Data mining is great for providing statistical answers to questions. But the statistics can only be used for "inductive reasoning". The conclusion is not a fact, it's only a probability. "Joe learned how to make bombs in the army", "Joe is in an anti-war group", "Mary does not know how to make bombs", does not yield "Joe is an anti-war terrorist." It might yield "Joe has more potential to be an anti-war terrorist than Mary does."
Now, if there are more facts and premises, such as "People who plant bombs are terrorists" and "Joe was convicted for planting a bomb in an Army Recruiting Center", then you can deduce that Joe is a terrorist. But you still cannot deduce that "the anti-war group is a terrorist organization" unless you include a the premise that "an organization that has a terrorist as a member is a terrorist organization." And that premise may or may not be valid.
"I don't believe the First Amendment is any guarantee to those who wish to disrupt the government," [Hutchins] said.
I'm boggled.
That one absolutely floored me, too. That such an anti-American person is in a position of power shows the system is truly broken. Even so, I would not label Hutchins with a "Terrorism-anti-American-statements" choice. A simple "Anti-American-statements" label would be much more accurate.
If they've done something illegal, then arrest and prosecute them. If they haven't, then they should be free to go about their lives.
You're absolutely right. Hutchins and Sheridan should be arrested and prosecuted for slander, making defamatory statements, and abridging the civil rights of the 53 people they falsely accused of a truly heinous crime. And they should certainly be given a fair trial, and if found innocent their records should be cleaned and they should be free to go about their lives.
But saying stuff like "the activists' names were entered into the state police database as terrorists partly because the software offered limited options for classifying entries" should be very compelling evidence of making the false accusation in the first place. Trying to fix it later should be strong evidence that the suspects knew what they had done was wrong, and therefore had malice aforethought. There was no good faith here. This was an outright criminal act designed to deny 53 people their rights as citizens.
I'd say those 53 defendants have a pretty solid case on their hands.
Rather than looking individually at each problem, consider it from a strategic point of view, or one in statistics: maybe 99% of the herd cooperates 100% of the time. The idea is to find a way to automate or at least control the outliers, and the rest of the work is done.
Or if the ranchers can control the animals' behavior 99% of the time through automation, maybe they're willing to accept the risk of not having full control. What are the chances that some cow will step on a rattlesnake and cause a stampede that can't be stopped by radio collars? If it happens only once every 10 years, it might be worth it.
Or what if the collars let fewer hands do the work?
I know that at the root of what you're saying is "animals have their own self-motivations". And sure, a good cowboy will understand what motivates a cow (a fence, a rattlesnake, a bull, a calf, etc.) and will use that knowledge to their advantage. They will know that cows can get spooked and how to calm them, how to herd them, how to corral them, etc. I'm not arguing that a trained hand knows these things well. What I'm saying is that there's a reason we have the phrase "herd mentality", and that is that cow behavior is essentially statistical in nature. Figure out how to usually drive the herd, how to control the beasts and you can use it "most of the time". Maybe you reduce costs. Maybe four cowboys are too expensive. Or maybe the tech isn't mature enough. Depends on some rancher's willingness to take a risk on these collars.
And how often do those ranch hands not show up because they're sick, or hungover, or whatever? How much do those hands cost the rancher because they're inexperienced and do the wrong thing? How often do hands get injured on the job and sue for damages? And you never have to pay unemployment insurance if you leave a radio collar on the shelf for a year.
On the other side, how much do you pay for maintenance contracts for radio collars, and how reliable are they? How expensive is it if your herd wanders off because of a technology failure? How expensive are the redundancies? How does a rancher know when his gadgets are failing?
The only real question is if the machinery can perform the task adequately. Notice I did not try to compare them to a human, just asking "can they actually do the task required"?
The answer is that an automaton is almost always cheaper in the long run than a human. The more you train a human, the better or faster they'll perform a task -- within certain limits -- but then they'll cost you more, and the expense is perpetual. Machinery, on the other hand, costs you dearly up front but is practically free over the long term. It's even more cost effective when you consider that it doesn't need training -- plug it in and it goes.
i agree that Motorola has a lot of problems (i owned a RAZR, and while it was kinda cool at first, the software was terribly crippled)
As you speculated later, the software was indeed crippled by your phone service provider. I purchased an unlocked Mobilinux-based Z6 ROKR directly from Motorola and it's not hobbled in any way. There are no restrictions on ring tones, images, transferring or importing any media. The Linux-based music app works quite well, and certainly doesn't suck like the "Digital Music Player" that comes standard on their KRZR and RAZR lines. The user interface is clean and sharp, much more professional than their previous offerings.
I also have no complaints about the audio quality, and their completely automated voice recognition is excellent, even in noisy environments. (It recognizes names that are entered in the contact list simply by their English pronunciations -- no training required.) And the EDGE speeds clock in about 115kbps -- nothing to complain about.
However, their Bluetooth software is still bug-prone, and the simple act of loading up and using their Java implementation causes my phone to have fits. This is my third Motorola phone in almost as many years, and I'm getting pretty tired of the lack of improvement in these areas.
Simply looking at their posting, it's likely part of an attempt to spam Google into boosting the page rank for some of those NYT opinion pieces, or maybe to try to relate some of those stories to each other. (I'm not visiting or posting those links out of principle, so it's hard to say for sure.) They could be attacking just the search engine or possibly Google News as well. We can only hope that the Google spiders walking Slashdot ignore all "Score -1" postings.
I'm not sure why they didn't put hrefs around their links, tho. Maybe they don't understand/., or maybe it's just a bot.
The "I'm not buying it" argument against evolution typically comes because humans do not innately have the understanding of the time scales involved, or of the nature of probability.
Humans don't typically pay attention long enough to visibly see evolution taking place. The evidence is there, but it may take effort to put it together, and they're unwilling to do that. (I keep thinking that a month-long process of breeding antibiotic-resistant e. coli and feeding it to them would go a long way toward helping them understand it, but that's just me.) And lots of people don't understand probability -- look at gamblers and their "systems", or even try to explain the Monty Hall problem to someone.
And of course lots of people don't understand simply because they can't or won't make the effort. I have less respect for those people.
If the tests don't provide coverage for ALL the business issues that the piece of software is supposed to solve, then you pass the tests, but will have no clue if you broke or not things apart.
That's not true at all. You have a lot of clues based on the tests that passed. You'll have confidence in the code that passed the tests. Overall, it's less to worry about. It's certainly of more benefit than no tests. And if you failed tests, you'll have 100% confidence that you broke something, and you can get it fixed.
Consider the end goal of unit tests would be 100% assurance that all code paths are covered, and that all behavior is tested. Not that it's realistic in many places, but that is the goal.
Now consider "Legacy Code" as defined by Michael Feathers -- code that does not have unit tests. You have no idea if a single line of it works, other than what your integration or end-user testing shows.
His book is a set of procedures you can try to insert tests into untested or hard-to-test code. His strategy is fairly simple: jam a test into the middle of the code. Chop something in half, place tests on either side of the wound, and see if it still works. Now you have a big ugly gash separating two halves of the code, but you have less code to execute when running the tests on each half. Now, using that line as a starting point, begin refactoring the code, and adding unit tests to prove your refactorings are safe. Sure, you'll still need end user testing and integration testing, but your developers can move forward faster, and you're more confident that your new changes will improve the code, rather than break stuff.
He's trying to help solve a hard problem: how do you refactor (which is inherently a test-driven-development strategy for improving your code) without tests that help ensure you didn't break anything? Once you've got tests, refactoring becomes easier. The more tests you have, the more assurance you have that your changes aren't breaking code. Ultimately, when you are 100% covered, you should be a lot closer to bug-free than you were when you started.
I'll tell you right now that on a large legacy project, most people couldn't tell you where to begin placing tests. Michael's book helps jump start the process. And while the end game might not be 100% test coverage of all the legacy code, it leaves you a lot better off than you were before you started writing the tests.
Define "useful". My phone has a 1600x1200 sensor, and while the lens is certainly no Canon L series, it's able to accurately capture the contents of a whiteboard after a meeting just fine. I have pictures of my friends displayed on the little screen, and I know who they are. And it does take good incriminating shots of drunk pals in the bar!
Sure, if I want good pictures I'm going to bring a real camera. But if I'm just out and about, I won't have the good camera with me -- the thing weighs twelve times the weight of the phone, after all. So I find the phone does just fine for my average point-and-shoot needs.
The OK camera you have with you is far more useful than the great camera you wish you had with you.
So get an Eye-Fi SD card for your own decent digital camera. 2GB storage, and it automagically scans for your network (or any network access point named the same thing, such as LINKSYS:-) and uploads your pictures as soon as it can to your machine or to flickr or wherever.
You don't even have to buy a crappy camera that sucks because it's got a crappy phone in it for no reason. You can buy a decent camera and still get the magic functionality out of it.
Although.. every phone should have gps, since they've all got it built-in anyway so the carriers don't have to put any effort into triangulating.
That's not completely correct.
Many phones that seem to have "GPS" actually have "Assisted-GPS (A-GPS)". A-GPS allows for a cheaper device where the phone has only the receiver but no 'smarts'. These phones send the raw satellite data to a service provider to perform the math required to locate the phone. The service provider is able to fill in a lot of missing data if poor reception at the phone is an issue (a frequent problem in urban canyons), and because it does know tower locations and satellite ephemeris, it can very quickly provide reasonable location data. The service then returns the coordinates to the phone.
While they're more accurate than simple triangulation from the towers, those obviously cost the carrier in terms of providing A-GPS servers.
I have a Motorola Z6, and it's pretty close to your description, (except trade "rugged" for "sleek".) The acoustics are great, and I have no reception problems -- it usually gets "more bars" than an iPhone on the same carrier (AT&T). The voice recognition dialing is automated based on the names you enter in your contact list, and has been remarkably precise. And the screen is crystal clear.
The software, not so much. The Bluetooth gets fairly easily confused, and the headphones don't keep a perfect connection when using the AD2P music player.
I shouldn't say it isn't rugged -- mine looks as good as the day I bought it, apart from a tiny 1/8" scratch on the trim ring -- but it certainly does not have the feel of a rubberized Army field model.
Thank you, this is the most insightful reply in this thread.
I *hate* it when people whine about their phone's features. "Oh, I don't need all that crap, I just need to make a phone call!" They're usually not telling the truth. They were in the cell phone store once upon a time buying a new phone, and they could have accepted the feature-poor 'freebie' phone. But the salesman said "same price, this one has a free camera." Then he said "This one has a better lens, and it's only $20 more." And any phone with a camera comes with a display large enough for a miniature web browser, and that's just software. So these people willingly bought the phones that they then complain about.
I also know that people complain a lot about the CDMA carriers locking their phones down. If you go the Verizon / Sprint route, you're going to have lockdown -- plain and simple. Find a GSM carrier instead, and buy an unlocked phone.
Of course, phones get my goat too. What really cheezes me off is when I buy a phone specifically for a set of features, and the features do not work as advertised. The Bluetooth hangs, or the browser screws up, or the music player takes half an hour to load, or it crashes, or whatever. Motorola is great for that.
I'm a geek. If I want a phone, it's because I want to phone someone, or I want them to phone me. I don't need all this other stuff.
If you're a geek you don't need all this other stuff because you already have it hanging from pouches on your belt or in your cargo pants, right? Palm Tungsten -- check. Pager -- check. Digital Elph camera -- check. Calculator (HP-41C, no doubt) -- check. iPod -- check. GPS -- check.
Maybe it's just me, but the all-in-one phone seems to be a lazier choice. And we geeks are a lazy lot, after all.
IBM invented FUD. I've seen them playing hardball, and they were once so good at it that they used to manipulate the outcome of the World Series of Manipulation. IBM has simply learned that FUD could get you so far and no further, and now there's more to be gained by being honest (or at least more honest than Microsoft.)
How fried? Do you mean disable (i.e. temporary blindness), cripple (permanent blindness), burn (seared flesh), or maim (severed limbs or death?) An ordinary mirror will work fine for the lower power attacks (as would a sheet of cardboard!) If you have a mirror made from the right kind of material (such as copper) it'll deflect any of these. But the kind powerful enough to take out a satellite, missile, or weapons platform? I have heard that no ordinary mirror would withstand them, but that copper sheeting could still protect the target.
The more important question is: are you going to be hiding behind a mirror when the weapon strikes? My understanding is the big laser weapons are pulse weapons, meaning all their energy is delivered virtually instantly. It's not like you can watch a laser beam come zapping in and then duck just before it hits you. That only works for Captain Kirk.
Perhaps it's simply marketing. Maybe it's test marketing. Maybe they're trying to gauge the public acceptance of chip delivery systems. Maybe they're trying to see if the commercial public will accept 128kbps MP3s over higher quality formats.
Or maybe because everyone on Slashdot has been screaming "you need to figure out a different sales model!" This is indeed different.
The point is they're trying something new. Anything new. It costs them almost nothing to try, they get to look cool doing it, they successfully generated "buzz" by trying (look, here we are discussing it.) I don't think its any deeper than that.
No kidding. If the average kid had to do homework as a part of grinding in WoW, considering how much time they spend in game we might end up with a nation of Doogie Howsers.
Of course, all the asian gold farmers would quickly have their PhDs, and we'd really just end up with a nation of kids all trained to send their money overseas.
Slashdot Mirror
Data mining is great for providing statistical answers to questions. But the statistics can only be used for "inductive reasoning". The conclusion is not a fact, it's only a probability. "Joe learned how to make bombs in the army", "Joe is in an anti-war group", "Mary does not know how to make bombs", does not yield "Joe is an anti-war terrorist." It might yield "Joe has more potential to be an anti-war terrorist than Mary does."
Now, if there are more facts and premises, such as "People who plant bombs are terrorists" and "Joe was convicted for planting a bomb in an Army Recruiting Center", then you can deduce that Joe is a terrorist. But you still cannot deduce that "the anti-war group is a terrorist organization" unless you include a the premise that "an organization that has a terrorist as a member is a terrorist organization." And that premise may or may not be valid.
"I don't believe the First Amendment is any guarantee to those who wish to disrupt the government," [Hutchins] said.
I'm boggled.
That one absolutely floored me, too. That such an anti-American person is in a position of power shows the system is truly broken. Even so, I would not label Hutchins with a "Terrorism-anti-American-statements" choice. A simple "Anti-American-statements" label would be much more accurate.
If they've done something illegal, then arrest and prosecute them. If they haven't, then they should be free to go about their lives.
You're absolutely right. Hutchins and Sheridan should be arrested and prosecuted for slander, making defamatory statements, and abridging the civil rights of the 53 people they falsely accused of a truly heinous crime. And they should certainly be given a fair trial, and if found innocent their records should be cleaned and they should be free to go about their lives.
But saying stuff like "the activists' names were entered into the state police database as terrorists partly because the software offered limited options for classifying entries" should be very compelling evidence of making the false accusation in the first place. Trying to fix it later should be strong evidence that the suspects knew what they had done was wrong, and therefore had malice aforethought. There was no good faith here. This was an outright criminal act designed to deny 53 people their rights as citizens.
I'd say those 53 defendants have a pretty solid case on their hands.
Ze Googles! Zey do nothing!
Rather than looking individually at each problem, consider it from a strategic point of view, or one in statistics: maybe 99% of the herd cooperates 100% of the time. The idea is to find a way to automate or at least control the outliers, and the rest of the work is done.
Or if the ranchers can control the animals' behavior 99% of the time through automation, maybe they're willing to accept the risk of not having full control. What are the chances that some cow will step on a rattlesnake and cause a stampede that can't be stopped by radio collars? If it happens only once every 10 years, it might be worth it.
Or what if the collars let fewer hands do the work?
I know that at the root of what you're saying is "animals have their own self-motivations". And sure, a good cowboy will understand what motivates a cow (a fence, a rattlesnake, a bull, a calf, etc.) and will use that knowledge to their advantage. They will know that cows can get spooked and how to calm them, how to herd them, how to corral them, etc. I'm not arguing that a trained hand knows these things well. What I'm saying is that there's a reason we have the phrase "herd mentality", and that is that cow behavior is essentially statistical in nature. Figure out how to usually drive the herd, how to control the beasts and you can use it "most of the time". Maybe you reduce costs. Maybe four cowboys are too expensive. Or maybe the tech isn't mature enough. Depends on some rancher's willingness to take a risk on these collars.
And how often do those ranch hands not show up because they're sick, or hungover, or whatever? How much do those hands cost the rancher because they're inexperienced and do the wrong thing? How often do hands get injured on the job and sue for damages? And you never have to pay unemployment insurance if you leave a radio collar on the shelf for a year.
On the other side, how much do you pay for maintenance contracts for radio collars, and how reliable are they? How expensive is it if your herd wanders off because of a technology failure? How expensive are the redundancies? How does a rancher know when his gadgets are failing?
The only real question is if the machinery can perform the task adequately. Notice I did not try to compare them to a human, just asking "can they actually do the task required"?
The answer is that an automaton is almost always cheaper in the long run than a human. The more you train a human, the better or faster they'll perform a task -- within certain limits -- but then they'll cost you more, and the expense is perpetual. Machinery, on the other hand, costs you dearly up front but is practically free over the long term. It's even more cost effective when you consider that it doesn't need training -- plug it in and it goes.
> radios that attach to an animal's head that allow a person at the other end...
hmm, where's this person supposed to be, say again?
At the south end of a north-facing cow, naturally.
i agree that Motorola has a lot of problems (i owned a RAZR, and while it was kinda cool at first, the software was terribly crippled)
As you speculated later, the software was indeed crippled by your phone service provider. I purchased an unlocked Mobilinux-based Z6 ROKR directly from Motorola and it's not hobbled in any way. There are no restrictions on ring tones, images, transferring or importing any media. The Linux-based music app works quite well, and certainly doesn't suck like the "Digital Music Player" that comes standard on their KRZR and RAZR lines. The user interface is clean and sharp, much more professional than their previous offerings.
I also have no complaints about the audio quality, and their completely automated voice recognition is excellent, even in noisy environments. (It recognizes names that are entered in the contact list simply by their English pronunciations -- no training required.) And the EDGE speeds clock in about 115kbps -- nothing to complain about.
However, their Bluetooth software is still bug-prone, and the simple act of loading up and using their Java implementation causes my phone to have fits. This is my third Motorola phone in almost as many years, and I'm getting pretty tired of the lack of improvement in these areas.
"Stay Alert! Trust No One! Keep Your Laser Handy!"
and
"Trust The Computer. The Computer is Your Friend."
We can see perfectly clear evidence of evolution in humans.
Go on ..
Nobody ever said evolution would produce a "better" animal, just one "more suited" to breed in the environment in which it lives.
Yeah. Think about that.
Simply looking at their posting, it's likely part of an attempt to spam Google into boosting the page rank for some of those NYT opinion pieces, or maybe to try to relate some of those stories to each other. (I'm not visiting or posting those links out of principle, so it's hard to say for sure.) They could be attacking just the search engine or possibly Google News as well. We can only hope that the Google spiders walking Slashdot ignore all "Score -1" postings.
I'm not sure why they didn't put hrefs around their links, tho. Maybe they don't understand /., or maybe it's just a bot.
The "I'm not buying it" argument against evolution typically comes because humans do not innately have the understanding of the time scales involved, or of the nature of probability.
Humans don't typically pay attention long enough to visibly see evolution taking place. The evidence is there, but it may take effort to put it together, and they're unwilling to do that. (I keep thinking that a month-long process of breeding antibiotic-resistant e. coli and feeding it to them would go a long way toward helping them understand it, but that's just me.) And lots of people don't understand probability -- look at gamblers and their "systems", or even try to explain the Monty Hall problem to someone.
And of course lots of people don't understand simply because they can't or won't make the effort. I have less respect for those people.
If the tests don't provide coverage for ALL the business issues that the piece of software is supposed to solve, then you pass the tests, but will have no clue if you broke or not things apart.
That's not true at all. You have a lot of clues based on the tests that passed. You'll have confidence in the code that passed the tests. Overall, it's less to worry about. It's certainly of more benefit than no tests. And if you failed tests, you'll have 100% confidence that you broke something, and you can get it fixed.
Consider the end goal of unit tests would be 100% assurance that all code paths are covered, and that all behavior is tested. Not that it's realistic in many places, but that is the goal.
Now consider "Legacy Code" as defined by Michael Feathers -- code that does not have unit tests. You have no idea if a single line of it works, other than what your integration or end-user testing shows.
His book is a set of procedures you can try to insert tests into untested or hard-to-test code. His strategy is fairly simple: jam a test into the middle of the code. Chop something in half, place tests on either side of the wound, and see if it still works. Now you have a big ugly gash separating two halves of the code, but you have less code to execute when running the tests on each half. Now, using that line as a starting point, begin refactoring the code, and adding unit tests to prove your refactorings are safe. Sure, you'll still need end user testing and integration testing, but your developers can move forward faster, and you're more confident that your new changes will improve the code, rather than break stuff.
He's trying to help solve a hard problem: how do you refactor (which is inherently a test-driven-development strategy for improving your code) without tests that help ensure you didn't break anything? Once you've got tests, refactoring becomes easier. The more tests you have, the more assurance you have that your changes aren't breaking code. Ultimately, when you are 100% covered, you should be a lot closer to bug-free than you were when you started.
I'll tell you right now that on a large legacy project, most people couldn't tell you where to begin placing tests. Michael's book helps jump start the process. And while the end game might not be 100% test coverage of all the legacy code, it leaves you a lot better off than you were before you started writing the tests.
Define "useful". My phone has a 1600x1200 sensor, and while the lens is certainly no Canon L series, it's able to accurately capture the contents of a whiteboard after a meeting just fine. I have pictures of my friends displayed on the little screen, and I know who they are. And it does take good incriminating shots of drunk pals in the bar!
Sure, if I want good pictures I'm going to bring a real camera. But if I'm just out and about, I won't have the good camera with me -- the thing weighs twelve times the weight of the phone, after all. So I find the phone does just fine for my average point-and-shoot needs.
The OK camera you have with you is far more useful than the great camera you wish you had with you.
So get an Eye-Fi SD card for your own decent digital camera. 2GB storage, and it automagically scans for your network (or any network access point named the same thing, such as LINKSYS :-) and uploads your pictures as soon as it can to your machine or to flickr or wherever.
You don't even have to buy a crappy camera that sucks because it's got a crappy phone in it for no reason. You can buy a decent camera and still get the magic functionality out of it.
Although.. every phone should have gps, since they've all got it built-in anyway so the carriers don't have to put any effort into triangulating.
That's not completely correct.
Many phones that seem to have "GPS" actually have "Assisted-GPS (A-GPS)". A-GPS allows for a cheaper device where the phone has only the receiver but no 'smarts'. These phones send the raw satellite data to a service provider to perform the math required to locate the phone. The service provider is able to fill in a lot of missing data if poor reception at the phone is an issue (a frequent problem in urban canyons), and because it does know tower locations and satellite ephemeris, it can very quickly provide reasonable location data. The service then returns the coordinates to the phone.
While they're more accurate than simple triangulation from the towers, those obviously cost the carrier in terms of providing A-GPS servers.
I have a Motorola Z6, and it's pretty close to your description, (except trade "rugged" for "sleek".) The acoustics are great, and I have no reception problems -- it usually gets "more bars" than an iPhone on the same carrier (AT&T). The voice recognition dialing is automated based on the names you enter in your contact list, and has been remarkably precise. And the screen is crystal clear.
The software, not so much. The Bluetooth gets fairly easily confused, and the headphones don't keep a perfect connection when using the AD2P music player.
I shouldn't say it isn't rugged -- mine looks as good as the day I bought it, apart from a tiny 1/8" scratch on the trim ring -- but it certainly does not have the feel of a rubberized Army field model.
Thank you, this is the most insightful reply in this thread.
I *hate* it when people whine about their phone's features. "Oh, I don't need all that crap, I just need to make a phone call!" They're usually not telling the truth. They were in the cell phone store once upon a time buying a new phone, and they could have accepted the feature-poor 'freebie' phone. But the salesman said "same price, this one has a free camera." Then he said "This one has a better lens, and it's only $20 more." And any phone with a camera comes with a display large enough for a miniature web browser, and that's just software. So these people willingly bought the phones that they then complain about.
I also know that people complain a lot about the CDMA carriers locking their phones down. If you go the Verizon / Sprint route, you're going to have lockdown -- plain and simple. Find a GSM carrier instead, and buy an unlocked phone.
Of course, phones get my goat too. What really cheezes me off is when I buy a phone specifically for a set of features, and the features do not work as advertised. The Bluetooth hangs, or the browser screws up, or the music player takes half an hour to load, or it crashes, or whatever. Motorola is great for that.
I'm a geek. If I want a phone, it's because I want to phone someone, or I want them to phone me. I don't need all this other stuff.
If you're a geek you don't need all this other stuff because you already have it hanging from pouches on your belt or in your cargo pants, right? Palm Tungsten -- check. Pager -- check. Digital Elph camera -- check. Calculator (HP-41C, no doubt) -- check. iPod -- check. GPS -- check.
Maybe it's just me, but the all-in-one phone seems to be a lazier choice. And we geeks are a lazy lot, after all.
Ha!
IBM invented FUD. I've seen them playing hardball, and they were once so good at it that they used to manipulate the outcome of the World Series of Manipulation. IBM has simply learned that FUD could get you so far and no further, and now there's more to be gained by being honest (or at least more honest than Microsoft.)
How fried? Do you mean disable (i.e. temporary blindness), cripple (permanent blindness), burn (seared flesh), or maim (severed limbs or death?) An ordinary mirror will work fine for the lower power attacks (as would a sheet of cardboard!) If you have a mirror made from the right kind of material (such as copper) it'll deflect any of these. But the kind powerful enough to take out a satellite, missile, or weapons platform? I have heard that no ordinary mirror would withstand them, but that copper sheeting could still protect the target.
The more important question is: are you going to be hiding behind a mirror when the weapon strikes? My understanding is the big laser weapons are pulse weapons, meaning all their energy is delivered virtually instantly. It's not like you can watch a laser beam come zapping in and then duck just before it hits you. That only works for Captain Kirk.
If so, what's the purpose of doing this?
Perhaps it's simply marketing. Maybe it's test marketing. Maybe they're trying to gauge the public acceptance of chip delivery systems. Maybe they're trying to see if the commercial public will accept 128kbps MP3s over higher quality formats.
Or maybe because everyone on Slashdot has been screaming "you need to figure out a different sales model!" This is indeed different.
The point is they're trying something new. Anything new. It costs them almost nothing to try, they get to look cool doing it, they successfully generated "buzz" by trying (look, here we are discussing it.) I don't think its any deeper than that.
RTFA. They're disconnecting it on October 25th.
The other related joke I remember from back then was "Has anyone else noticed the quality of lamers and newbies has gone downhill?"
No kidding. If the average kid had to do homework as a part of grinding in WoW, considering how much time they spend in game we might end up with a nation of Doogie Howsers.
Of course, all the asian gold farmers would quickly have their PhDs, and we'd really just end up with a nation of kids all trained to send their money overseas.