I don't buy some kind of personalized, reduced-rate sandwich card at Subway. They give me these little stamps... that I can redeem later when I have several of them.
Not anymore. Subway recently discontinued the Sub Club program because there were too many stolen and counterfeit stamps. Near the end of last year, my local Subway was robbed. All they took was the spool of stamps.
If the feds have enough evidence to suspect someone on a flight, then maybe they should have the authority to shut down the WiFi server for that flight. If the suspicions were wrong, then--at worst--a couple hundred people have been temporarily inconvenienced.
Technologically, it might even be easier to do. It's a faster response to a threat, so it seems more likely to actually STOP an attack than collecting a data stream of potential evidence that will be useless by the time it has been decrypted. As a bonus, nobody's privacy gets invaded simply because we didn't think we had time to get a warrant.
... since the camera is presumably looking at the heat coming from the veins...
That may or may not be true. They may be using a infrared light which may be reflected/scattered/absorbed differently by the veins and blood than by the surrounding tissue. I'd be surprised if you could get a good high-contract image of the veins just by sensing the radiant infrared.
Ever been to the hospital and had one of the clips on your finger that measure your pulse and the oxygen saturation of your blood? The clip shines different colors of light through your finger, usually with red and infrared LEDs. Hemaglobin with oxygen scatters one of the wavelengths more than the other. By comparing the ratio of the two colors that's transmitted, you can estimate what percentage of your hemaglobin (sp?) is currently carrying oxygen. Also, as blood surges through your veins, the amount of blood passing through the light increases and descreases. Using either or both of the waveforms, you can determine the pulse rate. That's the basic theory of pulse oximetry. Doing it well in practice requires significant signal processing.
[O]f America's 285 million TVs only 12 percent (33.6 million) are used for watching OTA broadcasts.... [A]pproximately 3 million (around 10 percent) aren't used for viewing broadcast television at all.
If 12% is 33.6 million, then how can 10% be 3 million? Perhaps the author meant 10% of the 12%, but--as written--that's a contradiction. The 12% is reportedly the group of TV sets used for OTA broadcast, and the 10% is a group that does NOT use any broadcast (OTA, cable, or satellite).
I'm surprised people are already trying to do things with the player. I can't even get the thing to run. It just hangs Firefox. I'm on a stock machine that meets all the system requirements.
This is one time I regret having tabbed browsing, as now I lose my context in several places rather than just the instance that the Google player hosed.
[T]he article wasn't about open source. It was about open standards.
The article says the minister's plan involved both open standards and open source.
The Minister, as part of the plan, has charged all government institutions, both at the national and local level, to by the end of 2005 have worked out a recommendation for the use of open source code in the public sector. Further by the end of 2006 every body of the public sector in Norway must have in place a plan for the use of open source code and open standards.
(...Ford isn't liable for criminals using their vehicles as bank robbery getaway cars). The difference, with Grokster, is the actual promotion of the service as a way to violate copyrights.
Last night I saw a Chrysler Crossfire commercial that touted the car as performance tested at 150 miles per hour, without any disclaimer about obeying traffic laws and driving safely. Are they promoting a car for illegal uses?
It was unanimous, but not in the "any P2P software is illegal" sense, but in the "if you push your p2p software as a means for infringing copyright, you're just as guilty"
Exactly correct. It's a pretty limited decision, specific to Grokster's business model.
Sounds dangerous to me. It's not a well-defined line. What's the standard for determining promotion?
The RIAA or MPAA could use accuse someone, say Bram Cohen of BitTorrent fame, that his product is intended to promote illegal distribution. If it's not true, the burden effectively shifts to the defendent to prove that the device does not promote infringement. Other sites certainly tout BitTorrent as a means of infringing copyrights, even if Bram doesn't.
I would, but I'm not sure who originally wrote it. I first heard it in 1985, and I doubt it was original then. Certainly long before Simcity 2000 Maxis.
"We don't serve strings in here," the bartender says. "Get out."
The string leaves, waits a few minutes, then slips back inside and takes a seat in a booth out of sight from the bartender. He twists his body around in a strange contortion, then begins to unravel a little.
At last the waitress comes by and takes his order.
"Wait a minute," the waitress says. "Are you a string?"
What exactly is wrong with making a checkmark in a circle beside the name of a candidate one wishes to vote for, and then counting such votes manually?
I'm not a fan of the current state of electronic voting machines, but the proponents make several arguments for a higher tech approach. Some of them are good reasons.
Accessibility: paper and pencil doesn't work well for the blind or people with certain physical handicaps. A highly touted feature of the electronic machines is that they can use other I/O methods to accomodate these people.
Multilingual: in places like California, ballots come in many languages. Having a machine that can present the same ballot in a selection of languages simplifies logistical problems of making sure you have the right number of each language at each polling place.
Complex Ballots: in a general election, there are usually several offices and issues to vote on. Futhermore, which issues you get to vote on depend upon which district/county/municipality you live in. If I want to vote near where I work (a different county than where I live), an electronic ballot machine can present me with the appropriate set of choices. Counting ballots with a single question by hand is easy, but when they get as complicated as they are here in California, the benefits of automation start to look appealing.
Speed: the media wants results faster. (I'm not sure why. It took weeks to sort out the 2000 presidential election, and the news outlets got tons of extra viewership because of it.) Electronic tallies are perceived as faster.
The OS will only run IE and Windows Media; everything else will be on an application server. I do not think this solves the actual problem. We have terminals.
Terminals connected to a central server can be a very efficient solution for many scenarios. No, you're not going to run high-end developer tools that way, but I could imagine a bunch of thin clients doing wordprocessing and the like without much trouble. Consider how much time your CPU is idle, even during work hours.
I'm not arguing for or against copyright infringement. I'm just calling BS on the tangential linguistic debate.
"Piracy" is an appropriate word in this context. Usage of "piracy" to mean copyright infringement is not a result of recent spin from the likes of BSA, RIAA, and MPAA. It has a history that predates these propaganda machines.
The GNU rant, which the parent of my original post linked, is counter-productive by trying to assert associations with kidnapping and murder. Hyperbole like that is more likely to set off BS detectors than to advance the cause. If you want to point out that copyright infringement is not the a crime like theft, fine. But this piracy argument seems more like an attack on language than a contribution for healthy debate on copyright issues.
Piracy is a confusing word you should avoid. The title of the story should read, "Software Sharing To Increase as the Internet Grows." Doesn't sound as bad, does it?
The GNU paragraphs you reference on piracy are a little over the top. The first definition of piracy in most dictionaries I've seen is simply "an act of robbery, especially on the high seas". That certainly doesn't imply kidnapping and murder as the GNU page likes to rant.
The second definition (even in decades-old dictionaries) is "copyright infringement". Thus "piracy" seems a perfectly serviceable word for this discussion. Repeated typing or saying "unauthorized copying", "copyright infringement", and similar terms becomes cumbersome.
Is it just me, or do the numbers not add up? On the one hand, this:
Currently, about one-third of software used is illegally made copies.... Worldwide revenue loss due to software piracy was estimated at $33 billion for 2004.
seems to suggest that the worldwide market is about $100 billion... per year.
You're assuming equal distributions of revenue and copyright infringement. The article pointed out that piracy rates vary dramatically in different parts of the world. It's also likely that software prices vary by market. Thus a pirated copy in one market may be have been valued differently than one in another. I don't think we have the data here to conclude that the market is $100 billion.
In fact, a little web searching yields a variety of values. One estimate I saw was $230 billion back in 2003.
The whole website's entirely Firefox compat[i]ble....
I'm not sure what version of Firefox you're running, but 1.0.4 had some rendering problems for me. Column breaks split through the lines of text, leaving the tops of the letters at the bottom of one column and the bottoms at the top of the next.
I love how you can click on the right half of the article to move to the next page, or left side to move back....
I was completely distracted from the article, trying to figure out what I was doing that caused the Next and Prev buttons to light up. When they aren't lit up, they look disabled. As I was moving the mouse (mostly) vertically, the Next button would light up. Very disconcerting and confusing and not particularly helpful. I'd rather have a button with some affordance for clicking. Or just lay the text out on one page and let me scroll! Scrolling can be done easily from the keyboard.
In my sci-fi novel, Kill Switch, long space flight is achieved by killing the astronauts under controlled circumstances, preserving the bodies, and having machines revive them at the destination. It works most of the time; occasionally, though, somebody ends up with amnesia.
Read carefully. This isn't just about targeting to certain demographics. They are also providing that specific but non-identifying demographic info back to the advertiser.
Microsoft would then provide the company with detailed information about the demographics of the people who clicked on its ads.
If I click one of these ads, Microsoft will report to the advertiser that the user clicking this ad is a 37-year old man in ZIP code 94542, and average household income in that ZIP code is $105,393. This report is the scary part.
I wonder how specific they get with the birthday. After all, 87% of Americans can be uniquely identified from gender, birthday, and ZIP code. So is it really non-identifying data?
I've been trying to install the.NET framework on my PCs for months so I can use this cool Logitech io gizmo whose driver and supporting software require the framework. Unfortunately, the official.NET installation package silently fails on my stock Windows boxen, so I've never been able to use it.
If this malware could successfully install it, I'd be golden!
So, I'm not surprised at all that that data [language preference] isn't there. If you want to be surprised by this, you should probably be surprised that the bank didn't choose to store your language preference in their database and then look it up when you swipe your card.
I'm sure things have changed a lot in how the ATM networks work, and such a scheme may be feasible now, but this wouldn't have fit the model they had when first introduced. Throughout the 1970s, my mother, father, and step-father all wrote code for banking terminal systems and some of the first ATMs. From them I learned:
There was one roundtrip to the bank's central computers after you had entered everything for the transaction. I assume this was for scalability. The ATM would collect your card number, PIN, and transaction request and send it as a single request the central computer. That's why they wouldn't tell you about a mistyped PIN until you've entered everything else for your transaction. Transactions were stored in a secondary database which were posted to your real account record overnight.
In the good old days, the bank didn't assign a PIN for you, store it in a database (which could be snooped by employees), printed it on paper (which could be discovered by anyone), and send it to you in the mail (which could be stolen). Instead, to activate your account, you went to your local branch. A teller would come out to the ATM with you, put his/her card into the machine, enter his/her PIN, then insert your card, and finally turn his/her back while you entered a PIN of your choice. PINs were hashed in the ATM and the bank only ever had the hash, not the original value.
How does this affect TiVo? My series one box uploads activity logs whenever it calls in for a listings update and subscription check. TiVo promises to use the data anonymously and in aggregate. Does this now become illegal?
I'd hope not because for most problems efficient *algorithms* are more important than efficient implementation.
For many problems, yes, but for most, perhaps not. Every challenging performance problem I've tackled in the past five years was due to thrashing in virtual memory, either because the execution path skips throughout process space or because there's poor locality in the data set.
It has literally been years since I've worked on a performance problem that had to do with a loop-based algorithm. With event-driven GUIs, layers and layers of abstraction, marshalling, defensive security-enhancing code, and inheritance trees that are dozens of layers deep mean you can have tens of thousands instructions each executed a small number of times. There simply isn't a way to improve the performance by changing an algorithm. To solve these problems, you have to make sure all those scattered instructions are in cache (or at least RAM) and that the branches are predictable. There's no inner loop left that's accounting for a large portion of the time.
Load times in particular are a big problem. With dozens and dozens of libraries each needing to be loaded, patched up, and paged out because they couldn't load at their preferred base addresses. Faith that they will only be paged in as needed is unfounded unless you're very careful to choose your base addresses well. (Or if your processor is adept with relative branches and your compiler and linker are smart enough to exploit that.)
Slashdot Mirror
Not anymore. Subway recently discontinued the Sub Club program because there were too many stolen and counterfeit stamps. Near the end of last year, my local Subway was robbed. All they took was the spool of stamps.
If the feds have enough evidence to suspect someone on a flight, then maybe they should have the authority to shut down the WiFi server for that flight. If the suspicions were wrong, then--at worst--a couple hundred people have been temporarily inconvenienced.
Technologically, it might even be easier to do. It's a faster response to a threat, so it seems more likely to actually STOP an attack than collecting a data stream of potential evidence that will be useless by the time it has been decrypted. As a bonus, nobody's privacy gets invaded simply because we didn't think we had time to get a warrant.
That may or may not be true. They may be using a infrared light which may be reflected/scattered/absorbed differently by the veins and blood than by the surrounding tissue. I'd be surprised if you could get a good high-contract image of the veins just by sensing the radiant infrared.
Ever been to the hospital and had one of the clips on your finger that measure your pulse and the oxygen saturation of your blood? The clip shines different colors of light through your finger, usually with red and infrared LEDs. Hemaglobin with oxygen scatters one of the wavelengths more than the other. By comparing the ratio of the two colors that's transmitted, you can estimate what percentage of your hemaglobin (sp?) is currently carrying oxygen. Also, as blood surges through your veins, the amount of blood passing through the light increases and descreases. Using either or both of the waveforms, you can determine the pulse rate. That's the basic theory of pulse oximetry. Doing it well in practice requires significant signal processing.
If 12% is 33.6 million, then how can 10% be 3 million? Perhaps the author meant 10% of the 12%, but--as written--that's a contradiction. The 12% is reportedly the group of TV sets used for OTA broadcast, and the 10% is a group that does NOT use any broadcast (OTA, cable, or satellite).
Crappy reporting.
I'm surprised people are already trying to do things with the player. I can't even get the thing to run. It just hangs Firefox. I'm on a stock machine that meets all the system requirements.
This is one time I regret having tabbed browsing, as now I lose my context in several places rather than just the instance that the Google player hosed.
The article says the minister's plan involved both open standards and open source.
Last night I saw a Chrysler Crossfire commercial that touted the car as performance tested at 150 miles per hour, without any disclaimer about obeying traffic laws and driving safely. Are they promoting a car for illegal uses?
Sounds dangerous to me. It's not a well-defined line. What's the standard for determining promotion?
The RIAA or MPAA could use accuse someone, say Bram Cohen of BitTorrent fame, that his product is intended to promote illegal distribution. If it's not true, the burden effectively shifts to the defendent to prove that the device does not promote infringement. Other sites certainly tout BitTorrent as a means of infringing copyrights, even if Bram doesn't.
I would, but I'm not sure who originally wrote it. I first heard it in 1985, and I doubt it was original then. Certainly long before Simcity 2000 Maxis.
A string walks into a bar and orders a drink.
"We don't serve strings in here," the bartender says. "Get out."
The string leaves, waits a few minutes, then slips back inside and takes a seat in a booth out of sight from the bartender. He twists his body around in a strange contortion, then begins to unravel a little.
At last the waitress comes by and takes his order.
"Wait a minute," the waitress says. "Are you a string?"
"No," the string says. "I'm a frayed knot."I'm not a fan of the current state of electronic voting machines, but the proponents make several arguments for a higher tech approach. Some of them are good reasons.
Yup!
Terminals connected to a central server can be a very efficient solution for many scenarios. No, you're not going to run high-end developer tools that way, but I could imagine a bunch of thin clients doing wordprocessing and the like without much trouble. Consider how much time your CPU is idle, even during work hours.
I'm not arguing for or against copyright infringement. I'm just calling BS on the tangential linguistic debate.
"Piracy" is an appropriate word in this context. Usage of "piracy" to mean copyright infringement is not a result of recent spin from the likes of BSA, RIAA, and MPAA. It has a history that predates these propaganda machines.
The GNU rant, which the parent of my original post linked, is counter-productive by trying to assert associations with kidnapping and murder. Hyperbole like that is more likely to set off BS detectors than to advance the cause. If you want to point out that copyright infringement is not the a crime like theft, fine. But this piracy argument seems more like an attack on language than a contribution for healthy debate on copyright issues.
The GNU paragraphs you reference on piracy are a little over the top. The first definition of piracy in most dictionaries I've seen is simply "an act of robbery, especially on the high seas". That certainly doesn't imply kidnapping and murder as the GNU page likes to rant.
The second definition (even in decades-old dictionaries) is "copyright infringement". Thus "piracy" seems a perfectly serviceable word for this discussion. Repeated typing or saying "unauthorized copying", "copyright infringement", and similar terms becomes cumbersome.
They have. They say 33% for 2004, but the president of the BSA testified before Congress that the rate was 42% in 2003. That's impressive.
Also, technically, the BSA doesn't "profit" (in a financial sense) from piracy. It's a non-profit organization.
You're assuming equal distributions of revenue and copyright infringement. The article pointed out that piracy rates vary dramatically in different parts of the world. It's also likely that software prices vary by market. Thus a pirated copy in one market may be have been valued differently than one in another. I don't think we have the data here to conclude that the market is $100 billion.
In fact, a little web searching yields a variety of values. One estimate I saw was $230 billion back in 2003.
I'm not sure what version of Firefox you're running, but 1.0.4 had some rendering problems for me. Column breaks split through the lines of text, leaving the tops of the letters at the bottom of one column and the bottoms at the top of the next.
I was completely distracted from the article, trying to figure out what I was doing that caused the Next and Prev buttons to light up. When they aren't lit up, they look disabled. As I was moving the mouse (mostly) vertically, the Next button would light up. Very disconcerting and confusing and not particularly helpful. I'd rather have a button with some affordance for clicking. Or just lay the text out on one page and let me scroll! Scrolling can be done easily from the keyboard.
In my sci-fi novel, Kill Switch, long space flight is achieved by killing the astronauts under controlled circumstances, preserving the bodies, and having machines revive them at the destination. It works most of the time; occasionally, though, somebody ends up with amnesia.
Keep reading. Also from the article:
In my book, that's where they cross the privacy line.
Read carefully. This isn't just about targeting to certain demographics. They are also providing that specific but non-identifying demographic info back to the advertiser.
If I click one of these ads, Microsoft will report to the advertiser that the user clicking this ad is a 37-year old man in ZIP code 94542, and average household income in that ZIP code is $105,393. This report is the scary part.
I wonder how specific they get with the birthday. After all, 87% of Americans can be uniquely identified from gender, birthday, and ZIP code. So is it really non-identifying data?
I've been trying to install the .NET framework on my PCs for months so I can use this cool Logitech io gizmo whose driver and supporting software require the framework. Unfortunately, the official .NET installation package silently fails on my stock Windows boxen, so I've never been able to use it.
If this malware could successfully install it, I'd be golden!
I'm sure things have changed a lot in how the ATM networks work, and such a scheme may be feasible now, but this wouldn't have fit the model they had when first introduced. Throughout the 1970s, my mother, father, and step-father all wrote code for banking terminal systems and some of the first ATMs. From them I learned:
There was one roundtrip to the bank's central computers after you had entered everything for the transaction. I assume this was for scalability. The ATM would collect your card number, PIN, and transaction request and send it as a single request the central computer. That's why they wouldn't tell you about a mistyped PIN until you've entered everything else for your transaction. Transactions were stored in a secondary database which were posted to your real account record overnight.
In the good old days, the bank didn't assign a PIN for you, store it in a database (which could be snooped by employees), printed it on paper (which could be discovered by anyone), and send it to you in the mail (which could be stolen). Instead, to activate your account, you went to your local branch. A teller would come out to the ATM with you, put his/her card into the machine, enter his/her PIN, then insert your card, and finally turn his/her back while you entered a PIN of your choice. PINs were hashed in the ATM and the bank only ever had the hash, not the original value.
How does this affect TiVo? My series one box uploads activity logs whenever it calls in for a listings update and subscription check. TiVo promises to use the data anonymously and in aggregate. Does this now become illegal?
For many problems, yes, but for most, perhaps not. Every challenging performance problem I've tackled in the past five years was due to thrashing in virtual memory, either because the execution path skips throughout process space or because there's poor locality in the data set.
It has literally been years since I've worked on a performance problem that had to do with a loop-based algorithm. With event-driven GUIs, layers and layers of abstraction, marshalling, defensive security-enhancing code, and inheritance trees that are dozens of layers deep mean you can have tens of thousands instructions each executed a small number of times. There simply isn't a way to improve the performance by changing an algorithm. To solve these problems, you have to make sure all those scattered instructions are in cache (or at least RAM) and that the branches are predictable. There's no inner loop left that's accounting for a large portion of the time.
Load times in particular are a big problem. With dozens and dozens of libraries each needing to be loaded, patched up, and paged out because they couldn't load at their preferred base addresses. Faith that they will only be paged in as needed is unfounded unless you're very careful to choose your base addresses well. (Or if your processor is adept with relative branches and your compiler and linker are smart enough to exploit that.)