Here, p(SPAM,sex,sexy), p(sex), p(sexy), p(sex|sexy) and p(sexy|sex) are known. The system finds p(SPAM|sex, sexy).
Generalizing, this system is finding p(SPAM|X0,X1..Xn)=p(SPAM,X0,...Xn)/(p(X0)p(X1|X0)...p(Xn|Xn-1,...,X0)) The Xi are the "most interesting", that is those for which (p(SPAM/Xi)-0.5)^2 is greater.
Re:European conspiracy theories...
on
World Cup Final
·
· Score: 1
There were quite a few more mistakes than a single "disallowed goal". Three of such "phantom" goals and several (non) offsides in actions which could easily led to a goal (goal keeper alone against striker, etc).
The problem is that, even though there was no conspiration, the referees had no experience (coming from countries with little football culture such as Uganda and Egypt). Only after this match did Blatter, FIFA's president, change its criteria for the choice of the referees. It's no accident the referee of the final was Italian (Collina, often considered the best one in the world).
BTW, here are some of the facts which could sustain the conspiration theory (not that I believe in it):
* one of FIFA's vicepresident is Korean and will try his chance in the next presidential elections of S. Korea
* Blatter won FIFA's election shortly before the World Cup, with Villar's (the president of the Spanish football federation) support.
* Blatter's rivals control the referee designation council, and would have forced the presence of non-experienced referees to punish Villar for supporting of Blatter
>The Spanish Talgo is in the works and will do 350 kph (218 mph).
Actually the Talgo is a rather older train... You meant the AVE (Alta Velocidad Española == Spanish High Speed; "ave" means "bird"), which runs commercially at about 300kph. It's been here since 1992, between Madrid and Sevilla, but the government is investing a lot on it at the moment (I think it was some ~4 billion $) as it looks forward to connect it with the french high speed railroad; the new trains will be faster.
BTW, you get a full refund if the train is delayed more than 5 minutes:-)
> It does nothing to increase the capacity of the available bandwidth.
THAT is the exact wording to use:-) The capacity of the available bandwidth of one link is not increased.
Having more capacity (in several links) is ANOTHER thing.
An analogy for those who still don't get it.. having more pipes means more water can flow through them all at the same time, but the capacity of ONE pipe is the same, no matter how many you put in your system.
> Slightly off-topic, perhaps, but the current limits of the radio spectrum are transient and purely technical.
You're just plain wrong, sorry.
>In theory we could cram an almost infinite number of bits into an almost infinitely small timeframe into an almost infinitely small frequency-range.
No. Read Shannon's papers. The capacity (bits per sec) is given by the bandwidth you use and the signal-to-noise ratio you have. What you can do is SPLIT the medium (ie "add more links") and this way the net total capacity of the network increases. But the capacity of each link is given by Shannon's law and this is a physical limit.
C (bits/s) = W(Hz) * log2 ( 1 + SNR )
You can't change that. This is NOT a transient or technological limit. We're already very close to it in some digital modulations.
The wording is quite poor... there's no increase of the bandwidth; you're just cutting the medium (the ether) into several segments. In each segment you can use (up to) all the available bandwidth of the medium.
This is what is happening: the air is one gigantic medium, just as a cable. If we all use the same cable to transmit, there's a limit in how many people can transmit at a time. If you cut it into n pieces, more people can speak at a time since the signal just goes through the necessary links, not ALL of the cable.
That's it. There's no bandwidth increase (in the physical sense) involved. It remains the same. But the total CAPACITY of the network grows if you're doing point-to-point communications, as in each segment you can use all the bandwidth of the ether you want.
This is just about the same thing as the Ethernet. The air is the ether. You can divide it and get several segments. If you only have one, collisions are frequent.
>I don't know that this is true. in that the 300 mhz of bandwidth is still 300 mhz of bandwidth.
But not all parts of the spectrum behave equally... The propagation of each frequency is quite different (see my anonymous post listing the max. distances and the different ways the waves propagate), and you absolutely prefer the shortwave range for some things. The antennae are different.
>You could very easily have AM radio in the gighertz band. 44khz band width (CD audio. etc) on a frequency of 4 giga hetrz. But it would be rather line of site, among other technical issues.
Plus the receptor would be much more expensive. You'd have to make a superheterodine repector with a number of intermediate frequencies, much more than the low-freq converter and the diode you need to make a simple AM radio... Anyway, you would not be able to put the AM signals as close one to another in the 4GHz as in the MF band... Plus the attenuation by fog and rain must be considered. This doesn't mix well with amplitude modulation...
>The very same technology that regulates printing in LAN's at universities can regulate the airwaves.
No it cannot. There's a reason for a single Ethernet segment having a maximum size... It's the way the MAC (medium access control) works.
Say you've got two hosts A and B, on opposite ends of your LAN (from the geographical POV), which want to send a frame to another one in the middle, C. MAC works this way: the sending host first listens to the ether (the medium), if it hears something, it just waits till it isn't used anymore (or rather a random amount of time); if not, it sends its packet. This seems to work fairly well, but, consider this: A sends a packet. It travels at c0/sqrt(epsilon) through the medium, which on a twisted pair cable is about 1.5*10^8m/s Before the signal gets to B, the latter decides to send another thing to C; it doesn't hear anything in the ether, and so sends its packet. Then C ends up receiving the sum of A and B's signals, that is, garbage. There's a collision and no transmission was successful, you've lost bandwidth and time. After a while, both A and B realize they didn't send their packet, and the thing must be restarted again, each host waiting for a random period of time before retry.
>Two people send a request to a printer to print a document at the same time; the printer doesn't know which to process first
There's a wrong assumption there: if both people send the request at the same time, using the same frequencies, the printer won't understand any of the requests, unless you use CDMA or another kind of multiplex access (and you're not using FDMA since you don't want this to be regulated nor TDMA, since they transmit at the same time). The sum of two valid signals is garbage, unless specifically designed not to be so.
The collision rate grows with the maximum propagation delay between two hosts, and the number of hosts (and the traffic their generate). It also grows with the time it takes to transmit a single packet.
Imagine how hardly collisions would hinder the performance of a wireless network being used at the same time by everybody in, say, a city. Plus why on earth would you want to have say a full 1GHz of spectrum (that is, several gigabits per sec depending of the digital modulation you use) to transmit the data of *a single user*?
Plus there's the issue of the ways each frequency propagates. As I said in another post, VLF gives worldwide coverage, LF and MF propagate via ground wave (although MF can also use ionosphere refraction), HF uses the ionosphere. VHF and UHF go on "regular" (spherical) waves, microwaves are line of sight. Each requires a different antenna, transmission power, etc...
Slashdot Mirror
He should be doing this:
. ..p(Xn|Xn-1,...,X0))
p(SPAM|sex,sexy) = p(SPAM,sex,sexy)/p(sex,sexy)=
= p(SPAM,sex,sexy)/(p(sex)*p(sexy|sex))=
= p(SPAM,sex,sexy)/(p(sexy)*p(sex|sexy))
Here, p(SPAM,sex,sexy), p(sex), p(sexy), p(sex|sexy) and p(sexy|sex) are known. The system finds p(SPAM|sex, sexy).
Generalizing, this system is finding p(SPAM|X0,X1..Xn)=p(SPAM,X0,...Xn)/(p(X0)p(X1|X0)
The Xi are the "most interesting", that is those for which (p(SPAM/Xi)-0.5)^2 is greater.
There were quite a few more mistakes than a single "disallowed goal". Three of such "phantom" goals and several (non) offsides in actions which could easily led to a goal (goal keeper alone against striker, etc).
The problem is that, even though there was no conspiration, the referees had no experience (coming from countries with little football culture such as Uganda and Egypt). Only after this match did Blatter, FIFA's president, change its criteria for the choice of the referees. It's no accident the referee of the final was Italian (Collina, often considered the best one in the world).
BTW, here are some of the facts which could sustain the conspiration theory (not that I believe in it):
* one of FIFA's vicepresident is Korean and will try his chance in the next presidential elections of S. Korea
* Blatter won FIFA's election shortly before the World Cup, with Villar's (the president of the Spanish football federation) support.
* Blatter's rivals control the referee designation council, and would have forced the presence of non-experienced referees to punish Villar for supporting of Blatter
It seems Talgo has made some things in the US... Its trains are being used between Portland and Vancouver. Amtrak and WSDOT bought some too. See also Talgo's site
>The Spanish Talgo is in the works and will do 350 kph (218 mph).
:-)
Actually the Talgo is a rather older train... You meant the AVE (Alta Velocidad Española == Spanish High Speed; "ave" means "bird"), which runs commercially at about 300kph. It's been here since 1992, between Madrid and Sevilla, but the government is investing a lot on it at the moment (I think it was some ~4 billion $) as it looks forward to connect it with the french high speed railroad; the new trains will be faster.
BTW, you get a full refund if the train is delayed more than 5 minutes
> It does nothing to increase the capacity of the available bandwidth.
:-) The capacity of the available bandwidth of one link is not increased.
THAT is the exact wording to use
Having more capacity (in several links) is ANOTHER thing.
An analogy for those who still don't get it..
having more pipes means more water can flow through them all at the same time, but the capacity of ONE pipe is the same, no matter how many you put in your system.
> Slightly off-topic, perhaps, but the current limits of the radio spectrum are transient and purely technical.
You're just plain wrong, sorry.
>In theory we could cram an almost infinite number of bits into an almost infinitely small timeframe into an almost infinitely small frequency-range.
No.
Read Shannon's papers.
The capacity (bits per sec) is given by the bandwidth you use and the signal-to-noise ratio you have. What you can do is SPLIT the medium (ie "add more links") and this way the net total capacity of the network increases. But the capacity of each link is given by Shannon's law and this is a physical limit.
C (bits/s) = W(Hz) * log2 ( 1 + SNR )
You can't change that. This is NOT a transient or technological limit. We're already very close to it in some digital modulations.
The wording is quite poor... there's no increase of the bandwidth; you're just cutting the medium (the ether) into several segments. In each segment you can use (up to) all the available bandwidth of the medium.
This is what is happening: the air is one gigantic medium, just as a cable. If we all use the same cable to transmit, there's a limit in how many people can transmit at a time. If you cut it into n pieces, more people can speak at a time since the signal just goes through the necessary links, not ALL of the cable.
That's it. There's no bandwidth increase (in the physical sense) involved. It remains the same. But the total CAPACITY of the network grows if you're doing point-to-point communications, as in each segment you can use all the bandwidth of the ether you want.
This is just about the same thing as the Ethernet. The air is the ether. You can divide it and get several segments. If you only have one, collisions are frequent.
>Why AM and wideband FM are still around for voice is more a toy and lack of education - than real need.
It's just because digital radios are so expensive nobody owns one. It'll take some time before *all* the old analog AM radios are thrown away...
>I don't know that this is true. in that the 300 mhz of bandwidth is still 300 mhz of bandwidth.
But not all parts of the spectrum behave equally... The propagation of each frequency is quite different (see my anonymous post listing the max. distances and the different ways the waves propagate), and you absolutely prefer the shortwave range for some things. The antennae are different.
>You could very easily have AM radio in the gighertz band. 44khz band width (CD audio. etc) on a frequency of 4 giga hetrz. But it would be rather line of site, among other technical issues.
Plus the receptor would be much more expensive. You'd have to make a superheterodine repector with a number of intermediate frequencies, much more than the low-freq converter and the diode you need to make a simple AM radio...
Anyway, you would not be able to put the AM signals as close one to another in the 4GHz as in the MF band...
Plus the attenuation by fog and rain must be considered. This doesn't mix well with amplitude modulation...
>The very same technology that regulates printing in LAN's at universities can regulate the airwaves.
No it cannot. There's a reason for a single Ethernet segment having a maximum size... It's the way the MAC (medium access control) works.
Say you've got two hosts A and B, on opposite ends of your LAN (from the geographical POV), which want to send a frame to another one in the middle, C. MAC works this way: the sending host first listens to the ether (the medium), if it hears something, it just waits till it isn't used anymore (or rather a random amount of time); if not, it sends its packet. This seems to work fairly well, but, consider this: A sends a packet. It travels at c0/sqrt(epsilon) through the medium, which on a twisted pair cable is about 1.5*10^8m/s Before the signal gets to B, the latter decides to send another thing to C; it doesn't hear anything in the ether, and so sends its packet. Then C ends up receiving the sum of A and B's signals, that is, garbage. There's a collision and no transmission was successful, you've lost bandwidth and time. After a while, both A and B realize they didn't send their packet, and the thing must be restarted again, each host waiting for a random period of time before retry.
>Two people send a request to a printer to print a document at the same time; the printer doesn't know which to process first
There's a wrong assumption there: if both people send the request at the same time, using the same frequencies, the printer won't understand any of the requests, unless you use CDMA or another kind of multiplex access (and you're not using FDMA since you don't want this to be regulated nor TDMA, since they transmit at the same time). The sum of two valid signals is garbage, unless specifically designed not to be so.
The collision rate grows with the maximum propagation delay between two hosts, and the number of hosts (and the traffic their generate).
It also grows with the time it takes to transmit a single packet.
Imagine how hardly collisions would hinder the performance of a wireless network being used at the same time by everybody in, say, a city. Plus why on earth would you want to have say a full 1GHz of spectrum (that is, several gigabits per sec depending of the digital modulation you use) to transmit the data of *a single user*?
Plus there's the issue of the ways each frequency propagates. As I said in another post, VLF gives worldwide coverage, LF and MF propagate via ground wave (although MF can also use ionosphere refraction), HF uses the ionosphere. VHF and UHF go on "regular" (spherical) waves, microwaves are line of sight. Each requires a different antenna, transmission power, etc...