing the voltage has a significant impact on losses. If we note that R and C are completely determined by the material (silicon) and the fabrication process, we can see that as long as the frequency is held constant, the voltage is a reasonable metric for comparing power consumption in silicon devices.
Of course this analysis is purely approximate since there are a lot of there things going in the devices. And I'm assuming complete capacitive discharge (independent of switching frequency), and didn't consider the changes in refresh rate to this DRAM device. But suffice it to say voltage is still a pretty good metric for comparison (until you actually build the thing and test it).
The ideal free market has no collusion or barriers to entry. Companies compete merely on merit (which is globally known to all consumers). Price is driven only by supply and demand. All goods are commodities, so there cannot be monopolies. Under this model the "invisible hand" WILL correct the market. This is really good, and I am not aware of another market model that can make this guarantee.
Yes net neutrality is against "laissez faire" capitalism. But this is necessary because laissez fair capitalism doesn't work in the real world. While traditional economic theory has shown that this sort of market provides the most efficient allocation of resources, the analysis makes several simplifying assumptions that limit its applicability. I could make a long list of these failed assumptions, but suffice it to say pure laissez faire capitalism is neither desirable nor workable.
Having said that, the laissez faire model can still be modified. If the government uses regulations to force the market to meet the assumptions of laissez faire capitalism, then capitalism can exist under those regulations, and will be at worst a local optimum. If you wish to call this socialism you may, but ideally the government should only interfere to ensure the market is competitive and only intervene when capitalism breaks down. This is the model currently employed by most modern countries, whether they admit it or not (ex. consumers have imperfect knowledge, so labeling regulations were added).
Ideally QoS would be combined with resource reservation schemes such as RSVP. Under these schemes the application requests the ability to send data at rate X with maximum latency Y/priority Z. Under these schemes the extra VoIP connections would only be allowed if there was already sufficient bandwidth to handle the traffic with low latency. In contrast, additional BULK traffic will generally be unconditionally accepted (the only limiting factor is the amount of buffer space available in intermediate routers).
This will enable the ISP to offer tiers of service: you can pay for 1 VoIP channel and 128kB/s BULK, or 3 video conferencing channels, 10 VoIP channels and 1MB/s BULK. If you want IP-TV, you will need to make sure you have enough video streams on your plan. People here are not likely to complain about that because (1) it is still net-neutral (2) it eliminates the "how unlimited" problem with ISPs now.
As for WiFi ISPs, I had one last year. Great service and price, $25/month for fast high quality internet access. Never had any problem with them. Unfortunately they went out of business, and I had the alternative of paying $60/month for cable internet or $60/month for ADSL. There are poorly supported ISPs everywhere, and I find it hard to believe your experience with one ISP's oversaturated infrastructure is evidence that shared infrastructure cannot work.
Net-neutrality is not anywhere near communism. (Under communist internets, streams resources reserve YOU!). Under communism there is centralized planning and management of the economy. This is actually a pretty close description of the Non-Net Neutral internet: the telco's decide the priority of packets of different types to different destinations. They decide whether or not you can set up your own IP-TV broadcasting company or make the next youtube. Of course since this is a company deciding this you might describe it as fascism.
Net-neutrality is capitalism. Collusion between ISPs and other companies and other non-competitive practices becomes more difficult. This forces the real world market to be a closer approximation of the ideal "free market".
Why not have computers keep track of TAI time and then translate to sidereal time for the UI. For most cases this would be a simple switch since they don't properly handle leap seconds anyway. Then, the addition or subtraction of leap seconds could be added to the UI.
That's why you don't make VoIP work "better" than bittorrent, you make it work "different" than bittorrent. With QoS your VoIP (Real-time streams) would get say, a fixed 9kb/s or whatever of "Guaranteed Low Latency" (TM). And your bittorrent (BULK Traffic) would get what ever is left over, but makes no guarantee of when your packets will arrive. The point is that if you make bittorrent act like VoIP, it will be limited to the real-time rate which should be slower than the BULK rate.
Of course the gotcha there is the "should be". If the telco's are cheap and don't upgrade, then even QoS can't stop the brownouts. But then again if the telco's don't upgrade, there'll be brownouts anyway...
It's not a bribe--there is no individual that is receiving the money. Instead, Microsoft would spend $$ marketing the TCS's classmate PC in Africa (if they choose windows).
Things like this occur in businesses every day in America. This is neither illegal or even shady. As usual the slashdot summary is slightly biased.
The point is there is a difference between interactive traffic and bulk traffic. While your 4.5 GB Linux ISO download doesn't care so much that it suffers from 250ms of latency while waiting in low priority queue for an opening, someone's VOIP or streaming video would (buffering....) be (buffering...) un (buffering...) useable with that much delay.
Ideally QoS would adjust latency based on application while keeping bandwidth allocations fair. So instead of simply ``letting your connection suffer'' QoS really allows them to use flexible limits on bandwidth and latency. Without QoS the only way to guarantee some level of service to to sompletely separate the individual users. Segregating everyone's pipes to guarantees that no one interferes with anyone else (and wastes alot of the bandwidth allocated to the email checkers). With QoS, they can let your bulk transfer borrow some bandwidth from some one who isn't using all of their's, and when they do send a VOIP packet you "return the favor" by letting them ahead in line. You benefit from the extra available bandwidth, and they benefit from the reduced latency.
Voted to allow bigoted Alabama judge to post Ten Commandments in courtroom, as free expression is just one of those things we used to care about..
The issue in this case was not a small wall hanging obtained with personal funds. In this case it was a large monument obtained using several thousand dollars of state funds. It was moved in during the middle of the night without notifying anybody. I'm for free expression as much as the next person, and if Roy wanted to use his personal funds to procure a sign or something for his office that's fine, but using state funds to purchase a gigantic monument and placing it in the courthouse lobby of the AL state supreme court kind of crosses the line a bit, no?
And, although the media primarily focused on the religion aspects, that isn't what bothered most people. The problem was the reckless misappropriation of government funds and the clandestine procedures. If they have funded it some other way and gone through normal channels I'm not sure it would have been such a big issue.
Well, you see: 1) an iPod shuffle has about as much processing power than a GPS satellite. 2) both the iPod shuffle and the GPS satellites merely play information from their playlist: The GPS satellite is programmed to repeat the time plus whatever orbit information the ground station sent within the last 2-6 hours.
I just wish my iPod could sync wirelessly from geosynchronous orbit =P
Of course, that's physical entropy, and I don't know that it's the same as "information entropy."
You are probably referring the thermodynamical entropy, which is based on continuous state distributions. While in most cases continuous and discrete solutions are closely related (allowing summations to be replaced with integrals and vice versa), it has been shown in this case that these notions of entropy are not comparable (the limit of the discrete entropy as the number of divisions goes to infinity also goes to infinity, whereas the continuous integral is always finite). They are called the same since when information theory was first introduced (in Claude Shannon's 1948 paper -- which I might suggest you read [also his 1949 paper on cryptography is good too.]).
Of course there is also entropy in Statistical Mechanics (which starts from classical and quantum mechanical principles under a microscopic scale and uses statistical methods to examine what effects these have on a macroscopic scale [in essence explaining classical thermodynamics]). And this entropy is based on the total number of possible microstates a system can have (and since states in quantum mechanics is discrete this is finite), this definition of entropy is compatible with the information theoretical definition of entropy (but whether they are actually the same has not been proven).
In Information Theory entropy is usually measured in bits (minimum expected compressed size of an observation in binary digits) or nats (minimum expected compressed size of an observation in base e digits).
A truly random number would be a number that has no statistical correlation with any other observable quantity (past or future). I'm not sure it is possible to generate a truly random number since any physical method would create the potential for other data to be observed. Of course given the correlations and all possible observation the correlations could be removed, but this will never be practical. Additionally all that is really needed for cryptographic purposes is that the number have no statistical correlations with any quantity observed by any receiver, which is much simpler (and arguably what this article is about).
You could have done the "./configure && make && make install" in Redhat as well.
Most likely not. Since Redhat and other non-source based distributions don't include the header files necessary for compilation you have to go through the game of:
1. $./configure
**Error Lib Foo not found
2. Check if Foo.so or Foo.a is in $LD_LIBRARY_PATH, if not
a. Search for the name of the package that provides Lib Foo and install it. Then continue
3. Check if the header files for Lib Foo are in $INCLUDE_PATH, if not
a. Search for which package provides the header files for Lib Foo (usually ${answer to 2.a}-dev) and install it
4. $./configure
a. If that doesn't work try $./configure --with-foo-lib-path $PATH_TO_LIBFOO --with-foo-include-path $PATH_TO_FOO_HEADERS
**Error Lib Bar not found
Set Foo=Bar && Goto 2
one strategy that often worked was to sneak up (or unhide) a devourer which would cast dark swarm (no hits from air) and then plague. That one will happily suck down the life of enemy units, which to add to it your terran ally can sneak up a science-vessel, dump an EMP, and basically drop that annoying carrier-rush-force down to nearly nothing for life
Yeah, the best counter I found for that was to stasis your carrier fleet so they don't lose life from the plague and can recharge their shields. That of course leaves you very vulnerable, but is better than letting the fleet be destroyed because (1) they're already in the base (2) the enemy has to guard them.
Of course what was really fun is to MC zerg and recall a full supply of Ultralisks under your carrier fleet, but that only really works in money maps.
I'm not so sure it's entirely due to the community. Gentoo is just retroactively Murphy's Law compliant: Everything that can go wrong in Linux, has gone wrong (for someone with Gentoo).
This is also a hot topic in wireless networks. In multi-hop wireless networks it can enable nodes to forward many packets in a single transmission. A neat paper about this is here.
Also the coding people are going crazy about this too. There were a couple papers showing that network coding is a simple extension of linear block codes.
You actually could do it with a voice call (use modems). You could even do it by hand using morse code (or a higher order encoding) using button presses. Assuming strong enough encryption this would be unbreakable, and there would be no voices to determine who called whom (though it might be prudent to disconnect the microphone).
Actually your technical details are somewhat inaccurate:
T1 lines send digital signals with almost NO current. This is due to the balanced encoding used on the line. There are two primary encodings used in North America (Europe has their own standard): B8ZS and AMI. These encodings ensure that the number of positive signals sent are roughly the same as number of the negative signals sent, resulting in an average DC voltage close to zero. While I don't doubt your anecdote about techs using their fingers to test if a line is live, the signal they experience has more in common with AC than DC.
The electrical specifications of a T1 show that it uses {-12, 0 12}V DC as the signaling alphabet. This is not the "hundreds of DC volts" you claim (maybe you were confused with the POTS system which uses 90V RMS ringer signal).
I don't know much about the politics of the system (I've only designed endpoint equipment and had little interaction with customers), but I know your technical details were rather specious. Do you have any evidence to back your other claims?
Most modern browsers can show the characters. Even IE 6 and older versions of mozilla have no problem displaying them (though older versions of mozilla may require some tweaking). The way special characters are encoded in mail was designed to be compatible with already deployed servers (with some special tags and something similar to the base64 encoding used for attachments). These servers don't see anything other than plain 7bit ASCII text, so it is unlikely it became garbled during transit. The most likely cause is either the sender is using a poorly configured mail client (that isn't setting the codepage and escaping things correctly), or you are using a poorly configured mail client (that isn't respecting the codepage specified).
Actually 100% redundancy is a good bit better than two copies. If, instead of losing the entire 1st disk, you scratch the first half of both disks you can still recover everything. If the disks were just identical copies, you would be SOL...
Actually grandparent is roughly correct (minus hyperbole). It is an FCC requirement that the end user not be able to modify the device to cause harmful interference. The actual phrasing is just that it be reasonably difficult...
In practice this means device manufacturers 1) use nonstandard antenna connectors so it is difficult to use an antenna that wasn't certified for operation with the device. 2) Limit power and other controls via one of the following methods
a) Place limits in firmware (closed firmware)
b) Place limits in the driver (closed driver)
Note that there are other solutions to this problem, but these are just the most common ones. The other options are mostly hardware, and since the restrictions vary for different markets (Europe and Japan), most companies find it cheaper to produce one hardware model and just change the limits in software.
Slashdot Mirror
Wrong. OP is correct. All things being equal you can fit the same amount of data in 700-720MHz as in 2.5-2.52GHz. As another poster mentioned the difference is not capacity, but instead reusability.
Maybe you should study how routing protocols work.
For the type of ubiquitous network described, geographic routing should work plenty well, and could easily be implemented. If not there is a large amount of other protocols (AODV, Fish-eye Routing, etc) that have been developed to deal with the scalability problems.
Of course my real concern with this approach is bandwidth. In my opinion there is no way to cram that much data into the available wireless spectrum. It's been proven that the per-user bandwidth-distance ratio of such a network is monotonically decreasing as the number of users increases [1]. The fact is we will always need land-lines, since they have the unique and desirable properties of (1) not interfering with each other, and (2) being able to cross really long distances with relatively small amounts of power.
References:
1. P. Gupta and P. R. Kumar, "The capacity of wireless networks," IEEE Trans. Inform. Theory, vol. 46, pp. 388-404, Mar. 2000
I'm also an EE, and while this is not my area (it's not yours either!), I think you are simply trying to be over-pedantic.
First of all context is everything. If I say that this line is at 5V, then someone in the power field of EE would think there was really 7.07 V (peak) on it since they alway deal with RMS. Different fields can make different assumptions: in the digital field I can roughly assume that the voltage in my circuits is in 1 of two states (well mostly).
The leakage losses occur because silicon is an imperfect insulator. Even when transistors are 'off' (or switching) some current leaks through to the drain and bulk. This doesn't depend on the amount of current in the transistor doing useful work or even on the switching frequency, but only on the voltage. The actual power loss depends on the layout and operation of each transistor (with really complex interactions among them). There isn't really a simple resistor (in fact most models include 3 or more), but I was trying to give the layman's version. I may have goofed with the formula and I should have written V^2/R (though this is still far from accurate).
The power lost through switching is not reactive power. Reactive power is useful if you are planning a power distribution network, but not so much when you are calculating heat generation (since reactive power doesn't create heat). The switching losses occur because of the way CMOS logic works. When the pair of transistors changes from a 1 to a 0 the charge built up on the source of the NMOS transistor is dumped to ground (and from 0 to 1 with the PMOS dumped to Vcc). This charge is due to (among other things) the capacitance across the transistor, and when it is dumped through the NMOS transistor, all that energy is lost. The formula for energy in a capacitor is 0.5 C V^2, and since this amount of energy is lost every switching cycle, the power lost is 0.5 f C V^2. This is not the complete picture as there are other losses (and some devices shutoff portions of the chip not in use [clock gating, etc]).
Yes, the current will vary with the voltage, but there's really not need to over-specify. As long as you are using silicon processes the parameters are going to be roughly the same (though it would have been nice if they mentioned the fab process or the scale). If you can calculate the power with P=V^2/R and everybody knows R, why bother to provide I? Also since the second power of voltage is in all those equations, halving the voltage means you can more than quadruple the frequency with the same power consumption (okay not really, due to the transistor switching times), so small changes in frequency are insignificant compared to changes in voltage.
The difference is we are talking about semiconductor devices. Losses from these semiconductor devices are primarily due to leakage and switching. As long as we're still using silicon, leakage will be roughly 0.5 V^2/R, no matter how much current you pump through the transistors. Switching losses occur in when logic changes from 1's to 0's due to the capacitance of the transistors. The power lost here is roughly 0.5 f C V^2, where f is the switching frequency and C is the capacitance (material dependent). The V^2's means that reducing the voltage has a significant impact on losses. If we note that R and C are completely determined by the material (silicon) and the fabrication process, we can see that as long as the frequency is held constant, the voltage is a reasonable metric for comparing power consumption in silicon devices.
Of course this analysis is purely approximate since there are a lot of there things going in the devices. And I'm assuming complete capacitive discharge (independent of switching frequency), and didn't consider the changes in refresh rate to this DRAM device. But suffice it to say voltage is still a pretty good metric for comparison (until you actually build the thing and test it).
The ideal free market has no collusion or barriers to entry. Companies compete merely on merit (which is globally known to all consumers). Price is driven only by supply and demand. All goods are commodities, so there cannot be monopolies. Under this model the "invisible hand" WILL correct the market. This is really good, and I am not aware of another market model that can make this guarantee.
Yes net neutrality is against "laissez faire" capitalism. But this is necessary because laissez fair capitalism doesn't work in the real world. While traditional economic theory has shown that this sort of market provides the most efficient allocation of resources, the analysis makes several simplifying assumptions that limit its applicability. I could make a long list of these failed assumptions, but suffice it to say pure laissez faire capitalism is neither desirable nor workable.
Having said that, the laissez faire model can still be modified. If the government uses regulations to force the market to meet the assumptions of laissez faire capitalism, then capitalism can exist under those regulations, and will be at worst a local optimum. If you wish to call this socialism you may, but ideally the government should only interfere to ensure the market is competitive and only intervene when capitalism breaks down. This is the model currently employed by most modern countries, whether they admit it or not (ex. consumers have imperfect knowledge, so labeling regulations were added).
Ideally QoS would be combined with resource reservation schemes such as RSVP. Under these schemes the application requests the ability to send data at rate X with maximum latency Y/priority Z. Under these schemes the extra VoIP connections would only be allowed if there was already sufficient bandwidth to handle the traffic with low latency. In contrast, additional BULK traffic will generally be unconditionally accepted (the only limiting factor is the amount of buffer space available in intermediate routers).
This will enable the ISP to offer tiers of service: you can pay for 1 VoIP channel and 128kB/s BULK, or 3 video conferencing channels, 10 VoIP channels and 1MB/s BULK. If you want IP-TV, you will need to make sure you have enough video streams on your plan. People here are not likely to complain about that because (1) it is still net-neutral (2) it eliminates the "how unlimited" problem with ISPs now.
As for WiFi ISPs, I had one last year. Great service and price, $25/month for fast high quality internet access. Never had any problem with them. Unfortunately they went out of business, and I had the alternative of paying $60/month for cable internet or $60/month for ADSL. There are poorly supported ISPs everywhere, and I find it hard to believe your experience with one ISP's oversaturated infrastructure is evidence that shared infrastructure cannot work.
Net-neutrality is not anywhere near communism. (Under communist internets, streams resources reserve YOU!). Under communism there is centralized planning and management of the economy. This is actually a pretty close description of the Non-Net Neutral internet: the telco's decide the priority of packets of different types to different destinations. They decide whether or not you can set up your own IP-TV broadcasting company or make the next youtube. Of course since this is a company deciding this you might describe it as fascism.
Net-neutrality is capitalism. Collusion between ISPs and other companies and other non-competitive practices becomes more difficult. This forces the real world market to be a closer approximation of the ideal "free market".
Why not have computers keep track of TAI time and then translate to sidereal time for the UI. For most cases this would be a simple switch since they don't properly handle leap seconds anyway. Then, the addition or subtraction of leap seconds could be added to the UI.
That's why you don't make VoIP work "better" than bittorrent, you make it work "different" than bittorrent. With QoS your VoIP (Real-time streams) would get say, a fixed 9kb/s or whatever of "Guaranteed Low Latency" (TM). And your bittorrent (BULK Traffic) would get what ever is left over, but makes no guarantee of when your packets will arrive. The point is that if you make bittorrent act like VoIP, it will be limited to the real-time rate which should be slower than the BULK rate.
Of course the gotcha there is the "should be". If the telco's are cheap and don't upgrade, then even QoS can't stop the brownouts. But then again if the telco's don't upgrade, there'll be brownouts anyway...
It's not a bribe--there is no individual that is receiving the money. Instead, Microsoft would spend $$ marketing the TCS's classmate PC in Africa (if they choose windows).
Things like this occur in businesses every day in America. This is neither illegal or even shady. As usual the slashdot summary is slightly biased.
The point is there is a difference between interactive traffic and bulk traffic. While your 4.5 GB Linux ISO download doesn't care so much that it suffers from 250ms of latency while waiting in low priority queue for an opening, someone's VOIP or streaming video would (buffering....) be (buffering...) un (buffering...) useable with that much delay.
Ideally QoS would adjust latency based on application while keeping bandwidth allocations fair. So instead of simply ``letting your connection suffer'' QoS really allows them to use flexible limits on bandwidth and latency. Without QoS the only way to guarantee some level of service to to sompletely separate the individual users. Segregating everyone's pipes to guarantees that no one interferes with anyone else (and wastes alot of the bandwidth allocated to the email checkers). With QoS, they can let your bulk transfer borrow some bandwidth from some one who isn't using all of their's, and when they do send a VOIP packet you "return the favor" by letting them ahead in line. You benefit from the extra available bandwidth, and they benefit from the reduced latency.
And, although the media primarily focused on the religion aspects, that isn't what bothered most people. The problem was the reckless misappropriation of government funds and the clandestine procedures. If they have funded it some other way and gone through normal channels I'm not sure it would have been such a big issue.
It appears I'm wrong. The satellites are in "Medium Earth Orbit".
Well, you see:
1) an iPod shuffle has about as much processing power than a GPS satellite.
2) both the iPod shuffle and the GPS satellites merely play information from their playlist: The GPS satellite is programmed to repeat the time plus whatever orbit information the ground station sent within the last 2-6 hours.
I just wish my iPod could sync wirelessly from geosynchronous orbit =P
Of course there is also entropy in Statistical Mechanics (which starts from classical and quantum mechanical principles under a microscopic scale and uses statistical methods to examine what effects these have on a macroscopic scale [in essence explaining classical thermodynamics]). And this entropy is based on the total number of possible microstates a system can have (and since states in quantum mechanics is discrete this is finite), this definition of entropy is compatible with the information theoretical definition of entropy (but whether they are actually the same has not been proven).
In Information Theory entropy is usually measured in bits (minimum expected compressed size of an observation in binary digits) or nats (minimum expected compressed size of an observation in base e digits).
A truly random number would be a number that has no statistical correlation with any other observable quantity (past or future). I'm not sure it is possible to generate a truly random number since any physical method would create the potential for other data to be observed. Of course given the correlations and all possible observation the correlations could be removed, but this will never be practical. Additionally all that is really needed for cryptographic purposes is that the number have no statistical correlations with any quantity observed by any receiver, which is much simpler (and arguably what this article is about).
1. $./configure
**Error Lib Foo not found
2. Check if Foo.so or Foo.a is in $LD_LIBRARY_PATH, if not
a. Search for the name of the package that provides Lib Foo and install it. Then continue
3. Check if the header files for Lib Foo are in $INCLUDE_PATH, if not
a. Search for which package provides the header files for Lib Foo (usually ${answer to 2.a}-dev) and install it
4. $./configure
a. If that doesn't work try $./configure --with-foo-lib-path $PATH_TO_LIBFOO --with-foo-include-path $PATH_TO_FOO_HEADERS
**Error Lib Bar not found
Set Foo=Bar && Goto 2
Of course what was really fun is to MC zerg and recall a full supply of Ultralisks under your carrier fleet, but that only really works in money maps.
I'm not so sure it's entirely due to the community. Gentoo is just retroactively Murphy's Law compliant: Everything that can go wrong in Linux, has gone wrong (for someone with Gentoo).
This is also a hot topic in wireless networks. In multi-hop wireless networks it can enable nodes to forward many packets in a single transmission. A neat paper about this is here.
Also the coding people are going crazy about this too. There were a couple papers showing that network coding is a simple extension of linear block codes.
You actually could do it with a voice call (use modems). You could even do it by hand using morse code (or a higher order encoding) using button presses. Assuming strong enough encryption this would be unbreakable, and there would be no voices to determine who called whom (though it might be prudent to disconnect the microphone).
I stand corrected. I've only dealt with "dry" T1. I was not aware span powered t1 existed (I've only seen that with the *DSLs).
T1 lines send digital signals with almost NO current. This is due to the balanced encoding used on the line. There are two primary encodings used in North America (Europe has their own standard): B8ZS and AMI. These encodings ensure that the number of positive signals sent are roughly the same as number of the negative signals sent, resulting in an average DC voltage close to zero. While I don't doubt your anecdote about techs using their fingers to test if a line is live, the signal they experience has more in common with AC than DC.
The electrical specifications of a T1 show that it uses {-12, 0 12}V DC as the signaling alphabet. This is not the "hundreds of DC volts" you claim (maybe you were confused with the POTS system which uses 90V RMS ringer signal).
I don't know much about the politics of the system (I've only designed endpoint equipment and had little interaction with customers), but I know your technical details were rather specious. Do you have any evidence to back your other claims?
Most modern browsers can show the characters. Even IE 6 and older versions of mozilla have no problem displaying them (though older versions of mozilla may require some tweaking). The way special characters are encoded in mail was designed to be compatible with already deployed servers (with some special tags and something similar to the base64 encoding used for attachments). These servers don't see anything other than plain 7bit ASCII text, so it is unlikely it became garbled during transit. The most likely cause is either the sender is using a poorly configured mail client (that isn't setting the codepage and escaping things correctly), or you are using a poorly configured mail client (that isn't respecting the codepage specified).
Actually 100% redundancy is a good bit better than two copies.
If, instead of losing the entire 1st disk, you scratch the first half of both disks you can still recover everything. If the disks were just identical copies, you would be SOL...
Actually grandparent is roughly correct (minus hyperbole). It is an FCC requirement that the end user not be able to modify the device to cause harmful interference. The actual phrasing is just that it be reasonably difficult...
In practice this means device manufacturers
1) use nonstandard antenna connectors so it is difficult to use an antenna that wasn't certified for operation with the device.
2) Limit power and other controls via one of the following methods
a) Place limits in firmware (closed firmware)
b) Place limits in the driver (closed driver)
Note that there are other solutions to this problem, but these are just the most common ones. The other options are mostly hardware, and since the restrictions vary for different markets (Europe and Japan), most companies find it cheaper to produce one hardware model and just change the limits in software.