IIRC, RFID nowadays has failure rates between 10% and 40% - and even though it would be incredibly revolutionary if i could get an exact tally of my inventory by just walking through the aisles with an RFID reader once, a failure rate of even 5% would be way to high
But what's cheaper, hiring lots of employees to count every single item in the store (probably with close to a 5% error rate) Or hiring one person to walk through the store all day. A 5% error rate for one reading, quickly drops for multiple readings, so you just have one employee walk through all the isles and keep repeating the path. Over the course of the day you can statistically bring that 5% down to a reasonable number.
Alright, but from what other people have pointed out in this discussion RFID has its own set of problems so even if you could get 5 cents per tag, how would you prevent the RFID errors from being just about as bad (i.e. not all tags in the box respond to the ping, certain items in the box interfere with the signals, etc) as the barcodes?
There are two benefits to RFID that would make the problem much smaller than with bar codes. 1) The labor required to read the tag is much smaller. Say you get a palette off the truck and what to make sure it's all there. You just ping it with the reader and it reports what's there. But since you can query ALL of the items in the palette with the same effort, you can make measurement a couple of times. That way a tag has a chance to respond on a seconds or thirds (or 100th) read. and 2) RFID identifies individual items, unlike a bar code which identifies classes of items. So say you read the palette and you only get 49 items and you expected 50 in your palette. By the missing tag, you know exactly what's missing and can either look for it pretty quickly, or tell the shipper to send another one.
I *still* get freaked out every morning when I go downstairs to the kitchen and turn on the fluorescent light. It flickers a little before fully lighting and it reminds me of the flickering lights in doom.
Well initially it wouldn't change anything. But as thieves learn that the resale value of an iPod is essentialy $0 because it can't be recharged, then theft will go down. Of course, it doesn't stop someone from taking it just because they're an asshole and just want to take something from you. But it does stop the thief who's looking for money.
If you know how to have an https connection to Gmail, please let me know. I've never connected to Gmail (or Google Calendar) through https, although I've always wanted to.
what's not to trust about the wholesale price? it's set on the commodities market: http://money.cnn.com/data/commodities/. As of the time of this comment, unleaded gas is going for $2.24/gal
here's an outdated page (from '05), but it took me 3 seconds of googling to find. it says oil profit margins are about 6.8%, close to my 5% estimate.
http://www.gravmag.com/oil.html#dollar
Oil companies make about $0.10/gal profit, which is about a 5% profit margin given the wholesale cost of gasoline being around $2.25/gal as of this morning. That may sound like a bunch, but you need to remember that the Federal Gas Tax is $0.185/gal, and the average state tax is $0.25/gal, not to mention local taxes. So guess who's really raking it in on gasoline sales?
Unless the research is from government funding. All of my funding comes from the NSF or NIST, and since the money is taxpayer's money research that comes from it cannot be copywritten. So in addition to giving the Journal the transfer-of-copyright form, I also send them a form that says technically I have no copyright to transfer to them.
Yes, those machines do measure the mechanical resistance and see what your power output is - and that's the information I'd really want anyway. However, most people want to know how many snickers bars they burnt off, so they calibrate the power output that they measure to the actual amount of energy you're burning, and that's why it's so inaccurate because people cover a wide range of efficiencies.
You're right that we'd be burnt to charcoal if we weren't good at converting chemical energy (fat) into mechanical energy. I meant to say, that the conversion from chemical energy into useful mechanical energy was inefficient. There's lots of mechanical energy that's orthogonal to the useful energy for any given exercise.
Those calories are kilocalories. So your average power is actually 2,800W. But we know that can't be right. First those calorie calculators are very innacurate, they just get you in the ball park. Second, that's the amount of energy you're burning to produce the motion, but your body is very inefficient. So you may be burning 2,800J/s of energy, but the usable output of work could easily be 1% of that. And then as you said, you have to worry about the efficiency of the generator.
Example in point: I can't speak for anywhere else, but in the US and Canada, I understand that when adjusted for inflation, the average income has gone down $800 in the last 20 years.
I've heard similar figures to this too and if it is true, could the reason be that we have more people trying to share the same resources? What I mean is that at the lowest levels economics is really just a way for people with unlimited wants to find a way to distribute limited resources. The world isn't getting any bigger, there's a finite amount of stuff we can pull out of the ground, etc. However the population keeps growing. So we all (on average) get a smaller chuck of the pie. This would make it look like income goes down because you need to pay more for the same stuff you got before now that more people want it.
I'm also going to a Sharp (32D40U) TV but I'm going DVI -> HDMI on my receiver and the HDMI on the receiver -> HDMI on the TV. I don't know if it matters or not, but the Mac Mini reports the name of my reveiver as the display name and not the name of the TV.
That's odd, I have a 1366x768 LCD Flat Panel and the Mac Mini Looks great on it. When I turn off the the overscan, I get a small black border around the edge of the screen so that the signal coming from the Mac Mini actually uses 1280x720 of the pixels on the monitor. Granted, I'm not using all the pixels on the TV, but it's a 1:1 mapping so it's as clear as can be. Are you going DVI from the Mac Mini to HDMI on the TV or are you going from VGA on the Mac Mini to Component on the TV?
Heh Heh. The tow ball on my truck ripped the hood off of a VW Sirocco when he rear ended me after tailgating me on icy roads. My repair bill was $125 for a new bumper (I wouldn't have replaced it, but his insurance was paying, so I took it). His car was totaled.
I think there's a difference between mandating all public buildings to have wheelchair ramps and changing the money so blind people can use it. In the first case, you're making a regulation that affects a private business. If I own a business and don't want to have wheelchair ramps, that may make me a jackass, but it shouldn't be illegal. And disabled people can choose not to do business with me. However, the government owns the monopoly on money. If people are forced to live under our government and use the money that they says is the legal tender, then by all means it should be such that it's usable by as many people as possible with whatever disability. There's no excuse for the government to be issuing money that blind people can't use. Any interaction with a private business should be under the terms of the business and you, any interaction with the government needs to be lowest common denominator so that everyone who has to live under the rule of the government can also get the benefit that their taxes are supposedly paying for.
I forgot about the layering that the GE community can provide. Want to bet how long it will be until someone makes a layer in GE that reads the virtual earth data and redisplays it? I'm not sure how the layers in GE work, but I bet someone just needs to have a server the GE queries and then an automated ActiveX control that will grab the virtual earth data and send it back to GE.
How do we eradicate this problem? What strategies do we use next?
one solution would be to use word pairs or word triples in the same bayesian code. so rather than just tokenizing by single "words", have pairs. this doesn't increase the work required on the client end very much, but it makes is much much harder to come up with valid "phrases". right now, spammers just have to put unspammy words in their emails, but with a small change to the bayesian filter, they'd have to come up with unspammy phrases, the longer the phrase you choose, the larger the search space is for the spammer. i think paul graham mentions this in his article (it's been a while since i read it), and it would be quite easy to implement.
The reason you can't buy a tiny vehicle in the US (other than, for example, the Honda Insight) is that historically tiny vehicles have performed poorly in the market. Our safety regulations are neither "insane" nor are they particularly stringent - EuroNCAP, for example, is a considerably harder test than the NHTSA test (and more comparable to the IIHS tests).
EuroNCAP may be a harder test than the NHTSA test, I don't doubt that, as the NHTSA test is crap (and the results don't mean much anyway). I think (and probalbly the OP since we seem to agree on most of the points in this thread) the point is that the specific laws are dumb. If the government wants to say, "cars must be able to protect the passengers with xx% reliablilty in a head on collision of yy mph...". That's fine with me. What I don't like is when the government says "you must put airbags in a car". They're legislating the wrong thing. If they specify a broad result (like people shouldn't get hurt in car crashes) that's fine, let it up to the individual companies to do research and find the best way to meet the law. There are lots of cars in Europe that are as safe or safer than ours, but don't have airbags (and would cost too much to retrofit them), so we don't get them over here. That's just one example, but there are tons of rediculous laws regulating the auto industry which serve mainly to protect the established players (did you know it costs a car company between 1 and 3 million dollars per model to do the NHTSA testing!)
I don't like the idea of a "safe harbor" or anything like that. If I give my money to a bank and they lose it, even through a "genuine mistake", I get it back. Likewise, I expect that if I give information to a company, and they lose it, they are liable for any harm that comes from that loss. The trouble is that when the governemnt gets involved, then the lawyers at the companies will get involved and they'll look for loopholes and such. There have been a couple of laws passed in the last couple of years that give protection to the companies (Why do you think the submitter was notified of the data loss? Not because the company cares about the submitter, but they get legal protection if they notify of the loss), what we need is to not have those laws and let it up to people to bring civil cases against the companies that lose the data. Yes it will be expensive, but after a few precidents are set, then it'll be easier for the little guy to go after the big companies that lose the info.
I didn't get a chance to read all 650+ comments (and probably no one will read mine since it's posted so late). But a common theme in the 200 or so comments I read was that people ask a question after they already RTFM and are told to RTFM and they get mad. What you really need to do when you ask a question is tell people what you already did. It does two things 1) it lets them know that you are trying to solve the problem - not just expecting to be given the answer and 2) it let them know not to waste their time telling you something you already know. We may think ESR is crazy, but there are two things I really agree with him on 1) his opinion of guns and gun control, and 2) how to ask quesions: http://www.catb.org/~esr/faqs/smart-questions.html. It not just in the linux world that you do this, but anytime you need to ask an expert for help - you tell them what you tried and what didn't work. e.g. when I take my car into the shop for a problem I can't fix I tell the mechanic what I already tried and what the results were, they're usually very appreciatve because it saves them time and they can get to the real problem (and it saves me money in hourly labor charges!).
yeah, i got a copy of that back in the windows 3.1 days, so it was pre 95. i remember that is was shareware and it said that if you bought the full version it would decompress the file if you deleted the original. i never believed it, and if it did really work, then it'd be around today.
on a CD it would damage x bits. Since a dual layer DVD holds about 14 CDs, i'm guessing the same scratch would take out 14x bits.
it may even be worse that that. i know a CD (at least in audio format) has error correcting bits, thats why you can still play a scratched cd with no issues (until the scratch is big enough to screw the error correcting bits too). i don't know about dvd, but i've never heard of any error correcting bits there. maybe an expert can fill us in.
Re:apologies, slightly off-topic...[but only a lit
on
Chemical Words List
·
· Score: 1
well it's because there's really no way to quantify smell. you can talk about color even if you're color blind. all you do is define the wavelengh (or mixtures) and you can uniquely identify a color. likewise you can quanitfy sound even if you're deaf. middle C is 440 Hz (or is it A, i don't know i don't remember my music theory). how do you quantify a smell? the only thing that makes a smell unique (other than your subjective perception) is the molecule itself.
Slashdot Mirror
IIRC, RFID nowadays has failure rates between 10% and 40% - and even though it would be incredibly revolutionary if i could get an exact tally of my inventory by just walking through the aisles with an RFID reader once, a failure rate of even 5% would be way to high
But what's cheaper, hiring lots of employees to count every single item in the store (probably with close to a 5% error rate) Or hiring one person to walk through the store all day. A 5% error rate for one reading, quickly drops for multiple readings, so you just have one employee walk through all the isles and keep repeating the path. Over the course of the day you can statistically bring that 5% down to a reasonable number.
Alright, but from what other people have pointed out in this discussion RFID has its own set of problems so even if you could get 5 cents per tag, how would you prevent the RFID errors from being just about as bad (i.e. not all tags in the box respond to the ping, certain items in the box interfere with the signals, etc) as the barcodes?
There are two benefits to RFID that would make the problem much smaller than with bar codes. 1) The labor required to read the tag is much smaller. Say you get a palette off the truck and what to make sure it's all there. You just ping it with the reader and it reports what's there. But since you can query ALL of the items in the palette with the same effort, you can make measurement a couple of times. That way a tag has a chance to respond on a seconds or thirds (or 100th) read. and 2) RFID identifies individual items, unlike a bar code which identifies classes of items. So say you read the palette and you only get 49 items and you expected 50 in your palette. By the missing tag, you know exactly what's missing and can either look for it pretty quickly, or tell the shipper to send another one.
I *still* get freaked out every morning when I go downstairs to the kitchen and turn on the fluorescent light. It flickers a little before fully lighting and it reminds me of the flickering lights in doom.
Well initially it wouldn't change anything. But as thieves learn that the resale value of an iPod is essentialy $0 because it can't be recharged, then theft will go down. Of course, it doesn't stop someone from taking it just because they're an asshole and just want to take something from you. But it does stop the thief who's looking for money.
If you know how to have an https connection to Gmail, please let me know. I've never connected to Gmail (or Google Calendar) through https, although I've always wanted to.
what's not to trust about the wholesale price? it's set on the commodities market: http://money.cnn.com/data/commodities/. As of the time of this comment, unleaded gas is going for $2.24/gal
here's an outdated page (from '05), but it took me 3 seconds of googling to find. it says oil profit margins are about 6.8%, close to my 5% estimate. http://www.gravmag.com/oil.html#dollar
Oil companies make about $0.10/gal profit, which is about a 5% profit margin given the wholesale cost of gasoline being around $2.25/gal as of this morning. That may sound like a bunch, but you need to remember that the Federal Gas Tax is $0.185/gal, and the average state tax is $0.25/gal, not to mention local taxes. So guess who's really raking it in on gasoline sales?
Unless the research is from government funding. All of my funding comes from the NSF or NIST, and since the money is taxpayer's money research that comes from it cannot be copywritten. So in addition to giving the Journal the transfer-of-copyright form, I also send them a form that says technically I have no copyright to transfer to them.
Yes, those machines do measure the mechanical resistance and see what your power output is - and that's the information I'd really want anyway. However, most people want to know how many snickers bars they burnt off, so they calibrate the power output that they measure to the actual amount of energy you're burning, and that's why it's so inaccurate because people cover a wide range of efficiencies.
You're right that we'd be burnt to charcoal if we weren't good at converting chemical energy (fat) into mechanical energy. I meant to say, that the conversion from chemical energy into useful mechanical energy was inefficient. There's lots of mechanical energy that's orthogonal to the useful energy for any given exercise.
Those calories are kilocalories. So your average power is actually 2,800W. But we know that can't be right. First those calorie calculators are very innacurate, they just get you in the ball park. Second, that's the amount of energy you're burning to produce the motion, but your body is very inefficient. So you may be burning 2,800J/s of energy, but the usable output of work could easily be 1% of that. And then as you said, you have to worry about the efficiency of the generator.
I've heard similar figures to this too and if it is true, could the reason be that we have more people trying to share the same resources? What I mean is that at the lowest levels economics is really just a way for people with unlimited wants to find a way to distribute limited resources. The world isn't getting any bigger, there's a finite amount of stuff we can pull out of the ground, etc. However the population keeps growing. So we all (on average) get a smaller chuck of the pie. This would make it look like income goes down because you need to pay more for the same stuff you got before now that more people want it.
I'm also going to a Sharp (32D40U) TV but I'm going DVI -> HDMI on my receiver and the HDMI on the receiver -> HDMI on the TV. I don't know if it matters or not, but the Mac Mini reports the name of my reveiver as the display name and not the name of the TV.
That's odd, I have a 1366x768 LCD Flat Panel and the Mac Mini Looks great on it. When I turn off the the overscan, I get a small black border around the edge of the screen so that the signal coming from the Mac Mini actually uses 1280x720 of the pixels on the monitor. Granted, I'm not using all the pixels on the TV, but it's a 1:1 mapping so it's as clear as can be. Are you going DVI from the Mac Mini to HDMI on the TV or are you going from VGA on the Mac Mini to Component on the TV?
Heh Heh. The tow ball on my truck ripped the hood off of a VW Sirocco when he rear ended me after tailgating me on icy roads. My repair bill was $125 for a new bumper (I wouldn't have replaced it, but his insurance was paying, so I took it). His car was totaled.
I think there's a difference between mandating all public buildings to have wheelchair ramps and changing the money so blind people can use it. In the first case, you're making a regulation that affects a private business. If I own a business and don't want to have wheelchair ramps, that may make me a jackass, but it shouldn't be illegal. And disabled people can choose not to do business with me. However, the government owns the monopoly on money. If people are forced to live under our government and use the money that they says is the legal tender, then by all means it should be such that it's usable by as many people as possible with whatever disability. There's no excuse for the government to be issuing money that blind people can't use. Any interaction with a private business should be under the terms of the business and you, any interaction with the government needs to be lowest common denominator so that everyone who has to live under the rule of the government can also get the benefit that their taxes are supposedly paying for.
I forgot about the layering that the GE community can provide. Want to bet how long it will be until someone makes a layer in GE that reads the virtual earth data and redisplays it? I'm not sure how the layers in GE work, but I bet someone just needs to have a server the GE queries and then an automated ActiveX control that will grab the virtual earth data and send it back to GE.
one solution would be to use word pairs or word triples in the same bayesian code. so rather than just tokenizing by single "words", have pairs. this doesn't increase the work required on the client end very much, but it makes is much much harder to come up with valid "phrases". right now, spammers just have to put unspammy words in their emails, but with a small change to the bayesian filter, they'd have to come up with unspammy phrases, the longer the phrase you choose, the larger the search space is for the spammer. i think paul graham mentions this in his article (it's been a while since i read it), and it would be quite easy to implement.
EuroNCAP may be a harder test than the NHTSA test, I don't doubt that, as the NHTSA test is crap (and the results don't mean much anyway). I think (and probalbly the OP since we seem to agree on most of the points in this thread) the point is that the specific laws are dumb. If the government wants to say, "cars must be able to protect the passengers with xx% reliablilty in a head on collision of yy mph...". That's fine with me. What I don't like is when the government says "you must put airbags in a car". They're legislating the wrong thing. If they specify a broad result (like people shouldn't get hurt in car crashes) that's fine, let it up to the individual companies to do research and find the best way to meet the law. There are lots of cars in Europe that are as safe or safer than ours, but don't have airbags (and would cost too much to retrofit them), so we don't get them over here. That's just one example, but there are tons of rediculous laws regulating the auto industry which serve mainly to protect the established players (did you know it costs a car company between 1 and 3 million dollars per model to do the NHTSA testing!)
I don't like the idea of a "safe harbor" or anything like that. If I give my money to a bank and they lose it, even through a "genuine mistake", I get it back. Likewise, I expect that if I give information to a company, and they lose it, they are liable for any harm that comes from that loss. The trouble is that when the governemnt gets involved, then the lawyers at the companies will get involved and they'll look for loopholes and such. There have been a couple of laws passed in the last couple of years that give protection to the companies (Why do you think the submitter was notified of the data loss? Not because the company cares about the submitter, but they get legal protection if they notify of the loss), what we need is to not have those laws and let it up to people to bring civil cases against the companies that lose the data. Yes it will be expensive, but after a few precidents are set, then it'll be easier for the little guy to go after the big companies that lose the info.
I didn't get a chance to read all 650+ comments (and probably no one will read mine since it's posted so late). But a common theme in the 200 or so comments I read was that people ask a question after they already RTFM and are told to RTFM and they get mad. What you really need to do when you ask a question is tell people what you already did. It does two things 1) it lets them know that you are trying to solve the problem - not just expecting to be given the answer and 2) it let them know not to waste their time telling you something you already know. We may think ESR is crazy, but there are two things I really agree with him on 1) his opinion of guns and gun control, and 2) how to ask quesions: http://www.catb.org/~esr/faqs/smart-questions.html . It not just in the linux world that you do this, but anytime you need to ask an expert for help - you tell them what you tried and what didn't work. e.g. when I take my car into the shop for a problem I can't fix I tell the mechanic what I already tried and what the results were, they're usually very appreciatve because it saves them time and they can get to the real problem (and it saves me money in hourly labor charges!).
yeah, i got a copy of that back in the windows 3.1 days, so it was pre 95. i remember that is was shareware and it said that if you bought the full version it would decompress the file if you deleted the original. i never believed it, and if it did really work, then it'd be around today.
on a CD it would damage x bits. Since a dual layer DVD holds about 14 CDs, i'm guessing the same scratch would take out 14x bits.
it may even be worse that that. i know a CD (at least in audio format) has error correcting bits, thats why you can still play a scratched cd with no issues (until the scratch is big enough to screw the error correcting bits too). i don't know about dvd, but i've never heard of any error correcting bits there. maybe an expert can fill us in.
but emacs is an operating system...*duck*
well it's because there's really no way to quantify smell. you can talk about color even if you're color blind. all you do is define the wavelengh (or mixtures) and you can uniquely identify a color. likewise you can quanitfy sound even if you're deaf. middle C is 440 Hz (or is it A, i don't know i don't remember my music theory). how do you quantify a smell? the only thing that makes a smell unique (other than your subjective perception) is the molecule itself.