According to article, sabotage of critical software poses a significant risk:
IT security consulting firm Cyber Defense Agency (CDA) has warned the US military, government and "critical infrastructure agencies" against using outsourced commercial software which could be tampered with by terrorists.
..."Outsourced commercial software poses a silent but significant security risk to the defence and welfare of the US," says Sami Saydjari, president of CDA. "The chances of strategic damage from a cyber-terrorist attack on the US increases the longer it takes to remedy the risks posed by outsourced software."
Never mind pulling the network cable. The vector may be lurking in the code, waiting for the time or circumstances that launches the attack. This is a way to attack "secure" systems without external connections, which is how critical systems generally will be implemented. If you can't trust the code, you can't trust the system made from it.
Cordless power tools are exempt from the cadmium restriction, so they can continue using NiCd cells. But if I read the article correctly, they must be removable and collected for recycling when you buy the replacement battery. More information and regulation history is available at the EU web site. On the whole, this is very much in line with the RoHS and WEEE directives. It's surprising they delayed implementation for as long they have.
At college age, contracting mumps is no joke. Mom was right to let us "get it over with" in grade school, back in those prehistoric, pre-MMR vaccine days. From the Wiki:
While symptoms are generally not severe in children, in teenagers and adults, the symptoms can be more severe and complications such as infertility or subfertility are relatively common, although still rare in absolute terms.
Very true! Old-timers who started on CP/M etc. had this first reaction to MS/DOS: "It isn't quite right". The syntax and command names were almost the same, but not quite. What we already knew was "right", and anything else was "wrong". The same applies to Windows vs. any older version of Windows vs. Mac OS vs. unix x-terminals vs. Gnome vs. KDE...you get the idea. What you learned first is imprinted in your bain as "the way it's supposed to be"; anything that makes you deviate from what you "know" to be right is unpleasant.
This isn't about "right" or "wrong", it's about human nature. What you learn first is what you will be comfortable with. If your first system uses linux, oo.o, and gimp you'll be comfortable using them. If your first system has Vista, MS Office and Photoshop then those will be your reference point. Any change will be uncomfortable, and it is human nature avoid discomfort.
The author spent some time expounding the metaphor linux:windows::motorcycle:automobile. I prefer linux:windows::hot-rod:minivan, with apologies to anyone who hot-rods their minivan. Linux is for people who spend time under the hood, looking for a new tweak to coax a little more from their machine. Windows is for those who just want to get someplace. Linux is for people who are their own mechanic. Windows is for those who head for the dealer when the "check engine" light comes on. The average person has no idea why anybody would ever want to work so hard setting up their computer just so. The average linux user has no idea how anybody could drive a Windows machine without dying of boredom or embarrassment, if not both.
Different folks, different strokes. If Linux was just another brand of Windows, would you still want it? Choice, my friends, choice: that's what free software is all about!
In the US auto industry, car manufacturers are required to produce replacement parts for a model for seven years. Thank goodness the aftermarket parts industry comes out with lower priced, longer lasting replacements. $300 for a factory-replacement fuel pump that's identical to the one that failed in less than four years? No thank you, I'll fix my 20 year old rust bucket with parts from the NAPA store.
The problem with software and their attached EULAs is that there isn't a NAPA store. When it breaks you can't fix it, in fact the EULA claims it's illegal to diagnose the problem (reverse engineering). Only the supplier's mechanics are allowed to open the hood and take a look. All you can do is buy a new one when the dealer says they quit fixing your model.
The sweet spot for best efficiency depends on the design. The bulk of the loss is of three types: static power (the same power at any load, making efficiency worse at low power), constant voltage losses (loss increases linearly for efficiency independent of load), and resistive losses (loss increase as the square of load, reducing efficiency at high power). Best efficiency can occur at, below, or even above the maximum rated power output: it depends on how much loss of each type occurs in that particular design.
These are the same factors that lead to the well-known conclusion that a transformer's efficiency is maximum at the point where core loss (static loss) is equal to the winding loss (resistive loss). It's an optimization exercise from your calculus class, find where the derivative is zero. The theory is fine, but the system operating point usually is dictated by other things (cost, size, and heat). If rated power corresponds to best efficiency, it's likely just a coincidence.
Rohn 25G topwers on hinged bases are not self-supporting at any height. They must always be guyed or bracketed to a building.
Also, since he mentioned paying Canadian dollars for it, ice loading will limit the self-supporting height. With 1/2" ice load the best you can do is 20 feet, in a low wind area and a small antenna. Even in a mild climate, the safe limit for self-supported 25G is 40 feet.
Do what the manufacturer says. They know where the failure point is. Overloaded towers will fail, it's a matter of when--not if.
The article hypes up to 30% power savings. Even if the DC system is perfect (100% efficiency), they are comparing it to AC systems no better than 70% efficient. If the DC system really has 90% efficiency end-to-end, the comparison is to an AC system less than 63% efficient.
That was pretty good for 1970, but not today! A 300W PC power supply at 63% efficiency has about 475W input, which means 175W loss. 175W would turn it into a crispy critter--more heat than a high end CPU, with less effective heat removal.
Assuming 70-80% efficiency for today's off-the-shelf power supplies and perhaps 15% potwntial power savings would be realistic, but that streches out the financial pay-back time. Not that it would make a big difference: most equipment buyers don't care how much power is wasted, as long as the servers keep running. Power conservation matters only when the circuit breakers trip or the system shuts down from overheating. The utility bills get paid from another department's budget.
The true benefit is moving some heat generation out to the DC power source, away from the servers. That improves the server room cooling budget, so you can squeeze in a few more racks before paying for a major facility upgrade. But this is a one-time shot. The next stage must be reductions of load power consumption. There isn't enough loss in a 90% efficient power supply to cover the next round of upgrades.
Apple embedded a marker in the code to help prove literal copying of Apple code. While hidden signatures in code may have roots in the authors' pride, they also serve a useful purpose in infringement cases. This snippet of text is completely irrelevant to the operation of OS/X, and the courts would presume that nobody except Apple would put it there. Apple isn't the first company to use them, nor will they be the last.
A KWHr from the computer is equivalent to a KWHr of resistance heating, but not necessarily to a KWHr in a heat-pump. Heat pumps can be more effective than resistance heat, especially if the source heat exchanger is close to room temperature. Ground-source heat pumps use the relatively stable sub-soil temperature to achieve impressive results, many times better than resistive heat.
Of course, this is an option for the landlord, not the renter. So turn your thermostat down as far as you can stand, get an automatic set-back thermostat if you don't have one, and learn to love warm clothes and thick blankets.
But if companies violate the license, the consequences can be more severe than they think. If companies are violating the GPL, they don't have the right to use that software. And if they don't have the right to use the software, they're violating federal law if they claim that they do.
The GPL concerns copyright: the ability to legally distribute the work to others. Companies may use software under the GPL license in any way they see fit; attempts to limit usage are the stuff of shrink-wrap EULA restrictions. The GPL specifically disclaims usage restrictions (see section 0 of the license: http://www.gnu.org/licenses/gpl.txt.
The GPL violations mentioned in the article apply to distribution of embedded systems and derviative works containing GPL code, not to companies using GPL software internally. Keep that distinction in mind to understand the context of the article.
Under previous US law, patents were valid 17 years from date of issue. Under present law, the term is 20 years from date of filing. During the transition between old and new law the term is either 20 years after filing or 17 years after the issue date, which ever is later. After the patent expires the inventioon is in the public domain, but one must be careful of similar inventions with later patent dates.
Try buying a computer from almost any large chain store. They tell you explicity that if you have a warranty claim, call the manufacturer--don't bring it back to the store. They don't accept returns if the package has been opened. Stores have been burned too often by thieves buying a system, taking out new good components, and returning it full of old inferior or broken ones. It's a pain for the honest customers, but you can't blame them for protecting themselves.
Counterfeiting the complete retail product is only part of the problem. Every level of sub-contractors can be taken in by counterfeits, too. Salvaged and rejected ICs are re-marked to look like new grade-A parts, low grade PSU capacitors are reborn with high end wrappers, not to mention the counterfeit safety approval markings. Longer supply chains mean more opportunities for inferior material to find its way into a product. Less control of the supply chain is the dark side to "outsourcing".
Quite right! For example, we bought our daughter a low-end Compaq system for homework, etc. It came with 256MB memory, XP/home, Symantec security 90 day trial, Office 90 day trial, and various utilities H-Paq includes and loads at start-up. Booting that system was very slow, and Office was sluggish to the point of being unusable. Adding another 512MB SIMM made it a much better running system.
Mere ability to boot and open an application is meaningless. To be useful, the system must be responsive. That means adequate physical memory; swapping to virtual memory "works", but constant swapping kills responsiveness: the application opens a document file, the application swaps out so the virus scanner has room to work, the application swaps back in, finds the document contains an embedded object, opens that file, repeat ad infinitum.
For the average user, response to keyboard and mouse inputs defines the system performance. Having enough memory determines if they experience good performance, no matter which OS you're using. The key point, which the article side-steps, is how much memory is needed for good performance, not the minimum to run it at all.
Some of that spectrum will go to first responders -- police, fire and public safety officials -- so they can better communicate with one another. Breakdowns in emergency communication slowed the response to the September 11 terrorist attacks and Hurricane Katrina. New spectrum should help.
That means duplication or replacement of most existing public safety radio equipment. Since the departments barely have the budgets to maintain the existing systems, where will the money come from to buy, install, and maintain all the new ones? Interoperability was a goal even before 9/11. It still hasn't happened because nobody is willing to pay for it. New spectrum won't help without the equipment to use it.
The rest of the spectrum will be auctioned off to the highest bidders -- probably tech companies. The sale of this valuable, scarce real estate is expected to bring in about $10 billion, maybe more. That will help reduce the federal budget deficit...Scheduled for 2008, the auction will be the biggest spectrum sale since a 1994-95 spectrum auction. That sale helped boost the mobile phone industry, boosting the number of cell phone subscribers in the U.S. from 24 million to 200 million. It also helped drive down the cost of wireless minutes from an average of 47 cents a minute to 9 cents a minute, according to analysis from financial services firm Stifel Nicolaus.
First, recall the huge expenditures needed for new public safety radio equipment. That alone is likely to consume all the auction revenue.
Second, recall the telcom bust that followed the '94-'95 land grab. The survivors remember the financial bloodbath that resulted from that bidding war, and are unlikely to spend so profligately again. The principle of supply and demand strongly suggests that declining air-time prices are symptoms of excess capacity. Why would the telcoms pay billions for more, when they need huge discounts to sell what they already have?
Like all operating systems of the time, the short names reflected the hardware limitations of the computers running it and the skill expected of the person using it. Both core memory and disk space were very expensive, and thus very limited. File systems restricted the length of file (command) names to conserve space in the directories. Terminal speeds were slow by today's stanmdards: 9600 baud (960 char/sec) for local users, 300-2400 baud (30-240 char/sec) over a modem. Long commands meant poor throughput. The users were expected to learn that OS and its commands, and use them regularly: it was targeted for skilled users, not the average Joe.
Unix offered an extraordinary luxury for its day by not limiting file names to six to eight characters (plus a short "file type" suffix that was the user generally doesn't have to type). Brevity was not only a matter of efficiency, but essential to avoid running out of room. The brevity ethic carried over into unix, probably by force of habit, even though the file name length limitation no longer demanded it. Microsoft followed the common practice of the day in MS/DOS and Windows through '98; only with NTFS did they adopt a file system with native support for filenames exceeding CP/M's "8.3" format.
And yes, many still do type with some kind of hunt-and-peck method--with that many years of practice, we old dogs are fast enough that retraining us is pointless.
Interference does not occur every time, but it has occurred and continues to occur in many documented cases. Deterioration of the aircraft systems, inconsistent (low cost) quality of personal electronic devices, the sheer number and variety of devices involved, and the relative locations of the "culprit" and "victim" equipment makes aircraft EMC issues difficult to solve. Increasing reliance on electronics for safety-critical flight systems (fly-by-wire, precision approach and auto-landing, "glass cockpit" instrument panels, etc.) has made interference more critical than ever before; in the the newest civil aircraft RF interference is a much more serious issue than it was for eletro-mechanical-hydraulic systems used in older planes. The airlines' built-in phones and entertainment devices have crucial improvements over consumer products: they are designed, built, and tested to stringent interference control standards, and the flight crew has custody of the "OFF" switches.
For example, I recently bought two 512MB RAM sticks, but only one qualified for the rebate. Besides explaining in part why 40% of the rebates go unclaimed, this establishes a market differentiation: home users get a rebate (eventually), while businesses are effectively shut out of the sale price. Like the $5 turkey sale before Thanksgiving, it benefits the home cook without attracting buyers from the local restaurant.
Another advantage to tankless heaters is that they can be located closer to the point-of-use than a large central heater. With a central heater, the water must flow for some time to flush the cold water out of the pipes between the tank and the shower or sink, and warm up the pipes so the water isn't cooled on the way. Several gallons are wasted before hot water reaches a tap at the far end of the plumbing.
The big deal is this: to build a computer you need parts, not press releases. Xeon is what Intel is selling, and Opteron is what AMD is selling, today. We'll see how next-generation parts (from both sides) measure up when they are in production. Many changes can happen between the marketing plan and production silicon.
Only 65km? On a clear day I can see Pike's Peak, 150km south and only 2.8km higher up.
The line-of sight range to an airship at 25km altitude is about 565km. LOS range to a location at the earth's surface is determined from Pythagoras' theroem. The earth's circumference is about 40,000km, so the earth's radius Re is about 6370km. At height h, LOS = sqrt((Re+h)^2-Re^2) = sqrt(2*Re*h+h^2). The textbook approximation LOS=sqrt(2*Re*h) is valid for Re>>h. Conclusion: they're probably using a narrow antenna beam to reduce the coverage area
You'd be much better off using a inverter, rather than an old UPS. Most low cost UPS units are designed for limited run time, not continuous duty operation. As designed and delivered, its output inverter will not over-heat before the internal battery runs down. This way they can use smaller heat sinks, omit cooling fans, etc. and sell a 500VA UPS for under $100. The inverter is for continuous duty use, so it needs better cooling and more rugged components--but it doesn't include the battery, charger, and power transfer relay that a UPS has.
Besides that, running a car engine to charge the battery wastes fuel. You can run far longer with a portable generator than by burning the same amount of fuel in your car engine.
Incoming bags have never used this system. From the article:
"United, Denver's busiest airline, has been using a stripped-down, simplified version of the network for its outgoing flights since the airport opened in 1995 - though 'enduring' is probably the better word, since regular breakdowns have continued despite years of tinkering.
Automation never worked for incoming flights, whose baggage has been moved by handlers from the beginning. And no other airline ever tried to use the error-prone system at all."
If you walk the bridge from Concourse A, you may waiting a lot longer than ten minutes! That's why if it's more than I can carry on, it ain't going. That's also why UAL/Ted wants to get gates on Concourse A. Frontier, with all their gates at close-in (by DIA standards) Concourse A, has a competitive edge. Many customers like having walk-in access, with a separate security screening station restricted to Concourse A passengers. Concourse A is a lot like going though the old Stapleton Field terminal, except it's a half hour closer to Kansas City!
Slashdot Mirror
Cordless power tools are exempt from the cadmium restriction, so they can continue using NiCd cells. But if I read the article correctly, they must be removable and collected for recycling when you buy the replacement battery. More information and regulation history is available at the EU web site. On the whole, this is very much in line with the RoHS and WEEE directives. It's surprising they delayed implementation for as long they have.
This isn't about "right" or "wrong", it's about human nature. What you learn first is what you will be comfortable with. If your first system uses linux, oo.o, and gimp you'll be comfortable using them. If your first system has Vista, MS Office and Photoshop then those will be your reference point. Any change will be uncomfortable, and it is human nature avoid discomfort.
Different folks, different strokes. If Linux was just another brand of Windows, would you still want it? Choice, my friends, choice: that's what free software is all about!
The problem with software and their attached EULAs is that there isn't a NAPA store. When it breaks you can't fix it, in fact the EULA claims it's illegal to diagnose the problem (reverse engineering). Only the supplier's mechanics are allowed to open the hood and take a look. All you can do is buy a new one when the dealer says they quit fixing your model.
These are the same factors that lead to the well-known conclusion that a transformer's efficiency is maximum at the point where core loss (static loss) is equal to the winding loss (resistive loss). It's an optimization exercise from your calculus class, find where the derivative is zero. The theory is fine, but the system operating point usually is dictated by other things (cost, size, and heat). If rated power corresponds to best efficiency, it's likely just a coincidence.
Also, since he mentioned paying Canadian dollars for it, ice loading will limit the self-supporting height. With 1/2" ice load the best you can do is 20 feet, in a low wind area and a small antenna. Even in a mild climate, the safe limit for self-supported 25G is 40 feet.
Do what the manufacturer says. They know where the failure point is. Overloaded towers will fail, it's a matter of when--not if.
That was pretty good for 1970, but not today! A 300W PC power supply at 63% efficiency has about 475W input, which means 175W loss. 175W would turn it into a crispy critter--more heat than a high end CPU, with less effective heat removal.
Assuming 70-80% efficiency for today's off-the-shelf power supplies and perhaps 15% potwntial power savings would be realistic, but that streches out the financial pay-back time. Not that it would make a big difference: most equipment buyers don't care how much power is wasted, as long as the servers keep running. Power conservation matters only when the circuit breakers trip or the system shuts down from overheating. The utility bills get paid from another department's budget.
The true benefit is moving some heat generation out to the DC power source, away from the servers. That improves the server room cooling budget, so you can squeeze in a few more racks before paying for a major facility upgrade. But this is a one-time shot. The next stage must be reductions of load power consumption. There isn't enough loss in a 90% efficient power supply to cover the next round of upgrades.
Apple embedded a marker in the code to help prove literal copying of Apple code. While hidden signatures in code may have roots in the authors' pride, they also serve a useful purpose in infringement cases. This snippet of text is completely irrelevant to the operation of OS/X, and the courts would presume that nobody except Apple would put it there. Apple isn't the first company to use them, nor will they be the last.
Of course, this is an option for the landlord, not the renter. So turn your thermostat down as far as you can stand, get an automatic set-back thermostat if you don't have one, and learn to love warm clothes and thick blankets.
The GPL concerns copyright: the ability to legally distribute the work to others. Companies may use software under the GPL license in any way they see fit; attempts to limit usage are the stuff of shrink-wrap EULA restrictions. The GPL specifically disclaims usage restrictions (see section 0 of the license: http://www.gnu.org/licenses/gpl.txt.
The GPL violations mentioned in the article apply to distribution of embedded systems and derviative works containing GPL code, not to companies using GPL software internally. Keep that distinction in mind to understand the context of the article.
Under previous US law, patents were valid 17 years from date of issue. Under present law, the term is 20 years from date of filing. During the transition between old and new law the term is either 20 years after filing or 17 years after the issue date, which ever is later. After the patent expires the inventioon is in the public domain, but one must be careful of similar inventions with later patent dates.
Counterfeiting the complete retail product is only part of the problem. Every level of sub-contractors can be taken in by counterfeits, too. Salvaged and rejected ICs are re-marked to look like new grade-A parts, low grade PSU capacitors are reborn with high end wrappers, not to mention the counterfeit safety approval markings. Longer supply chains mean more opportunities for inferior material to find its way into a product. Less control of the supply chain is the dark side to "outsourcing".
Mere ability to boot and open an application is meaningless. To be useful, the system must be responsive. That means adequate physical memory; swapping to virtual memory "works", but constant swapping kills responsiveness: the application opens a document file, the application swaps out so the virus scanner has room to work, the application swaps back in, finds the document contains an embedded object, opens that file, repeat ad infinitum.
For the average user, response to keyboard and mouse inputs defines the system performance. Having enough memory determines if they experience good performance, no matter which OS you're using. The key point, which the article side-steps, is how much memory is needed for good performance, not the minimum to run it at all.
That means duplication or replacement of most existing public safety radio equipment. Since the departments barely have the budgets to maintain the existing systems, where will the money come from to buy, install, and maintain all the new ones? Interoperability was a goal even before 9/11. It still hasn't happened because nobody is willing to pay for it. New spectrum won't help without the equipment to use it.
The rest of the spectrum will be auctioned off to the highest bidders -- probably tech companies. The sale of this valuable, scarce real estate is expected to bring in about $10 billion, maybe more. That will help reduce the federal budget deficit...Scheduled for 2008, the auction will be the biggest spectrum sale since a 1994-95 spectrum auction. That sale helped boost the mobile phone industry, boosting the number of cell phone subscribers in the U.S. from 24 million to 200 million. It also helped drive down the cost of wireless minutes from an average of 47 cents a minute to 9 cents a minute, according to analysis from financial services firm Stifel Nicolaus.
First, recall the huge expenditures needed for new public safety radio equipment. That alone is likely to consume all the auction revenue.
Second, recall the telcom bust that followed the '94-'95 land grab. The survivors remember the financial bloodbath that resulted from that bidding war, and are unlikely to spend so profligately again. The principle of supply and demand strongly suggests that declining air-time prices are symptoms of excess capacity. Why would the telcoms pay billions for more, when they need huge discounts to sell what they already have?
Like all operating systems of the time, the short names reflected the hardware limitations of the computers running it and the skill expected of the person using it. Both core memory and disk space were very expensive, and thus very limited. File systems restricted the length of file (command) names to conserve space in the directories. Terminal speeds were slow by today's stanmdards: 9600 baud (960 char/sec) for local users, 300-2400 baud (30-240 char/sec) over a modem. Long commands meant poor throughput. The users were expected to learn that OS and its commands, and use them regularly: it was targeted for skilled users, not the average Joe.
Unix offered an extraordinary luxury for its day by not limiting file names to six to eight characters (plus a short "file type" suffix that was the user generally doesn't have to type). Brevity was not only a matter of efficiency, but essential to avoid running out of room. The brevity ethic carried over into unix, probably by force of habit, even though the file name length limitation no longer demanded it. Microsoft followed the common practice of the day in MS/DOS and Windows through '98; only with NTFS did they adopt a file system with native support for filenames exceeding CP/M's "8.3" format.
And yes, many still do type with some kind of hunt-and-peck method--with that many years of practice, we old dogs are fast enough that retraining us is pointless.
This link (and its many references) are convincing: http://www.rvs.uni-bielefeld.de/publications/Inci
Myth: Upheld!
For example, I recently bought two 512MB RAM sticks, but only one qualified for the rebate. Besides explaining in part why 40% of the rebates go unclaimed, this establishes a market differentiation: home users get a rebate (eventually), while businesses are effectively shut out of the sale price. Like the $5 turkey sale before Thanksgiving, it benefits the home cook without attracting buyers from the local restaurant.
I checked all my LPs, and not a single one has a root kit.
Another advantage to tankless heaters is that they can be located closer to the point-of-use than a large central heater. With a central heater, the water must flow for some time to flush the cold water out of the pipes between the tank and the shower or sink, and warm up the pipes so the water isn't cooled on the way. Several gallons are wasted before hot water reaches a tap at the far end of the plumbing.
The big deal is this: to build a computer you need parts, not press releases. Xeon is what Intel is selling, and Opteron is what AMD is selling, today. We'll see how next-generation parts (from both sides) measure up when they are in production. Many changes can happen between the marketing plan and production silicon.
The line-of sight range to an airship at 25km altitude is about 565km. LOS range to a location at the earth's surface is determined from Pythagoras' theroem. The earth's circumference is about 40,000km, so the earth's radius Re is about 6370km. At height h, LOS = sqrt((Re+h)^2-Re^2) = sqrt(2*Re*h+h^2). The textbook approximation LOS=sqrt(2*Re*h) is valid for Re>>h. Conclusion: they're probably using a narrow antenna beam to reduce the coverage area
Besides that, running a car engine to charge the battery wastes fuel. You can run far longer with a portable generator than by burning the same amount of fuel in your car engine.
If you walk the bridge from Concourse A, you may waiting a lot longer than ten minutes! That's why if it's more than I can carry on, it ain't going. That's also why UAL/Ted wants to get gates on Concourse A. Frontier, with all their gates at close-in (by DIA standards) Concourse A, has a competitive edge. Many customers like having walk-in access, with a separate security screening station restricted to Concourse A passengers. Concourse A is a lot like going though the old Stapleton Field terminal, except it's a half hour closer to Kansas City!