If I throw it in the air and let it float down, what is the probability that, if it lands, it will land with that surface touching the ground?
I think I took care of that by making it a condition of the event occurring. If it doesn't land, the event has not occurred, and there is no probabilistic trial.
A real Möbius strip might land on edge.
Hence my statement that it is a ideal mobius strip, one with no edge. Thought experiments with unobtainium are just fine, thank you.
Besides which, you still can't fuck with p(1+1=2 for arithmetic over the set of real numbers) = 1.0.
I suspect that Joe Average on the street has no clue what SETI@Home is. And SETI@home isn't really "science"... I admire the project, and run their code, but there's a difference between search and research.
Couple reasons - once upon a time (some of it was taken out around STS-109, I believe) it still had extra instrumentation from its time as the shakedown vehicle. Also, there were certain modifications made to reduce weight over time, as the newer shuttles were built.
Let's say I have a mobius strip. A true mobius strip, no edges, just one connected surface.
If I throw it in the air and let it float down, what is the probability that, if it lands, it will land with that surface touching the ground?
p = 1.0.
p(1+1=2 for the set of real numbers)? 1.0
Your math sucks, by the way, or possibly you just can't read. He doesn't assume p=1 for anything in the coin example: if you can make 4 flips per minute, and you need a p(.5^4)=0.0625 event, it will on average take you eight trials (8 minutes, in this case). If you have 16 people doing this, on *average* at least one person will flip HHHH in the first minute. Where the fuck does a probability of one come in?
Nonetheless, AO can't push things beyond diffraction limits. If you want to improve your resolution past diffraction, you're going to have to increase the size of the scope.
Also, do you have a reference for your claim that Spitzer uses AO? I couldn't find a single mention of that fact, anywhere. In fact, I can't find a mention of *any* spaceborne imaging system that uses AO.
Why is it Okay for one group of people (the State) to capture and punish criminals, but not Okay for another group (the People) to do so?
It's called the social contract. Read up on your Enlightenment philosophy and you'll understand the answer to this question, but the simple form is:
Government (the State) is the result of an implied contract between the State and the People. In this contract, the people cede certain freedoms to the government in exchange for certain services performed by the government. For example, we cede our freedom to beat our neighbor bloody *for almost any reason*. In exchange, we receive a promise that those who break these aspects of the contract will be punished by the State. There is nothing inherently wrong with vigilante justice; HOWEVER, by participating in our state, you have agreed that you will not act in this fashion, and that if you do you are subject to the terms of the contract - to whit, prosecution for your actions.
By the way - the demands of fiction are not a good basis for reality. All of Shakespeare's tragic heros *die* before their time. Does that mean that you have to die before your time to be a hero?
Game theory has a very, very specific meaning which only rarely has anything to do with game design. Please use "game design" to refer to what you do, so Borel, von Neumann, Morgenstern, and the other game theory types can stop rolling over in their graves and/or beds.
Your pointless statement that "I don't think I need to read a book about it!"?
While your games may or may not be fun (I may or may not check them out), don't try to claim you weren't self-promoting. The only kind of prostitution anyone should have a problem with is the kind that isn't honest about it.
I was going to point that out in my comment, actually. This isn't astroturfing; astroturfing would be things like orchestrated letter-to-the-editor campaigns where the letters are sent by a small group purporting to represent many people (Parent's Television Council, anyone?), small communities of users swarming generally open polls to bias the results to their viewpoint, etc, etc.
There is a term for this obfuscated funding: it's called "buying your evidence".
Actually, old to this. I haven't touched a wireless mic in probably 8 or 9 years, because I've been too busy working as a signal processing and control engineer, which explains why I figured the wireless mic makers had gotten around to moving over to digital modulation methods by now like the rest of the world has.
2. Exactly. No one builds wireless to be as good as wired, because it would be incredibly expensive and have no market. Vicious circle - the available wireless is not good enough for studio use because there's no point in building wireless for the studio because studios won't use wireless for quality and expense reasons.
3. Wrong. Read any introductory text on information theory for why, but digital is inherently lossy, since you need an infinite number of bits to perfectly represent an analog signal. Whether you *care* about that lossiness is different, but digital can never be as good as analog.
4. 1914 is a long time ago. Get a recent book. Also, remember that there are multiple standard definitions, ranging from the communication/information theoretic "Width of the usable transmission channel" to the networking "Capacity of the communication channel" to the EE half-power bandwidth to many other definitions. Bandwidth simply means the frequency width of a certain channel between two endpoints; those endpoints may be defined in different ways. Commonly, they used half-power (-3dB) because it was the point at which the effective power output of the system was cut in half. For example, I can quote Franklin/Powell "Digital Control of Dynamic Systems", 1994, as saying "Bandwidth is the maximum frequency at which the output of a system will track a reference sinusoid in a satisfactory manner." That makes me no more right than you - although bandwidth is conventionally defined as -3dB, it is not inherently defined as -3dB.
5. Stop quoting a 90 year old book and assuming nothing has changed in the intervening time period. If somebody doesn't specify their bandwidth condition, they aren't reliable enough to trust their numbers.
Wireless microphone elements tend to be lower quality than wired microphone elements. This is nothing to do with wired vs. wireless, just the elements, but remains a correct statement, such that wired mics tend to be higher quality than wireless. Digital ALWAYS sounds worse than analog in the ideal case, again, by definition, as digital is inherently lossy. Properly designed digital may sound better than bad analog, and good digital *can* be indistinguishable from good analog. The point remains: studios use wired mics because they sound better than wireless. We can argue all we want about why, but the statement remains true.
Next, there's no such thing as a standard definition of bandwidth being -3dB. Bandwidth can also refer to the alias-free bandwidth (0 to 0.5sampling), the information capacity of a system, the frequency range of a communication channel, etc. You've been fooled just as badly as you think I have. The definition you're using is called the half-power bandwidth, and is valid for systems of any order. It's an extremely useful, extremely common definition, but by no means the only one, and for audio is less useful than you might think, as that "half power" means exactly what it says.
3dB points are a defined location. 6dB octave is not a DEFINITION, it is a mathematical truth. A first order system will roll off at 6 dB/octave because mathematically, that is what it has to do.
Half-power bandwidth of higher order systems is, indeed, defined as between the 3dB points. That was not my point. My point can be demonstrated by an example. I think we'd both agree that for audio, a flat frequency response is ideal. Look at a very high Q 2nd order system, then look at a low Q 2nd order system. Unlike first order systems, where the approach is specified completely by the order, in higher order systems you can have issues like resonance that do change how the approach looks. Hell, look at a first order system vs. a high order system with the same 3 dB points. The high order system will have a more ideal response because of their faster rolloff.
When a manufacturer uses the term bandwidth, don't put any heed into it unless they also explain *exactly* what their definition is. I'd much rather see a manufacturer who defines +/- because they've also stated that they don't have any local peaking greater than X dB within the specified band, not just that the first points where they roll off are at the band edges.
Better yet, ignore the numeric bandwidth and get a plot of the frequency response.
Learn just enough to figure out how to wrap it in your favored language and call the wrapped code, so you can use the aforementioned LAPACK/BLASPACK/whatever scientific library you need.
Recording studios don't use wireless because they sound worse. That was the point. The reasons they sound worse range from lower quality elements to, as you mentioned, additional processing stages.
Second, most electronics in audio gear are capable of more than the audio range for exactly the reason you quote - if you're looking for flat response across the audio range, your 3dB points need to be outside that range. However, 3dB is a damn meaningless statistic without knowing the order of a system, and for high-order systems (which real audio systems are, even if they resemble low order systems to a certain approximation) is more and more meaningless.
Third, the amplifier is the last worry when it comes to getting an accurate system transfer function. Its easy enough to get a +/- 0.1 dB response from 1 Hz up to 100 kHz in an amplifier, and in most any linear electronic stage; the response is dominated by speaker response and intentional nonlinearities induced by things like EQ.
(6dB/octave isn't a standard; its just the response of a 1st order system. Please never refer to it as standard again, especially since nearly all audio systems are at minimum effectively 2nd order, 12dB/octave.)
I don't think only the upper crust would received schooling (wasn't me said that, boss), but I'd be willing to bet most of what's traditionally considered "lower class" would not.
And there are a *lot* of lower class people out there.
No... see, all you've done in that case is changed the *analog* representation of a digital 1 or 0. The amplification is analog; the digital signal is unchanged, a sequency of zeros and ones.
Digital *is* a special case of analog. Any digital designer that tells you otherwise deserves to be kicked in the teeth.
That would be a class D amplifier, something I mentioned, using the speaker coil as the lowpass filter to take the high-frequency components of the switching signal and smooth them out into the desired waveform.
"Standard" Class D amps do the same thing, with an internal lowpass filter, rather than using the coil as a filter (which is generally a bad idea, since the speaker coil isn't going to be optimized to the characteristics of the switches).
And no, DACs do not generally convert a PWM signal into an analog signal. They generally take a PCM signal as input, and output an analog voltage. Some of them internally use a PWM signal in the conversion, but most won't accept one as input (I don't recall ever seeing any).
And actually, the better the speaker, often the worse that idea, as many speakers will have enough response to try to follow the high frequency switching, which is desirable for neither good sound nor the speakers' health.
It's okay; from experience, 99% of the staff hate U of M sports just as much as you do.:)
(I, on the other hand, would like your entire football team to choke and die... just like mine did at the Horseshoe. See ya in the Big House this year!)
I understand your points, certainly, and I'd be the last to advocate government-mandated computers.
I live in a large, large city, and grew up in a medium sized one, which might influence how I see these things, but its rare to see the sort of situation you mention. More common was seeing people on the corner sparing for change, sleeping in alleys, so forth. The thing is, instead of relying on people's good will, welfare mandates it. I don't have enough faith in people to expect them to take care of everyone - I spent time in the South with people who would welcome a troubled white family into their home, but wouldn't dream of helping out a black family with a damn thing, and I've seen inner city black families that go the other way. Charity just isn't suitable to provide the last-ditch net.
From whom will they be buying the computers? I'm assuming the government isn't going to give away G5 Power Macs. I'm figuring that, much like most social service programs, they'll be getting the bare bones - a P2-266 or some such, the sort of thing we regularly see posted on Slashdot as appropriate for 3rd world countries to receive. But enough to get a large portion of the real benefits of owning a computer, while not anything even close to the top of the line. Just like guv'mint cheese will never be mistaken for a fine brie.
So who would they buy computers from? I'm going to assume that, even if the computers available were more expensive, better quality computers would remain available, and their quality probably wouldn't drop much. Would $300 Walmart machines disappear? $500 Dell specials? Yeah, probably. But I doubt the people who currently buy Alienware boxes, who currently buy Apples and even high-end white boxes, would be satisfied with guv'mint silicon. I don't think there's any real evidence for your assertion that a government-mandated, government-provided, security net backup for some assumed "need" eliminates more expensive capitalist choices. Further, if the government can't build those computers cheaper than private industry, they can (and should!) allow private industry to build them, buy them from industry, and give them out as a baseline. Similar to how we handle defense - the US government doesn't build the hardware, we train and employ the people who use it. If computers were baseline, buying the computers and running the distribution and training infrastructure would be an appropriate way to run the program.
I see no reason to believe that a free market approach would successfully achieve *comprehensive* education. Being as we have, as a country, decided that everyone deserves a basic education, the free market approach to it seems designed to fail our desires.
Slashdot Mirror
Objection #1: " It might never land."
If I throw it in the air and let it float down, what is the probability that, if it lands, it will land with that surface touching the ground?
I think I took care of that by making it a condition of the event occurring. If it doesn't land, the event has not occurred, and there is no probabilistic trial.
A real Möbius strip might land on edge.
Hence my statement that it is a ideal mobius strip, one with no edge. Thought experiments with unobtainium are just fine, thank you.
Besides which, you still can't fuck with p(1+1=2 for arithmetic over the set of real numbers) = 1.0.
Didn't think so, but I felt he deserved a chance to back up his bullshit.
I suspect that Joe Average on the street has no clue what SETI@Home is. And SETI@home isn't really "science"... I admire the project, and run their code, but there's a difference between search and research.
Couple reasons - once upon a time (some of it was taken out around STS-109, I believe) it still had extra instrumentation from its time as the shakedown vehicle. Also, there were certain modifications made to reduce weight over time, as the newer shuttles were built.
Let's say I have a mobius strip. A true mobius strip, no edges, just one connected surface.
If I throw it in the air and let it float down, what is the probability that, if it lands, it will land with that surface touching the ground?
p = 1.0.
p(1+1=2 for the set of real numbers)? 1.0
Your math sucks, by the way, or possibly you just can't read. He doesn't assume p=1 for anything in the coin example: if you can make 4 flips per minute, and you need a p(.5^4)=0.0625 event, it will on average take you eight trials (8 minutes, in this case). If you have 16 people doing this, on *average* at least one person will flip HHHH in the first minute. Where the fuck does a probability of one come in?
Take your own advice.
Nonetheless, AO can't push things beyond diffraction limits. If you want to improve your resolution past diffraction, you're going to have to increase the size of the scope.
Also, do you have a reference for your claim that Spitzer uses AO? I couldn't find a single mention of that fact, anywhere. In fact, I can't find a mention of *any* spaceborne imaging system that uses AO.
Why is it Okay for one group of people (the State) to capture and punish criminals, but not Okay for another group (the People) to do so?
It's called the social contract. Read up on your Enlightenment philosophy and you'll understand the answer to this question, but the simple form is:
Government (the State) is the result of an implied contract between the State and the People. In this contract, the people cede certain freedoms to the government in exchange for certain services performed by the government. For example, we cede our freedom to beat our neighbor bloody *for almost any reason*. In exchange, we receive a promise that those who break these aspects of the contract will be punished by the State. There is nothing inherently wrong with vigilante justice; HOWEVER, by participating in our state, you have agreed that you will not act in this fashion, and that if you do you are subject to the terms of the contract - to whit, prosecution for your actions.
By the way - the demands of fiction are not a good basis for reality. All of Shakespeare's tragic heros *die* before their time. Does that mean that you have to die before your time to be a hero?
Game theory has a very, very specific meaning which only rarely has anything to do with game design. Please use "game design" to refer to what you do, so Borel, von Neumann, Morgenstern, and the other game theory types can stop rolling over in their graves and/or beds.
Provide credibility for what?
Your pointless statement that "I don't think I need to read a book about it!"?
While your games may or may not be fun (I may or may not check them out), don't try to claim you weren't self-promoting. The only kind of prostitution anyone should have a problem with is the kind that isn't honest about it.
More like "Linux would be a car so unfriendly you'd need to be a mechanic to buy it, much less drive it."
I was going to point that out in my comment, actually. This isn't astroturfing; astroturfing would be things like orchestrated letter-to-the-editor campaigns where the letters are sent by a small group purporting to represent many people (Parent's Television Council, anyone?), small communities of users swarming generally open polls to bias the results to their viewpoint, etc, etc.
There is a term for this obfuscated funding: it's called "buying your evidence".
Good lord.
He *announced* it was a troll post, and you still got suckered?
Go hide your face in shame.
4'33".
Anytime someone wants to argue about whether computer-generated procedural art is actually art, I point them to that.
(Alvin Lucier's "I Am Sitting In A Room" is also one of the finest pieces of process composition ever.)
Computers are just the tool used by the programmer to create procedural art; absolutely, it can be art.
Astroturf, in the literal meaning, is fake grass.
Astroturfing in the political sense is fake "grassroots" political activity.
Actually, old to this. I haven't touched a wireless mic in probably 8 or 9 years, because I've been too busy working as a signal processing and control engineer, which explains why I figured the wireless mic makers had gotten around to moving over to digital modulation methods by now like the rest of the world has.
2. Exactly. No one builds wireless to be as good as wired, because it would be incredibly expensive and have no market. Vicious circle - the available wireless is not good enough for studio use because there's no point in building wireless for the studio because studios won't use wireless for quality and expense reasons.
3. Wrong. Read any introductory text on information theory for why, but digital is inherently lossy, since you need an infinite number of bits to perfectly represent an analog signal. Whether you *care* about that lossiness is different, but digital can never be as good as analog.
4. 1914 is a long time ago. Get a recent book. Also, remember that there are multiple standard definitions, ranging from the communication/information theoretic "Width of the usable transmission channel" to the networking "Capacity of the communication channel" to the EE half-power bandwidth to many other definitions. Bandwidth simply means the frequency width of a certain channel between two endpoints; those endpoints may be defined in different ways. Commonly, they used half-power (-3dB) because it was the point at which the effective power output of the system was cut in half. For example, I can quote Franklin/Powell "Digital Control of Dynamic Systems", 1994, as saying "Bandwidth is the maximum frequency at which the output of a system will track a reference sinusoid in a satisfactory manner." That makes me no more right than you - although bandwidth is conventionally defined as -3dB, it is not inherently defined as -3dB.
5. Stop quoting a 90 year old book and assuming nothing has changed in the intervening time period. If somebody doesn't specify their bandwidth condition, they aren't reliable enough to trust their numbers.
Wireless microphone elements tend to be lower quality than wired microphone elements. This is nothing to do with wired vs. wireless, just the elements, but remains a correct statement, such that wired mics tend to be higher quality than wireless. Digital ALWAYS sounds worse than analog in the ideal case, again, by definition, as digital is inherently lossy. Properly designed digital may sound better than bad analog, and good digital *can* be indistinguishable from good analog. The point remains: studios use wired mics because they sound better than wireless. We can argue all we want about why, but the statement remains true.
Next, there's no such thing as a standard definition of bandwidth being -3dB. Bandwidth can also refer to the alias-free bandwidth (0 to 0.5sampling), the information capacity of a system, the frequency range of a communication channel, etc. You've been fooled just as badly as you think I have. The definition you're using is called the half-power bandwidth, and is valid for systems of any order. It's an extremely useful, extremely common definition, but by no means the only one, and for audio is less useful than you might think, as that "half power" means exactly what it says.
3dB points are a defined location. 6dB octave is not a DEFINITION, it is a mathematical truth. A first order system will roll off at 6 dB/octave because mathematically, that is what it has to do.
Half-power bandwidth of higher order systems is, indeed, defined as between the 3dB points. That was not my point. My point can be demonstrated by an example. I think we'd both agree that for audio, a flat frequency response is ideal. Look at a very high Q 2nd order system, then look at a low Q 2nd order system. Unlike first order systems, where the approach is specified completely by the order, in higher order systems you can have issues like resonance that do change how the approach looks. Hell, look at a first order system vs. a high order system with the same 3 dB points. The high order system will have a more ideal response because of their faster rolloff.
When a manufacturer uses the term bandwidth, don't put any heed into it unless they also explain *exactly* what their definition is. I'd much rather see a manufacturer who defines +/- because they've also stated that they don't have any local peaking greater than X dB within the specified band, not just that the first points where they roll off are at the band edges.
Better yet, ignore the numeric bandwidth and get a plot of the frequency response.
Learn just enough to figure out how to wrap it in your favored language and call the wrapped code, so you can use the aforementioned LAPACK/BLASPACK/whatever scientific library you need.
Wrong.
It has wider coverage because OnStar is AMPS-based, and the AMPS analog network is more widespread than the CDMA/GSM digital networks are.
However, OnStar will become CDMA based over the next couple years, wiping out that AMPS advantage.
(And yes, I know this for a fact, considering that the PCB layouts for OnStar MY06 are printed out in a folder next to my desk.)
Recording studios don't use wireless because they sound worse. That was the point. The reasons they sound worse range from lower quality elements to, as you mentioned, additional processing stages.
Second, most electronics in audio gear are capable of more than the audio range for exactly the reason you quote - if you're looking for flat response across the audio range, your 3dB points need to be outside that range. However, 3dB is a damn meaningless statistic without knowing the order of a system, and for high-order systems (which real audio systems are, even if they resemble low order systems to a certain approximation) is more and more meaningless.
Third, the amplifier is the last worry when it comes to getting an accurate system transfer function. Its easy enough to get a +/- 0.1 dB response from 1 Hz up to 100 kHz in an amplifier, and in most any linear electronic stage; the response is dominated by speaker response and intentional nonlinearities induced by things like EQ.
(6dB/octave isn't a standard; its just the response of a 1st order system. Please never refer to it as standard again, especially since nearly all audio systems are at minimum effectively 2nd order, 12dB/octave.)
I don't think only the upper crust would received schooling (wasn't me said that, boss), but I'd be willing to bet most of what's traditionally considered "lower class" would not.
And there are a *lot* of lower class people out there.
No... see, all you've done in that case is changed the *analog* representation of a digital 1 or 0. The amplification is analog; the digital signal is unchanged, a sequency of zeros and ones.
Digital *is* a special case of analog. Any digital designer that tells you otherwise deserves to be kicked in the teeth.
That would be a class D amplifier, something I mentioned, using the speaker coil as the lowpass filter to take the high-frequency components of the switching signal and smooth them out into the desired waveform.
"Standard" Class D amps do the same thing, with an internal lowpass filter, rather than using the coil as a filter (which is generally a bad idea, since the speaker coil isn't going to be optimized to the characteristics of the switches).
And no, DACs do not generally convert a PWM signal into an analog signal. They generally take a PCM signal as input, and output an analog voltage. Some of them internally use a PWM signal in the conversion, but most won't accept one as input (I don't recall ever seeing any).
And actually, the better the speaker, often the worse that idea, as many speakers will have enough response to try to follow the high frequency switching, which is desirable for neither good sound nor the speakers' health.
It's okay; from experience, 99% of the staff hate U of M sports just as much as you do. :)
(I, on the other hand, would like your entire football team to choke and die... just like mine did at the Horseshoe. See ya in the Big House this year!)
I understand your points, certainly, and I'd be the last to advocate government-mandated computers.
I live in a large, large city, and grew up in a medium sized one, which might influence how I see these things, but its rare to see the sort of situation you mention. More common was seeing people on the corner sparing for change, sleeping in alleys, so forth. The thing is, instead of relying on people's good will, welfare mandates it. I don't have enough faith in people to expect them to take care of everyone - I spent time in the South with people who would welcome a troubled white family into their home, but wouldn't dream of helping out a black family with a damn thing, and I've seen inner city black families that go the other way. Charity just isn't suitable to provide the last-ditch net.
From whom will they be buying the computers? I'm assuming the government isn't going to give away G5 Power Macs. I'm figuring that, much like most social service programs, they'll be getting the bare bones - a P2-266 or some such, the sort of thing we regularly see posted on Slashdot as appropriate for 3rd world countries to receive. But enough to get a large portion of the real benefits of owning a computer, while not anything even close to the top of the line. Just like guv'mint cheese will never be mistaken for a fine brie.
So who would they buy computers from? I'm going to assume that, even if the computers available were more expensive, better quality computers would remain available, and their quality probably wouldn't drop much. Would $300 Walmart machines disappear? $500 Dell specials? Yeah, probably. But I doubt the people who currently buy Alienware boxes, who currently buy Apples and even high-end white boxes, would be satisfied with guv'mint silicon. I don't think there's any real evidence for your assertion that a government-mandated, government-provided, security net backup for some assumed "need" eliminates more expensive capitalist choices. Further, if the government can't build those computers cheaper than private industry, they can (and should!) allow private industry to build them, buy them from industry, and give them out as a baseline. Similar to how we handle defense - the US government doesn't build the hardware, we train and employ the people who use it. If computers were baseline, buying the computers and running the distribution and training infrastructure would be an appropriate way to run the program.
I see no reason to believe that a free market approach would successfully achieve *comprehensive* education. Being as we have, as a country, decided that everyone deserves a basic education, the free market approach to it seems designed to fail our desires.
Pros:
4. Annoys all roommates simultaneously!
Cons:
4. Annoys all roommates simultaneously!