Carla Speed McNeil's "Finder" books are brilliant. She writes and draws them, and self-publishes. She's also done art for some other books (I know she did a run on Greg Rucka's "Queen and Country", for one.)
All of which goes to show why I prefer a hammer that can be used to nail together any sort of structure over a fancy gadget with a single "build doghouse" button.
Hammers are nice, but when all I really want is to build a doghouse, I like that "build doghouse" button.
Similarly, when I'm writing brand new code, I like having the ability to do dynamic linking, change various little bits, and so on. When I just want a word processing application to install, I like having the "install application" button.
A better (more expensive) design WOULD eliminate a lot of those false positives, but none of the actual devices available seem to use that better design. In other words, we already KNOW they willingly sacrifice accuracy to lower costs. To me, that is a good sign that I should place little trust in the manufacturer's intentions to produce a reliable machine.
You realize *all* test equipment has failure modes, right? As long as they're known and stated, there is nothing wrong with that; it isn't evidence of the manufacturer not wanting to produce a reliable machine, it's evidence of the manufacturer wanting to produce a reasonably reliable machine at a reasonable cost point, with acknowledgement that, like everything else we humans build, it isn't perfect.
The problem isn't with a breathalyzer capable of false positives. It's with a breathalyzer capable of false positives, where the circumstances that cause false positives are kept hidden.
A defendant facing trial on DUI charges has a right to examine each and every one of those possabilities and to rebut the various marketing claims stating that the whizzbang 2000 breathalyser is infallable.
Yep. But they don't need schematics or source code or any other detailed design information to do it, which has been my point all along. Adequate test records should show the known failure modes. In the course of test records, it is usually possible to deduce the mode of operation, so hiding that is sort of silly.
If the test records are inadequate, that is in and of itself a sufficient reason to reject *any* evidence produced by the suspect equipment.
For that particular example, yes, you'd have to show some reasonable probability. Do note though, sitting in direct sunlight on a hot southern day COULD send the device out of range. It would also be enough to send resistors and capacitors out of tolerance.
No offense, but you obviously don't understand how professional circuit designers work, especially on systems where accuracy is desirable/vital, which is true for any calibrated test gear. We know exactly what tolerance over temperature is for a device, and the design is generally done in such a fashion that within the specified operating range, the temperature induced change combined with the natural tolerance of the components will not throw anything outside of the specified accuracy. Of course, this assumes someone who cares about the quality of their product; I've kicked the asses of more than a couple circuit designers where I work for having the nerve to turn in a design that didn't meet spec at nominal, much less at temperature. Further, any design we've done that is significantly temperature-affected is not only temperature-compensated to the greatest degree possible, it is also designed to report a fault if the temperature is out of range.
It may not slip a reasonable effort at testing, but if breathalysers are built to the same standards as many radar guns, that doesn't mean much in practice.
Most radar guns are just fine, as long as they get a reasonable calibration routine. The problem is generally in the police's hands, as they don't tend to keep their cals up to date.
I see little reason to expect that a breathalyser is anything like accurate without good evidence and a careful examination.
And I see little reason to believe that a breathalyzer, backed up by design verification test records from the manufacturer and calibration records from the operator, is not accurate. Without those two items, I'd have my doubts, but with them, it is entirely likely that the machine *is* operating properly.
You also have to show the probability of a bit flip and that said probability is sufficient to constitute reasonable doubt. The probability of a bit flip within the operating range (for commercial hardware, this is usually quoted as -25 to +75C, and even with a 15C enclosure rise, it'd have to be a pretty goddamn hot day to go outside operating range) being as small as it is, it does *not* constitute reasonable doubt. Failure modes large enough to constitute reasonable doubt have a very hard time slipping by anything resembling a reasonable effort at testing.
If you have the technical details, you still won't have adequate information to determine all failure modes. You won't get code, so you won't be able to find race conditions, and even if you have schematics and code, a reasonably complex system can *still* generate failure modes you didn't expect. The only way to try to cover all failure modes is, in fact, via a combination of analytical procedures using the design to analyze failure modes and a battery of tests designed to find the unexpected modes. Since you have no way to determine that the manufacturer found all modes *whether or not you have details of equipment*, your argument fails to hold water.
Further, if you can see the case studies/test results, you'll have a general idea of method and technique used, and you can look at the test parameters to determine if test coverage was adequate to have been a reasonable attempt at finding all failure modes. The technical details, although nice, would simply be a gloss; given general knowledge of method (I need to know that it's an electrical test using conductivity rather than a chemical one; I don't need to know that they're using a Freescale DSP and that the third instruction in the program is a HCF opcode) and detailed knowledge of test procedure, I can accurately evaluate likelihood of failure.
Spun off, yes. Motorola proper is no longer in the chipmaking business, with certain ASIC exceptions in the realms of network switching and automotive electronics, and those are run on a fabless model (and often enough, built by Freescale, though not always.)
Analog processors not only do exist, but have existed for longer than digital. They're just a bitch to keep calibrated (yes, I have worked with analog computers; they're great for solving differential equations where you just need an engineering solution. Also, I am 23. So this is not a "back in my day" thing.) for any lengthy period of time.
Fun fact: analog synthesizers are basically analog computers, with certain specialized computations added in for convenience. And analog computers can be used as synths. It's all just math, in the end.
Depends. If you like explosive reactions, adding elemental sodium to a (presumably H2O-containing) biological sample, it may be *exactly* the result you want.
It may, in fact, be the result you want all over your lab, depending on how much you added.
the evidence accurately represents what actually happened.
evidence that is presented is accurate and that all the equipment used as evidence is functioning and being maintained properly
Correct me if I'm missing something, but did you just call bullshit on the grandparent and proceed to restate what he said in the first place?
The public has *no* need to know how the equipment works, even under your (arguably different) standard. All that's necessary is that the police can prove that the equipment functions accurately when properly calibrated (proven via case studies) and that the equipment was properly calibrated (proven via calibration logs). The mechanism is *not* essential public knowledge.
So why do I care whether a scientist is Christian, Buddhist, Muslim, etc? And if the scientist chooses to reveal that irrelevant fact to me, am I not right in having suspicions of an alterior motive in his/her points of view?
a) Christian Science is the name of the church; CSM is not a scientific paper, it just happens to have very solid coverage of science news. In and of itself, the Science portion of the name has essentially no connection to the issue.
b) Don't you owe it to yourself to check out whether those suspicions are at all valid? Especially when the website of the paper has a number of responses in their FAQ to just the issues you raise? Considering that CSM has won seven Pulitzer Prizes, it's pretty safe to say that they're doing something reasonably well.
Are you right in being suspicious? No. Not based on a name and a name alone, with no evidence of ulterior motives and significant evidence of validity and worth in the work produced.
Because it was founded by members of the Christian Science church. Being a member is not a requirement for working on the paper, but for historical reasons the name is retained; you know, kinda like how the Manchester Union-Leader spent quite some time not even remotely representing the views of the union, choosing instead to become a paper advocating practically fascist viewpoints. The YMCA has Christian in the name, despite being an essentially secular organization these days.
What's in a name, anyway? The Daily Show isn't actually daily; you gonna blame them for that? CNN is the Cable News Network, but a good portion of us read cnn.com; is it wrong of them to retain the title for their online, non-cable-affiliated branch? The Guardian doesn't guard shit, and BBC America is a contradiction in terms. It's a fucking title. They're all just fucking names!
The paper was started by Christian Scientists because they were being subjected to attack journalism in mainstream papers; since then, they've maintained a pretty good record as a general news source. Before writing them off, why don't you, I don't know, check into them? See if they have a decent rep as a good source of information? Rather than immediately writing them off because their name contains Christian?
If they were actually promoting the Christian Scientist viewpoint in their articles, I would be the first to write them off. They don't. Their science journalism is really as good as it gets in general news sources. Seriously. Get over your bigotry and give them a chance.
Hey, why don't we? Oh, because they don't have a record of being a reasonably good source of news on various subjects, INCLUDING CULTURE WHICH WOULD BE THE SUBJECT GAMING FALLS INTO.
If Carmack wants to start blogging on social issues, I'd be the last person to stop him. Doesn't mean I'd read it, but he's welcome to. If he proved to be an insightful, accurate writer, hey, maybe I'd read him. The CSM has already proven they're a reasonably accurate and fairly unbiased news source. Let's try to seperate the name of the paper from the content.
Not sure what the hell kind of non-profit you're talking about, but all the ones I've worked with, 24k was entry level (this was in SE Michigan, where 24k will let you survive; not well, but survive) and the president of the agency (mind you, the top administrator for a 400 full-time employee non-profit) made 80k at one, and (for around 600 employees) 100k at the other. Oh, and they were provided with a leased company car, on a 36 month lease, so maybe tack an additional 3k per year onto that. (They were generally cars on the order of a Chevy Malibu or Ford 500, and GM/Ford gave them at supplier discount.)
80k for running a 400 employee company is by no means above scale for the area.
Although this is like comparing apples and oranges
Exactly. I was comparing the modulation method used in standard CW Morse techniques with the equivalent communications bandwidth (not channel bandwidth) used in other (non DSB-SC) methods. Morse coding as a coding scheme, as opposed to the use of it as a modulation/transmission scheme, is incredibly inefficient and has next to no advantages over a properly constructed code other than usage history. In addition, while CW Morse can use a Class C amplifier, it will suffer greater effects regarding energy dissipation in the transmission circuit as compared to similar effects in an angle-modulation setup, where dissipation becomes minimal.
Thus, the bandwidth consumed by most common methods of sending morse code is really an issue of the keying nature of the transmitter --not the protocol itself.
You've just restated what I said originally. An "ideal" CW Morse signal is DSB-SC; like any SC signal, the odds of managing to fully suppress the carrier are low. The bandwidth consumed by a DSB-SC Morse scheme is solely a function of the message, which in this case is partially a function of the keying; this is determined by the protocol, as all DSB-SC systems have a bandwidth of 2x message bandwidth. As I originally stated, in DSB-SC Morse, the bandwidth consumed is dependent on the message bandwidth; an abrupt pulse requires a higher message bandwidth, plainly visible from the Fourier transform of the abrupt signal, which leads to a wider modulated bandwidth. A low rate pulse consumes little bandwidth, but carries little information, and is in addition difficult for a human to demodulate. We're in agreement here, even if you don't seem to think so.
Considering the protocol itself, if you knew when a dot or a dot space were about to happen, and you integrated over each individual period to measure the total energy of each element, then you're doing coherent, synchronous detection. Synchronous reception of a DSB signal with a DSB receiver does not have the overall link s/n penalty you'd get by receiving a DSB signal with an SSB receiver.
Actually, you wouldn't be doing coherent detection, only synchronous. In order to do coherent detection, you'd need to regenerate the carrier frequency locally with matched phase and downmix using that, directly reproducing your message signal. Integration over the period is equivalent to using an envelope detector, an inherently non-coherent method. (CCW is actually synchronous by design, not coherent).
Furthermore, you seem to be assuming that a human receiver is a single sideband receiver. Not true. A morse operator knows what speed of morse code is being sent. They integrate each dot and dash in their heads depending on what they're expecting with each dot rhythm beat. In other words, they're listening for the rhythms of various letters. It is not unreasonable to think of the human ear/brain combination as a synchronous reception device.
Actually, no. I was thinking of an asynchronous envelope detection method (as I believe were you, which is, for the record, useless for doing SSB demod) which would commonly be used for DSB-TC reception. It happens to work for DSB-SC/CW methods as well, so long as those methods maintain a modulation index = 1; happily, CW Morse has a modulation index of exactly 1.
Bandwidth of Morse is efficient, though dependent on the signaling rate; like any AM scheme, the occupied bandwidth expands linearly with the transmission bandwidth.
However, Morse transmitters are relatively inefficient in terms of their power usage, because they have to switch current on and off fairly rapidly (for the same reason, AM transmitters are generally less efficient than FM transmitters). This switching requires a less efficient amplifier design than that permitted by an FM transmitter, although the problem is less severe with Morse than with pure AM, since linearity isn't a concern. In addition, in Morse code (DSB-SC) all of the energy at the image frequency is wasted; a SSB scheme is more efficient, albeit harder to detect.
It is, however, the simplest modulated transmission method ever devised (a baseband transmission is simpler, but totally uninteresting.)
Then again, for certain types of games, a good gamepad really is better than the standard mouse-and-keyboard. Fighting games, sports games, and platformers all do better with a gamepad than mouse/keyboard.
The factor of 10 was really a statement regarding people who do understand why CRTs are better for some things. They'll switch if the cost scale becomes prohibitive, but for them, a factor of 10 is probably a decent assessment of what would be required to force a switch. For most people, I think that when LCDs are less than around 25-50% more than a roughly "equivalent" CRT (depending on the value a given person puts on aesthetics and desk space, and the desire to avoid fatty grunts) most people - meaning people who aren't interested in color, black level, and the like - will switch. But the people who deal in black level and color calibration will require more convincing.
It comes down to finding the right price - the price at which most people will find it worth their effort to bring it back for recycling. The rich lazy fuckers will be taken care of by poor people who need the cash, which may not be a very proud way of life, but it's better than starving.
Besides, when I was a broke Michigan student, it wasn't that hard to just walk down the street and see who had thrown out bags of returnable bottles. It paid for more than a couple dinners. (I have pride, but I can't eat it.)
Slashdot Mirror
Carla Speed McNeil's "Finder" books are brilliant. She writes and draws them, and self-publishes. She's also done art for some other books (I know she did a run on Greg Rucka's "Queen and Country", for one.)
Stop reading Marvel books, and you'll find plenty of realistically drawn characters. Including (*gasp*) women with realistic bodies.
All of which goes to show why I prefer a hammer that can be used to nail together any sort of structure over a fancy gadget with a single "build doghouse" button.
Hammers are nice, but when all I really want is to build a doghouse, I like that "build doghouse" button.
Similarly, when I'm writing brand new code, I like having the ability to do dynamic linking, change various little bits, and so on. When I just want a word processing application to install, I like having the "install application" button.
Flexibility is overrated, sometimes.
A better (more expensive) design WOULD eliminate a lot of those false positives, but none of the actual devices available seem to use that better design. In other words, we already KNOW they willingly sacrifice accuracy to lower costs. To me, that is a good sign that I should place little trust in the manufacturer's intentions to produce a reliable machine.
You realize *all* test equipment has failure modes, right? As long as they're known and stated, there is nothing wrong with that; it isn't evidence of the manufacturer not wanting to produce a reliable machine, it's evidence of the manufacturer wanting to produce a reasonably reliable machine at a reasonable cost point, with acknowledgement that, like everything else we humans build, it isn't perfect.
The problem isn't with a breathalyzer capable of false positives. It's with a breathalyzer capable of false positives, where the circumstances that cause false positives are kept hidden.
A defendant facing trial on DUI charges has a right to examine each and every one of those possabilities and to rebut the various marketing claims stating that the whizzbang 2000 breathalyser is infallable.
Yep. But they don't need schematics or source code or any other detailed design information to do it, which has been my point all along. Adequate test records should show the known failure modes. In the course of test records, it is usually possible to deduce the mode of operation, so hiding that is sort of silly.
If the test records are inadequate, that is in and of itself a sufficient reason to reject *any* evidence produced by the suspect equipment.
For that particular example, yes, you'd have to show some reasonable probability. Do note though, sitting in direct sunlight on a hot southern day COULD send the device out of range. It would also be enough to send resistors and capacitors out of tolerance.
No offense, but you obviously don't understand how professional circuit designers work, especially on systems where accuracy is desirable/vital, which is true for any calibrated test gear. We know exactly what tolerance over temperature is for a device, and the design is generally done in such a fashion that within the specified operating range, the temperature induced change combined with the natural tolerance of the components will not throw anything outside of the specified accuracy. Of course, this assumes someone who cares about the quality of their product; I've kicked the asses of more than a couple circuit designers where I work for having the nerve to turn in a design that didn't meet spec at nominal, much less at temperature. Further, any design we've done that is significantly temperature-affected is not only temperature-compensated to the greatest degree possible, it is also designed to report a fault if the temperature is out of range.
It may not slip a reasonable effort at testing, but if breathalysers are built to the same standards as many radar guns, that doesn't mean much in practice.
Most radar guns are just fine, as long as they get a reasonable calibration routine. The problem is generally in the police's hands, as they don't tend to keep their cals up to date.
I see little reason to expect that a breathalyser is anything like accurate without good evidence and a careful examination.
And I see little reason to believe that a breathalyzer, backed up by design verification test records from the manufacturer and calibration records from the operator, is not accurate. Without those two items, I'd have my doubts, but with them, it is entirely likely that the machine *is* operating properly.
You also have to show the probability of a bit flip and that said probability is sufficient to constitute reasonable doubt. The probability of a bit flip within the operating range (for commercial hardware, this is usually quoted as -25 to +75C, and even with a 15C enclosure rise, it'd have to be a pretty goddamn hot day to go outside operating range) being as small as it is, it does *not* constitute reasonable doubt. Failure modes large enough to constitute reasonable doubt have a very hard time slipping by anything resembling a reasonable effort at testing.
[cluebat]Byline, not biline.[/cluebat]
If you have the technical details, you still won't have adequate information to determine all failure modes. You won't get code, so you won't be able to find race conditions, and even if you have schematics and code, a reasonably complex system can *still* generate failure modes you didn't expect. The only way to try to cover all failure modes is, in fact, via a combination of analytical procedures using the design to analyze failure modes and a battery of tests designed to find the unexpected modes. Since you have no way to determine that the manufacturer found all modes *whether or not you have details of equipment*, your argument fails to hold water.
Further, if you can see the case studies/test results, you'll have a general idea of method and technique used, and you can look at the test parameters to determine if test coverage was adequate to have been a reasonable attempt at finding all failure modes. The technical details, although nice, would simply be a gloss; given general knowledge of method (I need to know that it's an electrical test using conductivity rather than a chemical one; I don't need to know that they're using a Freescale DSP and that the third instruction in the program is a HCF opcode) and detailed knowledge of test procedure, I can accurately evaluate likelihood of failure.
1919.
OK, he didn't build it, but his ideas were *tested*, which is equally important.
Spun off, yes. Motorola proper is no longer in the chipmaking business, with certain ASIC exceptions in the realms of network switching and automotive electronics, and those are run on a fabless model (and often enough, built by Freescale, though not always.)
Well, considering we won't let humans do that...
Analog processors not only do exist, but have existed for longer than digital. They're just a bitch to keep calibrated (yes, I have worked with analog computers; they're great for solving differential equations where you just need an engineering solution. Also, I am 23. So this is not a "back in my day" thing.) for any lengthy period of time.
Fun fact: analog synthesizers are basically analog computers, with certain specialized computations added in for convenience. And analog computers can be used as synths. It's all just math, in the end.
Depends. If you like explosive reactions, adding elemental sodium to a (presumably H2O-containing) biological sample, it may be *exactly* the result you want.
It may, in fact, be the result you want all over your lab, depending on how much you added.
the evidence accurately represents what actually happened.
evidence that is presented is accurate and that all the equipment used as evidence is functioning and being maintained properly
Correct me if I'm missing something, but did you just call bullshit on the grandparent and proceed to restate what he said in the first place?
The public has *no* need to know how the equipment works, even under your (arguably different) standard. All that's necessary is that the police can prove that the equipment functions accurately when properly calibrated (proven via case studies) and that the equipment was properly calibrated (proven via calibration logs). The mechanism is *not* essential public knowledge.
I'm a non-religious Jew, if it makes you happy.
So why do I care whether a scientist is Christian, Buddhist, Muslim, etc? And if the scientist chooses to reveal that irrelevant fact to me, am I not right in having suspicions of an alterior motive in his/her points of view?
a) Christian Science is the name of the church; CSM is not a scientific paper, it just happens to have very solid coverage of science news. In and of itself, the Science portion of the name has essentially no connection to the issue.
b) Don't you owe it to yourself to check out whether those suspicions are at all valid? Especially when the website of the paper has a number of responses in their FAQ to just the issues you raise? Considering that CSM has won seven Pulitzer Prizes, it's pretty safe to say that they're doing something reasonably well.
Are you right in being suspicious? No. Not based on a name and a name alone, with no evidence of ulterior motives and significant evidence of validity and worth in the work produced.
Because it was founded by members of the Christian Science church. Being a member is not a requirement for working on the paper, but for historical reasons the name is retained; you know, kinda like how the Manchester Union-Leader spent quite some time not even remotely representing the views of the union, choosing instead to become a paper advocating practically fascist viewpoints. The YMCA has Christian in the name, despite being an essentially secular organization these days.
What's in a name, anyway? The Daily Show isn't actually daily; you gonna blame them for that? CNN is the Cable News Network, but a good portion of us read cnn.com; is it wrong of them to retain the title for their online, non-cable-affiliated branch? The Guardian doesn't guard shit, and BBC America is a contradiction in terms. It's a fucking title. They're all just fucking names!
The paper was started by Christian Scientists because they were being subjected to attack journalism in mainstream papers; since then, they've maintained a pretty good record as a general news source. Before writing them off, why don't you, I don't know, check into them? See if they have a decent rep as a good source of information? Rather than immediately writing them off because their name contains Christian?
If they were actually promoting the Christian Scientist viewpoint in their articles, I would be the first to write them off. They don't. Their science journalism is really as good as it gets in general news sources. Seriously. Get over your bigotry and give them a chance.
Hey, why don't we? Oh, because they don't have a record of being a reasonably good source of news on various subjects, INCLUDING CULTURE WHICH WOULD BE THE SUBJECT GAMING FALLS INTO.
If Carmack wants to start blogging on social issues, I'd be the last person to stop him. Doesn't mean I'd read it, but he's welcome to. If he proved to be an insightful, accurate writer, hey, maybe I'd read him. The CSM has already proven they're a reasonably accurate and fairly unbiased news source. Let's try to seperate the name of the paper from the content.
MATLAB on Cell would be hot. Well. Assuming Cell implements fast double precision floating point math.
Seriously. MATLAB is about as non-branchy as you get on a PC, but the vectorization available on a Cell system would smoke.
Not sure what the hell kind of non-profit you're talking about, but all the ones I've worked with, 24k was entry level (this was in SE Michigan, where 24k will let you survive; not well, but survive) and the president of the agency (mind you, the top administrator for a 400 full-time employee non-profit) made 80k at one, and (for around 600 employees) 100k at the other. Oh, and they were provided with a leased company car, on a 36 month lease, so maybe tack an additional 3k per year onto that. (They were generally cars on the order of a Chevy Malibu or Ford 500, and GM/Ford gave them at supplier discount.)
80k for running a 400 employee company is by no means above scale for the area.
Although this is like comparing apples and oranges
Exactly. I was comparing the modulation method used in standard CW Morse techniques with the equivalent communications bandwidth (not channel bandwidth) used in other (non DSB-SC) methods. Morse coding as a coding scheme, as opposed to the use of it as a modulation/transmission scheme, is incredibly inefficient and has next to no advantages over a properly constructed code other than usage history. In addition, while CW Morse can use a Class C amplifier, it will suffer greater effects regarding energy dissipation in the transmission circuit as compared to similar effects in an angle-modulation setup, where dissipation becomes minimal.
Thus, the bandwidth consumed by most common methods of sending morse code is really an issue of the keying nature of the transmitter --not the protocol itself.
You've just restated what I said originally. An "ideal" CW Morse signal is DSB-SC; like any SC signal, the odds of managing to fully suppress the carrier are low. The bandwidth consumed by a DSB-SC Morse scheme is solely a function of the message, which in this case is partially a function of the keying; this is determined by the protocol, as all DSB-SC systems have a bandwidth of 2x message bandwidth. As I originally stated, in DSB-SC Morse, the bandwidth consumed is dependent on the message bandwidth; an abrupt pulse requires a higher message bandwidth, plainly visible from the Fourier transform of the abrupt signal, which leads to a wider modulated bandwidth. A low rate pulse consumes little bandwidth, but carries little information, and is in addition difficult for a human to demodulate. We're in agreement here, even if you don't seem to think so.
Considering the protocol itself, if you knew when a dot or a dot space were about to happen, and you integrated over each individual period to measure the total energy of each element, then you're doing coherent, synchronous detection. Synchronous reception of a DSB signal with a DSB receiver does not have the overall link s/n penalty you'd get by receiving a DSB signal with an SSB receiver.
Actually, you wouldn't be doing coherent detection, only synchronous. In order to do coherent detection, you'd need to regenerate the carrier frequency locally with matched phase and downmix using that, directly reproducing your message signal. Integration over the period is equivalent to using an envelope detector, an inherently non-coherent method. (CCW is actually synchronous by design, not coherent).
Furthermore, you seem to be assuming that a human receiver is a single sideband receiver. Not true. A morse operator knows what speed of morse code is being sent. They integrate each dot and dash in their heads depending on what they're expecting with each dot rhythm beat. In other words, they're listening for the rhythms of various letters. It is not unreasonable to think of the human ear/brain combination as a synchronous reception device.
Actually, no. I was thinking of an asynchronous envelope detection method (as I believe were you, which is, for the record, useless for doing SSB demod) which would commonly be used for DSB-TC reception. It happens to work for DSB-SC/CW methods as well, so long as those methods maintain a modulation index = 1; happily, CW Morse has a modulation index of exactly 1.
Bandwidth of Morse is efficient, though dependent on the signaling rate; like any AM scheme, the occupied bandwidth expands linearly with the transmission bandwidth.
However, Morse transmitters are relatively inefficient in terms of their power usage, because they have to switch current on and off fairly rapidly (for the same reason, AM transmitters are generally less efficient than FM transmitters). This switching requires a less efficient amplifier design than that permitted by an FM transmitter, although the problem is less severe with Morse than with pure AM, since linearity isn't a concern. In addition, in Morse code (DSB-SC) all of the energy at the image frequency is wasted; a SSB scheme is more efficient, albeit harder to detect.
It is, however, the simplest modulated transmission method ever devised (a baseband transmission is simpler, but totally uninteresting.)
Then again, for certain types of games, a good gamepad really is better than the standard mouse-and-keyboard. Fighting games, sports games, and platformers all do better with a gamepad than mouse/keyboard.
What he is proposing is called Frequency Shift Keying (FSK), and is somewhat commonly used by (amongst other things) IR remote controls.
Not for morse, though.
The factor of 10 was really a statement regarding people who do understand why CRTs are better for some things. They'll switch if the cost scale becomes prohibitive, but for them, a factor of 10 is probably a decent assessment of what would be required to force a switch. For most people, I think that when LCDs are less than around 25-50% more than a roughly "equivalent" CRT (depending on the value a given person puts on aesthetics and desk space, and the desire to avoid fatty grunts) most people - meaning people who aren't interested in color, black level, and the like - will switch. But the people who deal in black level and color calibration will require more convincing.
It comes down to finding the right price - the price at which most people will find it worth their effort to bring it back for recycling. The rich lazy fuckers will be taken care of by poor people who need the cash, which may not be a very proud way of life, but it's better than starving.
Besides, when I was a broke Michigan student, it wasn't that hard to just walk down the street and see who had thrown out bags of returnable bottles. It paid for more than a couple dinners. (I have pride, but I can't eat it.)