Target is getting serious about Linux laptops. Looking under "Laptops",
the first screen of "Featured items" has 3 colors of the EE PC, Linux
version, and some HP laptop. The XP version of the EE PC costs $100 more.
There's a clear future in little laptop machines that don't cost much and don't use much power, yet are powerful enough to do most of the things most users do with laptops. The low end x86 CPUs are finally good enough to power such machines.
The laptop manufacturers had a Detroit mentality of "more computer per computer". This kept laptop prices up and margins high. But, as it turns out, cramming enough CPU power into a laptop to run wind tunnel simulations isn't what users really need. Especially when the network connection is the bottleneck anyway. The actual uses for a 4-CPU laptop are somewhat limited.
The flood of low-cost laptops has just started. The EE PC set off a race for the bottom. In a year or two, laptops will come in blister packs at the drugstore, in the section with the calculators, electronic dictionaries, and other office supplies. From here on, it's all about lowering margins. Intel and Microsoft will be squeezed hard on price.
You can't set math on a Linotype. Math had to be hand-set, and this was not easy. Formulas had to be aligned with little brass spacers (what typesetters called "furniture"). Text that couldn't be machine-set was called "penalty copy", and involved large extra charges.
There are plenty of conformal coatings for this sort of thing. Fin-L-Kote is about $20/can.
I've used that on PC boards in robotics applications.
Automotive electronics are routinely conformal coated, usually with a rather thick coating.
The main problem is electrical contacts. Those have to be masked when the electronics are sprayed or dipped. The military/marine solution is gold-plated contacts, with everything else conformal coated.
This isn't going to work for something with a vented hard drive. Sealed hard drives are available for laptops.
Something's not right there. Cuil is missing obvious search hits. Searching on phrases and for multiple words comes up empty for many phrases that generate hits on Google. It even misses phrases on sites that it has definitely indexed.
Either there's an out and out bug, or there's a level of effort limit on difficult joins.
Can't tell which as yet.
Applied Materials, the largest maker of semiconductor fab machinery, makes fab gear for solar panels. Their CEO likes to show graphs of cost per watt vs. year, and there's a steady decline, at roughly the same rate as LCD panels. Applied Materials solar cell fabs are using technology borrowed from LCD panel fab, and they're now making 5 square meters of panel at a time. The machinery for manufacturing such huge panels is appropriately large, and that's part of what's bringing the cost down. Despite much hype, no single improvement has produced a big drop in panel cost. But the cumulative effect of continuous improvement is working.
Applied Materials people make the point that installation is now half the cost of the completed solar system, and the solar industry needs to move beyond the "guy with a pickup truck" level of installation. Bigger panels reduce installation cost, and they're working on panels that are roofs themselves, instead of being installed on top of roofs.
The actual rate of price drop is maybe a factor of 2 per decade. Which isn't bad. As the Applied Materials solar division head says, "This is a great business. Everybody else's costs are going up, and ours are going down. And we're nowhere near market saturation."
Well, that's progress. I've criticized SETI@Home for looking for "carriers" signals with a large fixed-frequency component. They need to get beyond that. AM and FM signals have carriers (Analog TV is AM video with an FM audio subcarrier), and as a result, 80% of the signal energy is wasted. None of the more modern digital transmission systems have strong carriers.
The more efficient a transmission system, the more it looks like white noise if you don't know how to decode it. If there's some big repetitive component like a carrier, or the horizontal and vertical retrace intervals in analog TV, it's inefficient. The FCC wouldn't approve any new transmission system which wasted bandwidth like that, and the old ones that do are being phased out.
So SETI systems that look for carriers are looking for civilizations advanced enough to generate high-power RF signals, but not advanced enough to use more efficient digital modes. Our civilization went through that period in under a century. It's also fairly clear that nobody in our stellar neighborhood is continuously sending a strong RF carrier in our direction; that's been looked for.
Question: can the new SETI algorithm pick up an HDTV broadcast station?
I've been fortunate enough to work on most of the big problems in computer science at one time or another. CPU scheduling. Network congestion. Compiler optimization. Proof of correctness. Secure operating systems. Image processing. Mobile robotics. Game physics. I've done very well financially. I have an advanced degree from one of the big-name schools. So I can't complain personally. Going into computer science worked out very well for me.
But I'm from the previous generation of programmers. Programming today is mostly about dealing with yet another API with another thousand or so interfaces, some of which work. By the time you're dealing with the fifteenth system for putting widgets on a screen and processing incoming events, you get fed up. Especially since you can see all the ways in which the new ones got something wrong that was a solved problem a decade or two ago.
Most of the basic algorithmic problems have been solved. Not only have they been solved, the solutions have been packaged up so that you don't have to look at them. How often do you really need to open Knuth any more?
Computing is the "stationary engineering" of today. About a century ago, stationary engineering, the work of, literally, keeping the wheels of industry turning, reached this point.
In 1870 or so, stationary engineering was a growth job that needed smart people who understood the details of engines, generators, and steam. The basics of the field were still being figured out. Counterflow boilers and carbon commutator brushes were hot R&D topics. Just getting the machinery to work at all was tough, and there were serious reliability problems.
By 1910 or so, most of the hard problems had been solved; big steam turbines and generators were working reliably, and plant operation didn't require much innovative thinking. Today, stationary engineering is a union job that few people even realize exists.
Computing is now about where stationary engineering was in 1910. Everything pretty much works, and most of what's going on is not that innovative. We're now dealing with scaling issues, which is where electric power was in 1910.
There's interesting stuff going on in robotics, parts of AI, statistical methods, and the handling of very large databases. We need small numbers of smart people to push that forward.
There are areas of software engineering that need real engineering talent, like video compression and graphics, although such narrow, well-bounded problems tend to move into hardware.
But we no longer need computer scientists just to run a data center or to set up business applications and web sites. Just careful, well-trained technicians.
That's what employers want, and that's what most of the students want. Most of the schools are willing to accommodate them.
It's a difficult issue. I have a dedicated server at APlus in Phoenix, and for the first six months, they didn't have any of the passwords for the box. Then they had a big outage and had to move the servers to another data center, and asked the users to tell them the root password so the could shut down the server, move it, and reconfigure the networking. So now they have the root password, and they did use it once without asking me first when I called in with a later problem.
It's not a big issue for this particular application, because it doesn't have any proprietary or personal data and it doesn't do credit card transactions. But for anyone selling something, it could be a very big deal.
This is to some extent a lack of Linux system administration capability. There's no standard way to give out a permission that allows only the operations a co-location facility might need to perform - startup, shutdown, IP address change, and maybe encrypted backup. APlus uses the Plesk control panel, which can do most of those things, but its security isn't designed to give the co-location operator a limited login.
Every time some big player starts pushing Linux, Microsoft makes it go away. Wal-Mart has sold Linux machiens twice, then backed off; they no longer seem to sell any Linux machines. Fry's has stopped selling Linux machines.
Go to the Dell site and try to find a Linux laptop. It's quite hard to find. Even when you finally find the Ubuntu page, for which you will probably have to search, the first thing you see is "Not sure Open Source is for You? The main thing to note is that when you choose open source you don't get a Windows® operating system. If you're here by mistake and you are looking for a Dell PC with Windows, please use the following link..."
You never get to a Linux PC from the "laptop selector", even though Dell does offer them.
Google StreetView now has all of the major U.S. cities covered. Except the Washington, D.C. area.
Of the top forty metropolitan areas in the US, Google has all of them covered except #8, the Washington D.C. area, and #20, the Baltimore area. There's no StreetView data for a 75-mile radius around Washington. They've covered Wilmington, DE and Richmond, VA, both about 100 miles from Washington, but that's as close as they get.
They're working on rural areas of California. They've worked down to Knoxville, TN, Greenville, NC, and Boise, IH. So it can't be accidental that they've avoided Washington.
True. Standard DVDs upconverted to 1080p in the DVD player and fed across an HDMI link look about as good as Blu-Ray discs. They look better than real HD signals that have passed through analog conversion on their way to the display.
But how many of those trillion pages have unique, useful content? E-mail is over 95% spam, and the web is getting there.
There were about 153 million registered domains at the beginning of the year. The ones from the spam-friendly registrars are mostly junk. Tim Bernars-Lee said in 2006 that web junk was becoming a major problem, and it's become worse since then.
If you throw out all the anonymous but commercial domains (we call them "bottom-feeders"), as we do with SiteTruth, the Web looks a lot better. Search engines are getting stricter about this. You don't see that many "landing pages" in Google any more. Bad news for companies like Marchex, the publicly traded web spammer that cranks out all those junk "What you need, when you need it" sites.
"The mass trials are going well. There will be fewer Russians, but better ones." - Greta Garbo in Ninotchka.
Having run a DARPA Grand Challenge team, I've been through most of this line of reasoning.
I'm rather less optimistic.
First, Templeton writes "The cost of accidents is arguably the single largest component of the per-mile cost of driving a vehicle", but doesn't provide justification for that statement. Total US gasoline consumption costs about $600 billion per year. The American Automobile Association says that US auto accidents cost about $164 billion per year.
Second, while we can do automatic driving in a situation where all the players are reasonably well-behaved vehicles, we're a long way from being able to do it safely in a populated area.
Today's robot vehicle technologies have minimal "situational awareness". That's one of the hardest problems in AI. Right now, sensing systems are up to recognizing "obstacle" and "moving car-like thing". Pedestrian and bicyclist behavior prediction is a ways off.
The whole section on robot vehicles with incredible evasive ability is bogus. Vehicles are limited by inertia and maneuvering room. Cutting the reaction time from 500ms to 50ms would help some. Half of all collisions would be prevented if braking started 500ms sooner, according to a Mercedes study. Chain collisions are an artifact of human reaction time; with minimal inter-vehicle coordination, all the cars in a lane could come to a fast stop without colliding.
But evasive action requires room.
Most of the estimates of huge savings come not from automatic driving but from electric cars.
Especially little lightweight electric cars. You can get little electric cars now; I'm in Silicon Valley and I see them now and then. But they're about as common as Segways.
Zipcar indicates that the car sharing concept can work. With automatic driving, the car could be delivered to you, so it could be used in less-dense areas than central cities. But it's really for people who only need a car occasionally. Zipcar is $10/hour.
You're describing Demo 97. That had serious industrial and government backing, it worked, and it went nowhere.
It was essentially a trackway scheme. Cars followed markers (permanent magnets) in the pavement, could measure the distance to cooperating cars ahead and behind, and had some minimal radar-based obstacle detection. It required automatic-only dedicated lanes.
On the Denver FBI's page for the escape, the picture of Davidson is a dead link. (There are press photos, but a Government picture can be put into Wikipedia.) It's embarrassing for the FBI to issue a wanted notice for a prison break with no picture.
The Denver FBI office doesn't seem to have a contact e-mail address.
Blogger is popular for spam redirects, because it's possible to turn a Blogger page into a redirect. Typical example: "Looking for a R0lex repl1ca?... Where? At http://www.mitch83393.blogspot.com/" (Google already got this one as a TOS violation, but they're throwaway blogs generated by programs. There will be a new one in a few minutes.) Spammers do this to get their message through filters that check for spam links.
This is a generic problem with Google's free services. Spammers and scammers now use GMail to get throwaway mail accounts, Blogger for an open redirector, YouTube to host advertising videos, AdWords to advertise scams, and Google Checkout to collect the money. It's full-service evil.
For the last two, Google has a business relationship, but doesn't seem to be validating their customers well enough. The use of Google Checkout for spam and attack tools is especially disturbing. Try, for example, searching for "craiglist posting". Note the ads with Google Checkout links. There, Google is an active participant in collecting the money and is profiting from the transaction.
The original article was about ads on social networking sites, which have very low value.
This has been discussed over on Search Engine Watch. Google AdWords has "exclusion lists", lists of sites where you don't want your ads to appear. If you have expensive per-click ads, adding MySpace and Facebook to the exclusion list cuts your ad cost without impacting revenue much.
You don't want your ad for expensive watches or mortgage refinancing on MySpace; Google will make money from you, but you won't make money.
Remember, 10% of the users produce 50% of the clicks but don't buy much.
That 10% seem to be heavy social networking site users. You can buy their clicks, but they don't buy your product. This is a huge money drain on advertisers. There's much discussion in advertiser forums like Search Engine Watch about what to do about this. Tools have been developed to help advertisers filter out the non-producing ad sites. Google was fighting this; their terms of service prohibit AdWords advertisers from exchanging ad performance data. But ways have been developed to do it anyway.
The result of this has been 1) the price of ads on social networking sites is very low, and 2) the ad quality on them is terrible. We track this with SiteTruth AdRater. If you install AdRater (a Firefox plug-in, requires Greasemonkey), a rating icon appears atop each Google ad.
Try this on Myspace, and almost all the ads will come up with a red "do-not-enter" sign. Then try, say, Bloomberg, and you'll see plenty of legit companies. The contrast is striking.
Myspace ads seem to be mostly links to ad farms, bottom feeder dating services (one is telling me "Three of your friends have a crush on you", even though I'm not logged into Myspace), and similar junk. Myspace just did a site redesign, and there are far fewer Google ad slots than before. This probably reflects unsold inventory.
Traffic alone is not enough. The users have to buy. Advertisers have now figured this out.
This could easily turn into spam. The problem is that this may not be considered a "call" under the Do Not Call list rules for cell phones. It might be legal to spam via this route. Uh oh.
Re:Eliminating the need for server virtualization
on
IT Jobs To Drop In 2009
·
· Score: 3, Interesting
Since we like to partition out at least one server per app since app vendors like to point fingers at each other....
That's the problem. Ask "Can you install this app as an ordinary user, without root privileges?"
If the answer is no, then the TCO of the application just went way up.
After decades of annoying flicker, strobing, and bad pans at 24FPS, we finally got LCD panels that don't flicker at all. Some monitors even time-interpolate to get the frame rate up, and framefree compression is just starting to work. Now people want to crud it up with alternating-frame stereo. Bleah.
Stereo vision doesn't really do much beyond about 4m or so, and it scales badly for anything that isn't its real size.
There are some very cute 3D systems that are sensitive to head position, so you can move your head and have the scene adjust accordingly. But that doesn't work in theaters.
"Server virtualization" is an admission that system administration is badly designed. After all, you're not going to get any more work out of the machine than you would running multiple processes. Usually, you get less.
Part of the problem is that Linux is still locked into the old UNIX user/group/everybody model of security, with an all-powerful "root". Virtualization is a way of working around that limitation.
OpenVZ and Linux-VServer are efforts to get around this problem by adding another level of administrative containerization. The performance is better, since you're not going through two layers of operating systems.
Many of the problems come from the fact that some major applications are coded as singletons.
For example, Apache assumes there is only one instance of itself per machine. (Yes, this can be worked around, but it's not easy.) So do most mail handling programs. If you can install and run an application without running as "root", it probably doesn't need virtualization.
Slashdot Mirror
Just click here and follow the official U.S. Government approved directions.
Target is getting serious about Linux laptops. Looking under "Laptops", the first screen of "Featured items" has 3 colors of the EE PC, Linux version, and some HP laptop. The XP version of the EE PC costs $100 more.
There's a clear future in little laptop machines that don't cost much and don't use much power, yet are powerful enough to do most of the things most users do with laptops. The low end x86 CPUs are finally good enough to power such machines.
The laptop manufacturers had a Detroit mentality of "more computer per computer". This kept laptop prices up and margins high. But, as it turns out, cramming enough CPU power into a laptop to run wind tunnel simulations isn't what users really need. Especially when the network connection is the bottleneck anyway. The actual uses for a 4-CPU laptop are somewhat limited.
The flood of low-cost laptops has just started. The EE PC set off a race for the bottom. In a year or two, laptops will come in blister packs at the drugstore, in the section with the calculators, electronic dictionaries, and other office supplies. From here on, it's all about lowering margins. Intel and Microsoft will be squeezed hard on price.
No, not the program, the Linotype.
You can't set math on a Linotype. Math had to be hand-set, and this was not easy. Formulas had to be aligned with little brass spacers (what typesetters called "furniture"). Text that couldn't be machine-set was called "penalty copy", and involved large extra charges.
There are plenty of conformal coatings for this sort of thing. Fin-L-Kote is about $20/can. I've used that on PC boards in robotics applications. Automotive electronics are routinely conformal coated, usually with a rather thick coating.
The main problem is electrical contacts. Those have to be masked when the electronics are sprayed or dipped. The military/marine solution is gold-plated contacts, with everything else conformal coated.
This isn't going to work for something with a vented hard drive. Sealed hard drives are available for laptops.
Something's not right there. Cuil is missing obvious search hits. Searching on phrases and for multiple words comes up empty for many phrases that generate hits on Google. It even misses phrases on sites that it has definitely indexed.
Either there's an out and out bug, or there's a level of effort limit on difficult joins. Can't tell which as yet.
Applied Materials, the largest maker of semiconductor fab machinery, makes fab gear for solar panels. Their CEO likes to show graphs of cost per watt vs. year, and there's a steady decline, at roughly the same rate as LCD panels. Applied Materials solar cell fabs are using technology borrowed from LCD panel fab, and they're now making 5 square meters of panel at a time. The machinery for manufacturing such huge panels is appropriately large, and that's part of what's bringing the cost down. Despite much hype, no single improvement has produced a big drop in panel cost. But the cumulative effect of continuous improvement is working.
Applied Materials people make the point that installation is now half the cost of the completed solar system, and the solar industry needs to move beyond the "guy with a pickup truck" level of installation. Bigger panels reduce installation cost, and they're working on panels that are roofs themselves, instead of being installed on top of roofs.
The actual rate of price drop is maybe a factor of 2 per decade. Which isn't bad. As the Applied Materials solar division head says, "This is a great business. Everybody else's costs are going up, and ours are going down. And we're nowhere near market saturation."
Well, that's progress. I've criticized SETI@Home for looking for "carriers" signals with a large fixed-frequency component. They need to get beyond that. AM and FM signals have carriers (Analog TV is AM video with an FM audio subcarrier), and as a result, 80% of the signal energy is wasted. None of the more modern digital transmission systems have strong carriers.
The more efficient a transmission system, the more it looks like white noise if you don't know how to decode it. If there's some big repetitive component like a carrier, or the horizontal and vertical retrace intervals in analog TV, it's inefficient. The FCC wouldn't approve any new transmission system which wasted bandwidth like that, and the old ones that do are being phased out.
So SETI systems that look for carriers are looking for civilizations advanced enough to generate high-power RF signals, but not advanced enough to use more efficient digital modes. Our civilization went through that period in under a century. It's also fairly clear that nobody in our stellar neighborhood is continuously sending a strong RF carrier in our direction; that's been looked for.
Question: can the new SETI algorithm pick up an HDTV broadcast station?
I've been fortunate enough to work on most of the big problems in computer science at one time or another. CPU scheduling. Network congestion. Compiler optimization. Proof of correctness. Secure operating systems. Image processing. Mobile robotics. Game physics. I've done very well financially. I have an advanced degree from one of the big-name schools. So I can't complain personally. Going into computer science worked out very well for me.
But I'm from the previous generation of programmers. Programming today is mostly about dealing with yet another API with another thousand or so interfaces, some of which work. By the time you're dealing with the fifteenth system for putting widgets on a screen and processing incoming events, you get fed up. Especially since you can see all the ways in which the new ones got something wrong that was a solved problem a decade or two ago.
Most of the basic algorithmic problems have been solved. Not only have they been solved, the solutions have been packaged up so that you don't have to look at them. How often do you really need to open Knuth any more?
Computing is the "stationary engineering" of today. About a century ago, stationary engineering, the work of, literally, keeping the wheels of industry turning, reached this point. In 1870 or so, stationary engineering was a growth job that needed smart people who understood the details of engines, generators, and steam. The basics of the field were still being figured out. Counterflow boilers and carbon commutator brushes were hot R&D topics. Just getting the machinery to work at all was tough, and there were serious reliability problems. By 1910 or so, most of the hard problems had been solved; big steam turbines and generators were working reliably, and plant operation didn't require much innovative thinking. Today, stationary engineering is a union job that few people even realize exists.
Computing is now about where stationary engineering was in 1910. Everything pretty much works, and most of what's going on is not that innovative. We're now dealing with scaling issues, which is where electric power was in 1910.
There's interesting stuff going on in robotics, parts of AI, statistical methods, and the handling of very large databases. We need small numbers of smart people to push that forward. There are areas of software engineering that need real engineering talent, like video compression and graphics, although such narrow, well-bounded problems tend to move into hardware. But we no longer need computer scientists just to run a data center or to set up business applications and web sites. Just careful, well-trained technicians.
That's what employers want, and that's what most of the students want. Most of the schools are willing to accommodate them.
It's a difficult issue. I have a dedicated server at APlus in Phoenix, and for the first six months, they didn't have any of the passwords for the box. Then they had a big outage and had to move the servers to another data center, and asked the users to tell them the root password so the could shut down the server, move it, and reconfigure the networking. So now they have the root password, and they did use it once without asking me first when I called in with a later problem.
It's not a big issue for this particular application, because it doesn't have any proprietary or personal data and it doesn't do credit card transactions. But for anyone selling something, it could be a very big deal.
This is to some extent a lack of Linux system administration capability. There's no standard way to give out a permission that allows only the operations a co-location facility might need to perform - startup, shutdown, IP address change, and maybe encrypted backup. APlus uses the Plesk control panel, which can do most of those things, but its security isn't designed to give the co-location operator a limited login.
Every time some big player starts pushing Linux, Microsoft makes it go away. Wal-Mart has sold Linux machiens twice, then backed off; they no longer seem to sell any Linux machines. Fry's has stopped selling Linux machines.
Go to the Dell site and try to find a Linux laptop. It's quite hard to find. Even when you finally find the Ubuntu page, for which you will probably have to search, the first thing you see is "Not sure Open Source is for You? The main thing to note is that when you choose open source you don't get a Windows® operating system. If you're here by mistake and you are looking for a Dell PC with Windows, please use the following link..."
You never get to a Linux PC from the "laptop selector", even though Dell does offer them.
Google StreetView now has all of the major U.S. cities covered. Except the Washington, D.C. area. Of the top forty metropolitan areas in the US, Google has all of them covered except #8, the Washington D.C. area, and #20, the Baltimore area. There's no StreetView data for a 75-mile radius around Washington. They've covered Wilmington, DE and Richmond, VA, both about 100 miles from Washington, but that's as close as they get.
They're working on rural areas of California. They've worked down to Knoxville, TN, Greenville, NC, and Boise, IH. So it can't be accidental that they've avoided Washington.
One wonders why.
It's like an amateur version of Speed Racer. It looks like something somebody put together with Blender.
As a driving game, it looks like fun. As a movie, probably not.
True. Standard DVDs upconverted to 1080p in the DVD player and fed across an HDMI link look about as good as Blu-Ray discs. They look better than real HD signals that have passed through analog conversion on their way to the display.
But how many of those trillion pages have unique, useful content? E-mail is over 95% spam, and the web is getting there.
There were about 153 million registered domains at the beginning of the year. The ones from the spam-friendly registrars are mostly junk. Tim Bernars-Lee said in 2006 that web junk was becoming a major problem, and it's become worse since then.
If you throw out all the anonymous but commercial domains (we call them "bottom-feeders"), as we do with SiteTruth, the Web looks a lot better. Search engines are getting stricter about this. You don't see that many "landing pages" in Google any more. Bad news for companies like Marchex, the publicly traded web spammer that cranks out all those junk "What you need, when you need it" sites.
"The mass trials are going well. There will be fewer Russians, but better ones." - Greta Garbo in Ninotchka.
Having run a DARPA Grand Challenge team, I've been through most of this line of reasoning. I'm rather less optimistic.
First, Templeton writes "The cost of accidents is arguably the single largest component of the per-mile cost of driving a vehicle", but doesn't provide justification for that statement. Total US gasoline consumption costs about $600 billion per year. The American Automobile Association says that US auto accidents cost about $164 billion per year.
Second, while we can do automatic driving in a situation where all the players are reasonably well-behaved vehicles, we're a long way from being able to do it safely in a populated area. Today's robot vehicle technologies have minimal "situational awareness". That's one of the hardest problems in AI. Right now, sensing systems are up to recognizing "obstacle" and "moving car-like thing". Pedestrian and bicyclist behavior prediction is a ways off.
The whole section on robot vehicles with incredible evasive ability is bogus. Vehicles are limited by inertia and maneuvering room. Cutting the reaction time from 500ms to 50ms would help some. Half of all collisions would be prevented if braking started 500ms sooner, according to a Mercedes study. Chain collisions are an artifact of human reaction time; with minimal inter-vehicle coordination, all the cars in a lane could come to a fast stop without colliding. But evasive action requires room.
Most of the estimates of huge savings come not from automatic driving but from electric cars. Especially little lightweight electric cars. You can get little electric cars now; I'm in Silicon Valley and I see them now and then. But they're about as common as Segways.
Zipcar indicates that the car sharing concept can work. With automatic driving, the car could be delivered to you, so it could be used in less-dense areas than central cities. But it's really for people who only need a car occasionally. Zipcar is $10/hour.
You're describing Demo 97. That had serious industrial and government backing, it worked, and it went nowhere.
It was essentially a trackway scheme. Cars followed markers (permanent magnets) in the pavement, could measure the distance to cooperating cars ahead and behind, and had some minimal radar-based obstacle detection. It required automatic-only dedicated lanes.
On the Denver FBI's page for the escape, the picture of Davidson is a dead link. (There are press photos, but a Government picture can be put into Wikipedia.) It's embarrassing for the FBI to issue a wanted notice for a prison break with no picture.
The Denver FBI office doesn't seem to have a contact e-mail address.
Blogger is popular for spam redirects, because it's possible to turn a Blogger page into a redirect. Typical example: "Looking for a R0lex repl1ca? ... Where? At http://www.mitch83393.blogspot.com/" (Google already got this one as a TOS violation, but they're throwaway blogs generated by programs. There will be a new one in a few minutes.) Spammers do this to get their message through filters that check for spam links.
This is a generic problem with Google's free services. Spammers and scammers now use GMail to get throwaway mail accounts, Blogger for an open redirector, YouTube to host advertising videos, AdWords to advertise scams, and Google Checkout to collect the money. It's full-service evil.
For the last two, Google has a business relationship, but doesn't seem to be validating their customers well enough. The use of Google Checkout for spam and attack tools is especially disturbing. Try, for example, searching for "craiglist posting". Note the ads with Google Checkout links. There, Google is an active participant in collecting the money and is profiting from the transaction.
The original article was about ads on social networking sites, which have very low value. This has been discussed over on Search Engine Watch. Google AdWords has "exclusion lists", lists of sites where you don't want your ads to appear. If you have expensive per-click ads, adding MySpace and Facebook to the exclusion list cuts your ad cost without impacting revenue much. You don't want your ad for expensive watches or mortgage refinancing on MySpace; Google will make money from you, but you won't make money.
Remember, 10% of the users produce 50% of the clicks but don't buy much. That 10% seem to be heavy social networking site users. You can buy their clicks, but they don't buy your product. This is a huge money drain on advertisers. There's much discussion in advertiser forums like Search Engine Watch about what to do about this. Tools have been developed to help advertisers filter out the non-producing ad sites. Google was fighting this; their terms of service prohibit AdWords advertisers from exchanging ad performance data. But ways have been developed to do it anyway.
The result of this has been 1) the price of ads on social networking sites is very low, and 2) the ad quality on them is terrible. We track this with SiteTruth AdRater. If you install AdRater (a Firefox plug-in, requires Greasemonkey), a rating icon appears atop each Google ad. Try this on Myspace, and almost all the ads will come up with a red "do-not-enter" sign. Then try, say, Bloomberg, and you'll see plenty of legit companies. The contrast is striking.
Myspace ads seem to be mostly links to ad farms, bottom feeder dating services (one is telling me "Three of your friends have a crush on you", even though I'm not logged into Myspace), and similar junk. Myspace just did a site redesign, and there are far fewer Google ad slots than before. This probably reflects unsold inventory.
Traffic alone is not enough. The users have to buy. Advertisers have now figured this out.
This could easily turn into spam. The problem is that this may not be considered a "call" under the Do Not Call list rules for cell phones. It might be legal to spam via this route. Uh oh.
Since we like to partition out at least one server per app since app vendors like to point fingers at each other....
That's the problem. Ask "Can you install this app as an ordinary user, without root privileges?" If the answer is no, then the TCO of the application just went way up.
After decades of annoying flicker, strobing, and bad pans at 24FPS, we finally got LCD panels that don't flicker at all. Some monitors even time-interpolate to get the frame rate up, and framefree compression is just starting to work. Now people want to crud it up with alternating-frame stereo. Bleah.
Stereo vision doesn't really do much beyond about 4m or so, and it scales badly for anything that isn't its real size.
There are some very cute 3D systems that are sensitive to head position, so you can move your head and have the scene adjust accordingly. But that doesn't work in theaters.
"Server virtualization" is an admission that system administration is badly designed. After all, you're not going to get any more work out of the machine than you would running multiple processes. Usually, you get less. Part of the problem is that Linux is still locked into the old UNIX user/group/everybody model of security, with an all-powerful "root". Virtualization is a way of working around that limitation.
OpenVZ and Linux-VServer are efforts to get around this problem by adding another level of administrative containerization. The performance is better, since you're not going through two layers of operating systems.
Many of the problems come from the fact that some major applications are coded as singletons. For example, Apache assumes there is only one instance of itself per machine. (Yes, this can be worked around, but it's not easy.) So do most mail handling programs. If you can install and run an application without running as "root", it probably doesn't need virtualization.
Additionally, when I visited Tribe.com, I received a 404...
Try "tribe.net. Looks like somebody forgot to update "tribe.com", which "tribe.net" apparently owns but is in domain hold for bogus Whois info.