True for analog phones, but while it's possible to snoop on spread spectrum digital phones (available in all frequency bands, although 900 MHz units are becoming VERY hard to find in both analog and digital flavors), it's much more difficult and I do not know of any off-the-shelf reasonably priced scanner that can snoop FHSS or DSS digital phones. I'm not saying it's impossible, it's just MUCH more difficult to the point where it would require a determined attacker with lots of time and money, someone who probably has the resources to install a physical tap on your landline or snoop your ISP connection in the case of VoIP.
Also, modern digital cordless phones may use encryption, but I am not sure of this. Hmm, I should look into the specs of my new Uniden 5.8 GHz set.
FYI, 802.11a and 5.8 GHz cordless phones operate in different parts of the ISM bands up in that range, so won't interfere with each other. This is, of course, not true for 2.4 GHz phones and 802.11b/g gear, which love to clobber each other and get clobbered by microwave ovens.
Except for extremely high end test equipment, I do not know of anything that performs signal processing at sample rates on the order of 1-2 GSPS (gigasamples/second) or higher.
Even modern cell phones are *at most* doing direct IF sampling and IF signal generation. Upconversion to/downconversion from a 1-2 GHz carrier frequency is still done in the analog domain. Most likely, upconversion to/downconversion from IF is also done in the analog domain, with the only digital portion being baseband processing. Direct IF sampling/transmission is still rather expensive, too expensive for your average cell phone.
Unfortunately the article is now Slashdotted so I can't tell whether this 500 GHz chip was RF circuitry operating at a 500 GHz carrier frequency (20-30 GHz plus is common nowadays in this domain), or digital circuitry operating at a clock speed of 500 GHz, and if it was digital, how complex a circuit it was. It's much easier to clock a simple circuit at very high frequencies than a complex one like a CPU.
When I moved to my new apartment, due to previous experiences with cable modem companies (port blocking, stealth caps, etc) which are *typically* less prevalent on DSL (partly due to the fact that they have a much lower advertised cap to begin with), I was seriously considering DSL.
One problem: I wasn't in any particular hurry to get a phone line, as my cell phone was mostly sufficient for my needs and my plan was to go with VoIP for my secondary voice line if I ever started pushing the limits of my cell phone plan. Not a SINGLE DSL provider would give me a rate/price quote based on address alone, or even address, area code, and a guess as to the exchange. Every place I tried to shop for service wanted a *complete* landline phone number.
I'm sorry, but if you won't even give me a price quote before I obtain something I don't want/need, I don't plan on buying from you. As a result I'm with Time Warner RoadRunner, and (for the most part) pretty happy with them. They're much better than OptimumOffline.
Turn-on of incandescents is not instant. It's almost unnoticeable except with very high power incandescents (such as stage lights), but even for low power incandescents, it is enough to make a subconscious difference. I believe studies into benefits of LED brake lights showed that the instant turnon of LED brake lights was equivalent to a full car length's worth of reaction time at highway speeds.
Also, at least in the past, while LED traffic lights do save power, their primary benefit has been their long lifetime. The labor cost of replacing a traffic light bulb is nearly the cost of even the new LED replacement bulbs, so even at 2-3 times the initial purchase price, they pay for themselves in labor alone within 2-3 incandescent bulb change periods.
The submitter even said that the survey was of school teachers. 40% of school teachers holding a second job is FAR different than almost anyone else holding a second job. School teachers inherently have more free time to work a second job than the average 9-5er, ESPECIALLY during the summer.
I'm shocked the figure is only 40%, unless they're not counting summer-only jobs in that survery.
See the numerous previous comments about credit unions. My primary bank is now a local credit union, and I think 75% of the employees at my company do business with that CU (part of that is because they have a branch office and ATMs *in* our facility!).
Yes, that's right. If I want to hit one of my bank's ATMs, I walk down the hall and one floor down from my cube and it's right there. Need to do something that requires a teller in an office? 20 feet from the cafeteria.
Note that aluminum production is notorious for using lots of electrical power for the same reasons. As a result aluminum refining plants are almost always located close to power plants. Despite this, aluminum is quite cheap.
With 802.11g you can only serve a total of 45 users. 15 users per AP times 3 non-overlapping 11g channels = 45.
In theory you can use highend equipment with transmit power management to set up an effective microcellular network with more APs. If you want to do this, ALL hardware (clients and APs) must support automatic transmit power management. Multiple low-power APs will let you increase density, but this is difficult and expensive.
802.11a gives you 20 nonoverlapping channels to work with, but has its own issues to deal with.
Once you're talking about 802.11a gear or high-end properly placed/installed 802.11g gear for a high density network, you're better off just installing gigabit Ethernet switches and Cat5e and getting (easily) 10 times the real-world throughput or more.
"Speed. As far as I know, pre-N technology hasn't been fully adopted and the best you can do is 802.11g (54Mbps) basically half of what you would get with wired (100Mbps). Granted you rarely ever get the full 100Mbps, but you rarely ever get the full 54Mbps either."
It's worse than that. The CSMA/CA collision management protocol used by 802.11 is inherently less efficient than CSMA/CD used by wired Ethernet. The throughput of an 802.11 system will always be a lower fraction of the signaling rate than even shared Ethernet via a hub. Switched Ethernet? No comparison. Switched Ethernet can easily reach 90-95% of the signaling rate in terms of real throughput. 802.11 (any variant) is lucky to get 50% to a single user. With multiple users and small packets, this drops significantly.
"booting from known-good media, and either testing MD5 sums of each executable, or a wipe/re-install." You just contradicted yourself. There's nothing preventing someone from creating a rootkit-hunting LiveCD.
Yes, there are definately major issues to be dealt with for some sort of multicast support for BT (such as, for example, the fact that torrent users don't all have the same downstream bandwidth, and even when multicast support starts appearing in ISPs it'll take a long time for full penetration so the "legacy" way needs to be supported.) There's also the TCP streams vs. UDP datagrams issue.
It would actually be easier to implement this as a seperate protocol that was designed to augment BitTorrent. (i.e. getting data both via BitTorrent and via a multicast mechanism.) Either way, end-to-end multicast support on the public Internet would contribute to a massive reduction in backbone load from P2P, which has in the past been claimed to consist of 30-50% of all backbone traffic now. (I don't know how accurate this number really is though.)
If I recall correctly, supporting IP Multicast was required instead of optional for IPv6 implementations.
If this is the case, a multicast-aware BitTorrent would be THE killer app, IF IPv6 were deployed sufficiently for multicast torrents to be effective.
The way things are now, a multicast torrent would be pretty much the same trafficwise as the way things currently are for the backbone, since for the most part everyone is tunneling to one of a small handful of IPv6 brokers.
Not necessarily, because even some of the name brands can sometimes be borderline quality.
For example, Memorex media is almost always made by CMC Magnetics. In my experience, while they're far better than the likes of Princo, they are not a manufacturer I will trust. Even many of the generic noname brands are better than CMC-sourced discs, as in my experience anything made by Ritek is far better than CMC media, and Riteks are almost never sold as name-brand.
Verbatim media used to be high-quality almost all the time, but I heard that they started selling some batches of CMCs, so you never knew what you were getting.
Only a few of the name brands are guaranteed to give you the same original manufacturer every time because they ARE the original manufacturer. Off of the top of my head, the only two examples of this I can think of are Sony and TDK.
British or U.S, unless you've badly screwed up and didn't do any proofreading or you are making a comedy spoof of vampire movies, no one ever drives steaks through anyone's heart. It's STAKE, and the editors of this rag didn't even catch such an obvious error.
I'm inclined to disagree with that. I just finished my M.S. at Rutgers, where the majority of the students in the graduate program were either Chinese or Indian. (The lack of U.S. citizens in a graduate program at a state university where tuition is DIRT CHEAP for state residents and pretty inexpensive for others says bad things about U.S. attitudes towards higher education...) The Chinese students, for the most part, barely spoke English at all and I'm surprised any of them were able to complete their classwork given the difficulty of understanding and communicating with their professors and fellow classmates. The Indian students had at least been educated reasonably well in English so that they were understandable. The problem is that while they may have been educated in English since the 11th grade, they were most likely educated in English by a non-native English speaker. Think of the game of Telephone, where after being passed from person to person, a message is distorted so much as to be barely understandable. It's same thing with "second generation" (or more) English language education. Yes, it was possible to communicate with the Indian students and much easier than the Chinese students, but it was still *extremely* difficult due to the thick accents. Yes, even if one's grammar is perfect, a thick accent can make verbal communication extremely difficult. By the way, most of the Indian students in question could write extremely well. It was only their spoken language that was difficult to understand. Unfortunately, even in technical fields, verbal communication is important. In technical support fields where outsourcing is currently the most common, verbal communications is *EVERYTHING*.
You snuck one rather BAD thing in your list of things that are good about Cuba - Infant mortality is a Bad Thing.
That said, after seeing Yank Tanks - I am amazed at the resourcefulness of some of Cuba's citizens. If the US had auto mechanics as competent as the Cubans, our auto industry would go out of business because cars would last so damn long.:)
The SanDisk units do not use mechanical hard drives, but use solid state flash memory. Thus their direct competitor is the iPod nano, Apple's solid-state player.
iPod nano 2GB - $199 iPod nano 4GB - $249 iPod nano 6GB - does not exist The 2 and 4 GB SanDisk variants are $20 cheaper. Not sure how they compare on features, but in terms of price per GB, the SanDisks beat the iPod nano.
Obviously you didn't put much effort into figuring out what these satellites do.
For example, if you click on Cornell's ICE CubeSat, it takes you to the AMSAT info page for the sat, which has a link to Cornell's own page for the sat, which has LOTS of details on the design of the sat, and more importantly, the science package the sat is carrying. Most of the other university sats are also carrying some sort of science package (most of them are cameras I believe, but I'm not sure.)
To summarize it in basic terms, the Cornell ICE Cube is designed to take measurements of ionospheric phenomenon that have been causing problems with the GPS system in certain parts of the world at certain times. The phenomenon is known as scintillation and causes rapid fluctuations in the signal strength of GPS signals reaching the ground. ICE Cube is meant to take measurements of such phenomenon at high altitudes, eliminating any effects the lower layers of the atmosphere might have on the signal.
This is all assuming that the comm system works... I designed the first revision of the onboard transciever and I admit I didn't do the best job due to lack of time/experience. (I was a senior with only one semester to work on it.) From what I've heard the grad students assigned to clean up my mess did a pretty good job though.:)
"unless you mean because they're violating someone elses and afraid to get caught. Patents are, by definition and law, public disclosures."
You forget that it's possible to LICENSE someone else's patents. That is, you are given permission to legally use their patents under certain conditions.
For whatever reason, those conditions have apparently required closed source drivers. I don't know about NVidia's excuse, but this was ATI's reasoning for no longer providing full specifications of their chipsets.
I've never had any problems getting any NVidia GPU to work flawlessly under Linux.
Yeah, the closed-source aspect of NVidia's drivers may be annoying to some, but I don't mind closed-source drivers if they're high enough quality, and NVidia's drivers are one of the few examples of closed source software with high quality. (The associated closed-source games for Linux, specifically iD Software's products, comprise most of the other examples...)
Face it - due to patent issues out of the chipset manufacturers' control (classic example being S3 Texture Compression - S3TC was the beginning of ATI's transition from fully documented open source drivers to binary-only drivers with the open-source versions lacking critical features for 3D gaming), no chipset manufacturer can release open source drivers that support their card's full feature set, unless their card's feature set is massively crippled. (See Intel GMA-series integrated graphics as an example.)
Normally I would trust the company's own site the most as far as publicity and propaganda, but obviously they are not listing many of their older investments. Perhaps they may no longer have a stake in any of those other companies listed on the Wikipedia page.
Slashdot Mirror
True for analog phones, but while it's possible to snoop on spread spectrum digital phones (available in all frequency bands, although 900 MHz units are becoming VERY hard to find in both analog and digital flavors), it's much more difficult and I do not know of any off-the-shelf reasonably priced scanner that can snoop FHSS or DSS digital phones. I'm not saying it's impossible, it's just MUCH more difficult to the point where it would require a determined attacker with lots of time and money, someone who probably has the resources to install a physical tap on your landline or snoop your ISP connection in the case of VoIP.
Also, modern digital cordless phones may use encryption, but I am not sure of this. Hmm, I should look into the specs of my new Uniden 5.8 GHz set.
FYI, 802.11a and 5.8 GHz cordless phones operate in different parts of the ISM bands up in that range, so won't interfere with each other. This is, of course, not true for 2.4 GHz phones and 802.11b/g gear, which love to clobber each other and get clobbered by microwave ovens.
Except for extremely high end test equipment, I do not know of anything that performs signal processing at sample rates on the order of 1-2 GSPS (gigasamples/second) or higher.
Even modern cell phones are *at most* doing direct IF sampling and IF signal generation. Upconversion to/downconversion from a 1-2 GHz carrier frequency is still done in the analog domain. Most likely, upconversion to/downconversion from IF is also done in the analog domain, with the only digital portion being baseband processing. Direct IF sampling/transmission is still rather expensive, too expensive for your average cell phone.
Unfortunately the article is now Slashdotted so I can't tell whether this 500 GHz chip was RF circuitry operating at a 500 GHz carrier frequency (20-30 GHz plus is common nowadays in this domain), or digital circuitry operating at a clock speed of 500 GHz, and if it was digital, how complex a circuit it was. It's much easier to clock a simple circuit at very high frequencies than a complex one like a CPU.
FIOS used to be their Fiber To The Home service only, and not DSL. Their DSL offerings were marketed differently.
AMEN.
When I moved to my new apartment, due to previous experiences with cable modem companies (port blocking, stealth caps, etc) which are *typically* less prevalent on DSL (partly due to the fact that they have a much lower advertised cap to begin with), I was seriously considering DSL.
One problem: I wasn't in any particular hurry to get a phone line, as my cell phone was mostly sufficient for my needs and my plan was to go with VoIP for my secondary voice line if I ever started pushing the limits of my cell phone plan. Not a SINGLE DSL provider would give me a rate/price quote based on address alone, or even address, area code, and a guess as to the exchange. Every place I tried to shop for service wanted a *complete* landline phone number.
I'm sorry, but if you won't even give me a price quote before I obtain something I don't want/need, I don't plan on buying from you. As a result I'm with Time Warner RoadRunner, and (for the most part) pretty happy with them. They're much better than OptimumOffline.
Turn-on of incandescents is not instant. It's almost unnoticeable except with very high power incandescents (such as stage lights), but even for low power incandescents, it is enough to make a subconscious difference. I believe studies into benefits of LED brake lights showed that the instant turnon of LED brake lights was equivalent to a full car length's worth of reaction time at highway speeds.
Also, at least in the past, while LED traffic lights do save power, their primary benefit has been their long lifetime. The labor cost of replacing a traffic light bulb is nearly the cost of even the new LED replacement bulbs, so even at 2-3 times the initial purchase price, they pay for themselves in labor alone within 2-3 incandescent bulb change periods.
UTFA - Understand the FA.
The submitter even said that the survey was of school teachers. 40% of school teachers holding a second job is FAR different than almost anyone else holding a second job. School teachers inherently have more free time to work a second job than the average 9-5er, ESPECIALLY during the summer.
I'm shocked the figure is only 40%, unless they're not counting summer-only jobs in that survery.
See the numerous previous comments about credit unions. My primary bank is now a local credit union, and I think 75% of the employees at my company do business with that CU (part of that is because they have a branch office and ATMs *in* our facility!).
Yes, that's right. If I want to hit one of my bank's ATMs, I walk down the hall and one floor down from my cube and it's right there. Need to do something that requires a teller in an office? 20 feet from the cafeteria.
Note that aluminum production is notorious for using lots of electrical power for the same reasons. As a result aluminum refining plants are almost always located close to power plants. Despite this, aluminum is quite cheap.
"and, although I'm not a metalurgist, is there any reason that titanium couldnt replace steel almost entirely if it was cheap enough?"
In addition to being expensive, titanium is NOT easy to work with.
With 802.11g you can only serve a total of 45 users. 15 users per AP times 3 non-overlapping 11g channels = 45.
In theory you can use highend equipment with transmit power management to set up an effective microcellular network with more APs. If you want to do this, ALL hardware (clients and APs) must support automatic transmit power management. Multiple low-power APs will let you increase density, but this is difficult and expensive.
802.11a gives you 20 nonoverlapping channels to work with, but has its own issues to deal with.
Once you're talking about 802.11a gear or high-end properly placed/installed 802.11g gear for a high density network, you're better off just installing gigabit Ethernet switches and Cat5e and getting (easily) 10 times the real-world throughput or more.
"Speed. As far as I know, pre-N technology hasn't been fully adopted and the best you can do is 802.11g (54Mbps) basically half of what you would get with wired (100Mbps). Granted you rarely ever get the full 100Mbps, but you rarely ever get the full 54Mbps either."
It's worse than that. The CSMA/CA collision management protocol used by 802.11 is inherently less efficient than CSMA/CD used by wired Ethernet. The throughput of an 802.11 system will always be a lower fraction of the signaling rate than even shared Ethernet via a hub. Switched Ethernet? No comparison. Switched Ethernet can easily reach 90-95% of the signaling rate in terms of real throughput. 802.11 (any variant) is lucky to get 50% to a single user. With multiple users and small packets, this drops significantly.
"Rootkits CANNOT be reliably cleaned up."
False.
"booting from known-good media, and either testing MD5 sums of each executable, or a wipe/re-install."
You just contradicted yourself. There's nothing preventing someone from creating a rootkit-hunting LiveCD.
Yes, there are definately major issues to be dealt with for some sort of multicast support for BT (such as, for example, the fact that torrent users don't all have the same downstream bandwidth, and even when multicast support starts appearing in ISPs it'll take a long time for full penetration so the "legacy" way needs to be supported.) There's also the TCP streams vs. UDP datagrams issue.
It would actually be easier to implement this as a seperate protocol that was designed to augment BitTorrent. (i.e. getting data both via BitTorrent and via a multicast mechanism.) Either way, end-to-end multicast support on the public Internet would contribute to a massive reduction in backbone load from P2P, which has in the past been claimed to consist of 30-50% of all backbone traffic now. (I don't know how accurate this number really is though.)
If I recall correctly, supporting IP Multicast was required instead of optional for IPv6 implementations.
If this is the case, a multicast-aware BitTorrent would be THE killer app, IF IPv6 were deployed sufficiently for multicast torrents to be effective.
The way things are now, a multicast torrent would be pretty much the same trafficwise as the way things currently are for the backbone, since for the most part everyone is tunneling to one of a small handful of IPv6 brokers.
Not necessarily, because even some of the name brands can sometimes be borderline quality.
For example, Memorex media is almost always made by CMC Magnetics. In my experience, while they're far better than the likes of Princo, they are not a manufacturer I will trust. Even many of the generic noname brands are better than CMC-sourced discs, as in my experience anything made by Ritek is far better than CMC media, and Riteks are almost never sold as name-brand.
Verbatim media used to be high-quality almost all the time, but I heard that they started selling some batches of CMCs, so you never knew what you were getting.
Only a few of the name brands are guaranteed to give you the same original manufacturer every time because they ARE the original manufacturer. Off of the top of my head, the only two examples of this I can think of are Sony and TDK.
British or U.S, unless you've badly screwed up and didn't do any proofreading or you are making a comedy spoof of vampire movies, no one ever drives steaks through anyone's heart. It's STAKE, and the editors of this rag didn't even catch such an obvious error.
If you don't know what I'm talking about, RTFA.
I'm inclined to disagree with that. I just finished my M.S. at Rutgers, where the majority of the students in the graduate program were either Chinese or Indian. (The lack of U.S. citizens in a graduate program at a state university where tuition is DIRT CHEAP for state residents and pretty inexpensive for others says bad things about U.S. attitudes towards higher education...) The Chinese students, for the most part, barely spoke English at all and I'm surprised any of them were able to complete their classwork given the difficulty of understanding and communicating with their professors and fellow classmates. The Indian students had at least been educated reasonably well in English so that they were understandable. The problem is that while they may have been educated in English since the 11th grade, they were most likely educated in English by a non-native English speaker. Think of the game of Telephone, where after being passed from person to person, a message is distorted so much as to be barely understandable. It's same thing with "second generation" (or more) English language education. Yes, it was possible to communicate with the Indian students and much easier than the Chinese students, but it was still *extremely* difficult due to the thick accents. Yes, even if one's grammar is perfect, a thick accent can make verbal communication extremely difficult. By the way, most of the Indian students in question could write extremely well. It was only their spoken language that was difficult to understand. Unfortunately, even in technical fields, verbal communication is important. In technical support fields where outsourcing is currently the most common, verbal communications is *EVERYTHING*.
You snuck one rather BAD thing in your list of things that are good about Cuba - Infant mortality is a Bad Thing.
:)
That said, after seeing Yank Tanks - I am amazed at the resourcefulness of some of Cuba's citizens. If the US had auto mechanics as competent as the Cubans, our auto industry would go out of business because cars would last so damn long.
The SanDisk units do not use mechanical hard drives, but use solid state flash memory. Thus their direct competitor is the iPod nano, Apple's solid-state player.
iPod nano 2GB - $199
iPod nano 4GB - $249
iPod nano 6GB - does not exist
The 2 and 4 GB SanDisk variants are $20 cheaper. Not sure how they compare on features, but in terms of price per GB, the SanDisks beat the iPod nano.
Obviously you didn't put much effort into figuring out what these satellites do.
:)
For example, if you click on Cornell's ICE CubeSat, it takes you to the AMSAT info page for the sat, which has a link to Cornell's own page for the sat, which has LOTS of details on the design of the sat, and more importantly, the science package the sat is carrying. Most of the other university sats are also carrying some sort of science package (most of them are cameras I believe, but I'm not sure.)
To summarize it in basic terms, the Cornell ICE Cube is designed to take measurements of ionospheric phenomenon that have been causing problems with the GPS system in certain parts of the world at certain times. The phenomenon is known as scintillation and causes rapid fluctuations in the signal strength of GPS signals reaching the ground. ICE Cube is meant to take measurements of such phenomenon at high altitudes, eliminating any effects the lower layers of the atmosphere might have on the signal.
This is all assuming that the comm system works... I designed the first revision of the onboard transciever and I admit I didn't do the best job due to lack of time/experience. (I was a senior with only one semester to work on it.) From what I've heard the grad students assigned to clean up my mess did a pretty good job though.
"unless you mean because they're violating someone elses and afraid to get caught. Patents are, by definition and law, public disclosures."
You forget that it's possible to LICENSE someone else's patents. That is, you are given permission to legally use their patents under certain conditions.
For whatever reason, those conditions have apparently required closed source drivers. I don't know about NVidia's excuse, but this was ATI's reasoning for no longer providing full specifications of their chipsets.
The article probably is mistakenly referring to the GPU chips on the card and their support circuitry as seperate boards, when in fact they are not.
Pulling teeth?
I've never had any problems getting any NVidia GPU to work flawlessly under Linux.
Yeah, the closed-source aspect of NVidia's drivers may be annoying to some, but I don't mind closed-source drivers if they're high enough quality, and NVidia's drivers are one of the few examples of closed source software with high quality. (The associated closed-source games for Linux, specifically iD Software's products, comprise most of the other examples...)
Face it - due to patent issues out of the chipset manufacturers' control (classic example being S3 Texture Compression - S3TC was the beginning of ATI's transition from fully documented open source drivers to binary-only drivers with the open-source versions lacking critical features for 3D gaming), no chipset manufacturer can release open source drivers that support their card's full feature set, unless their card's feature set is massively crippled. (See Intel GMA-series integrated graphics as an example.)
It's two GPUs *on a single card*.
Not two seperate cards that require seperate PCI Express slots.
Normally I would trust the company's own site the most as far as publicity and propaganda, but obviously they are not listing many of their older investments. Perhaps they may no longer have a stake in any of those other companies listed on the Wikipedia page.