Of course. A 'system' consists of both hardware and software, and both have to work properly. Some systems even go a bit farther and try to compensate for 'wetware' failures -- that is, if the human who is controlling the system gives an order that the system decides must be a mistake, it will disregard it.
Fault tolerance is actually a much simpler concept than realtime, but just as hard to implement.:P
A fault tolerant system is a system in which the designers have anticipated some of the ways in which a system can fail, and designed the system to keep on functioning, at least at some level, in spite. An example of this would be redundant hardware -- if, for example, one power supply fails, the system has a backup power supply that will take over.
Fault tolerant systems are not necessarily realtime -- for example, a lot of servers have some degree of fault tolerance built in. Real time systems are not necessarily fault tolerant either, but in practice, since realtime systems are often used in places where a failure would be catastrophic (ie. the control system of an airplane) it would be a very wise idea to design some fault tolerance into them.
Since some people are asking what a realtime system/OS is, here goes:
A realtime system is a system in which the response time of the system to an external event is predictable and bounded. That is, if a user presses a button, a realtime system will guarantee that the system takes whatever action is associated with that button press in a given time, or less.
This does not necessarily mean that the system is fast or responsive. For example, the design spec could say that if the user presses the button, the system must respond in 5 seconds or less. As long as the system always responds within 5 seconds, it is still a realtime system, even though most people wouldn't think of it as such.
If the system fails to respond within the specified time frame, it is considered a design failure. As such, it is used mainly in things like industrial or vehicle control systems, where failing to respond within the specified time can have catastrophic consequences.
A realtime OS is simply an operating system designed for use in realtime systems. Because of strict constraints on response times, some harsh tradeoffs must be made -- ones that users of desktop and server operating systems would find hard to fathom.
For example, a realtime filesystem is very difficult, if not impossible, to design, because some filesystem operations can have a potentially unbounded completion time. Usually, any processor cache is disabled, because, although the cache greatly decreases the average execution time, it can actually _increase_ the worst-case execution time, and it makes it very difficult to determine how long a piece of code is going to take to execute.
The Heisenberg Uncertainty Principle basically puts an upper limit on the accuracy with which you can measure the position and the velocity of a given particle. So, you can measure the position very accurately, but then your velocity measurement goes all to hell, and vice versa. For a transporter-like device to function, it stands to reason that you would need to measure both very accurately.
But, who knows? We are talking about the future, after all, and there's a lot we don't know about physics, especially at the particle level. I'm certainly willing to suspend disbelief about pieces of future technology, even if modern science says that they shouldn't work. They are, after all, based on possible _future_ scientific knowledge, not modern.
What does get my goat, though, is the gross misunderstanding of the theory of evolution that the Star Trek writers show every time they write an episode having to do with evolution. (eg. the episode of TNG where everyone 'de-evolves', and the episode of Voyager where Janeway and Paris have their 'rate of evolution' increased, and evolve into lizard-things)
According to the theory of evolution, species evolve. Individuals don't. Darwin's book, after all, was called 'The Origin of Species'. Come on, guys, that should be a hint.
Ironically, I've even seen Creationists that seem to be able to grasp this concept. I wonder why the creators of Star Trek can't seem to.
Since these are both valid domain names, I suspect that even if the above scheme was ever postulated, it was treated with the neglect that it deserves.
Fifteen subjects, or 23%, reported that they play female characters because they are "more aesthetically pleasing" than male characters. That goes for real life too. The female form is a much more pleasing shape to look at than the male form. (Don't think I'm biased because I'm male, ladies. Anyone who's ever studied art knows this.)
Aside from the uncertainty involved in saying that you 'know' something related to aesthetics...
Anyway, I'm curious to know what the female perspective on this is. Which form do the ladies think is more aesthetically pleasing?
As a previous poster mentioned, there is almost certainly a cultural component to this preference. I'd be curious to know if there's a biological component as well.
It is therefore of the utmost importance to create a usable, clear window manager if Linux is ever to gain any significant acceptance among users who are not systems administrators. Its other strongpoints (stability, configurability) are of no significance in achieving this if a well-thought out interface isn't realized. Soon.
Well, it looks like you have your work cut out for you. Stop wasting time posting here, and get cracking at designing said window manager!
What I did complain about was that a lot of `theme-developers' seem to mistake excessive artiness for `fresh approach to designing user interfaces'. Clever image-manipulation does not make a good working environment. A good working environment consists of clear, logical metaphors and a layout that doesn't distract the user from the task he or she is undertaking.
Perhaps true. I know that I certainly don't like most of the overdone wm themes. But, obviously, some people do. They are created, and they are downloaded. If you don't like them, you don't have to care about them. Simple as that.
I would encourage you, however, to take a look at sawmill.themes.org. The sawmill users seem to be of a slightly different bent than enlightenment users, and there are about as many simple, minimalistic themes as there are gaudy, overdone themes.
The theming capabilities of window managers (and other programs) allow me to create the look and feel that _I_ want, rather than the look and feel that somebody else thinks I want. Why, pray tell, is this a bad thing?
The issue is identical to the free speech issue. If you have the freedom to say what you want, then it follows that everybody else does, too, even if they may say things that you don't like. If you don't like that, then there's always a number of totalitarian states around that you can emigrate to.
Most new interfaces (like Propagan(d?)a
Propaganda is an interface? Gosh, last time I looked, it was just a collection of abstract background tiles.
What is needed right now is for someone to come up with the one brilliant idea that will change the whole paradigm.
And I take it that you're volunteering for the task?
Dumping yet another set of overworked GIMP-manipulations on the web just isn't the way to go.
You seem to have this strange idea that Propaganda was intended to be a revolutionary concept in user-interface design. Are you perhaps taking the 'JFK' motif of the website a little too literally?
Propaganda is a collection of eye-candy backgrounds. If you like it, then use it. If you don't like it, then don't use it. Whining about it will only waste time that could be better spent designing that paradigm-changing user interface that you volunteered to create.:)
My question is: what does this have to do with anything? There is no DFT performed in the signal path of a consumer digital audio system. About the only place I can see a DFT being used in a piece of consumer audio equipment is to implement a fancy spectrum analyser on the front panel.
BTW, what exactly do you mean by 50% of the signal? It doesn't seem to have a lot of meaning.
Well yeah, that's assuming your DAC is doing a nice job of reconstructing the sine wave. If you sample any waveform at 120 samples/sec with a fundamental of 60Hz and that only has harmonic overtones (120,180, etc) the output will always be the same. As far as I know, most DAC's don't do this very well at their frequency limit, in which case the output would be more "connect-the-dots" and you'd have a triangle waveform with some slightly rounded-off tips.
Certainly true. I should have prefixed that with 'theoretically'. In practice, no device performs very well near the limits of its range.
Most audio pros work with the digital audio at 96k and then downsample right before the CD master. On the other hand, my SBLive insists on outputting 48khz, which makes that card useless for digital transfers to/from DAT that wind up on CD. I just do everything at 44.1k (on another soundcard) and I'm happy.
I think you misunderstood what I was saying. The oversampling I was talking about occurs in the A/D converter, and the end user never sees it. As it turns out, it's easier to sample at a rate so high that aliasing simply won't occur, then use a digital filter for antialiasing, then convert the sample rate down to the user requested sample rate than it is to create an analog filter with the sharp cutoff required for the antialiasing filter.
For example, to get that 96kHz sample rate, the A/D chip might sample at 250kHz or 500kHz, then use an FIR lowpass filter (which can be made with a very sharp cutoff, and, as a bonus, can be made so that they don't mess with the phase of the signal like analog filters do) for antialiasing. Then they might downsample the signal to 96kHz.
All this happens on the A/D converter chip. (Obviously, it has to have a DSP processor -- essentially a special-purpose CPU -- integrated if it's going to do this.) The downsample to 44.1kHz just before the CD mastering stage is an entirely separate step.
Actually, any recording method, be it analog or digital, will lose some of the information in the original signal, and distort some of the remaining information. I think the thing is that LP fans actually like the sound of the distortion introduced into the signal.
Considering that compressing a digital audio signal with an MP3 encoder will discard about 90% of the information in the original signal while only having a slight impact on sound quality, you can see that the minute amount of information lost in the quantization process shouldn't amount to a hill of beans.
OK, this is interesting. Basically, you've stated that the difference isn't in the recording technology itself, but rather in the pre- and post-recording stages which surround the recording itself.
With 16-bit audio, the best S/N ratio you can hope for is 96db, am I correct? I haven't seen the math behind this, I've just read this in several places as being the theoretical S/N ratio for 16-bit digital audio.
I'm not going to go through the derivation of this -- check out a DSP textbook if you want to know that -- but the formula for the SQNR (Signal to Quantization Noise Ratio) for a digital system is:
SQNR = 6.02B + 1.76 dB
where B is the number of bits per sample.
Therefore, the theoretical maximum SNR for a 16 bit digital system is 98.08 dB. For a 24 bit system, it's 146.24 dB, and for a 32 bit system, it's 194.4 dB.
There are actually upper limits on the SNR of an analog system, too, that result from the effects of thermal noise, but I don't know what they are, off-hand. Also, you never truly have infinite bandwidth.
The thing is, I find that comparing devices that can only exist in theory to be rather pointless.
As the matter of fact, it has become commonplace to oversample the heck out of a signal in the front end of the A/D converter (ie. sample at several times the Nyquist rate), do the anitaliasing filtering with a digital filter (which have much nicer characteristics than analog filters), then downsample the signal to 44.1kHz. This gets around some of the stickier problems associated with sampling.
Interestingly enough, with your 60Hz sine wave example, you got the situation backwards. If you sample a 60Hz triangle wave at 120Hz, playing it back will result in a 60Hz sine wave. The reason? A 60Hz sine wave consists of only a single harmonic -- that is, if you took the Fourier transform of it, you'd end up with a single spike at 60Hz. The 60Hz triangle wave, on the other hand, consists of a fundamental harmonic at 60Hz, and a diminishing series of harmonics at higher frequencies. Since all of these higher harmonics would be lost in the sampling process, you'd end up with only the fundamental harmonic -- a 60Hz sine wave.
However, those who are willing to put up with the extra work, and spend a lot of money for good equipment, end up with a sound you don't find on CD.
I beg to differ. I suspect that you only _think_ that you have better sound quality.
Do you remember back in the late 80's/early 90's when people swore that colouring the edge of a CD with a green marker improved the sound quality? I think the same thing is at work here.
The ones who told you it was perfect reconstruction were not scientists, they were corporate marketers attempting to replace the LP in popular media with the CD. Sure worked, didn't it? Even got many people believing the mathematically, provably wrong claims of no distortion.
Kindly refrain from constructing straw men. I don't recall anyone ever saying that there was no distortion in practical applications -- just that the distortion was much less than was possible with comparable analog systems. If you think that this isn't the case, then kindly provide some proof.
BTW, the claim of no distortion is provably correct, but only assuming a bunch of ideal conditions that never occur in the real world.
Did I say consumer-level? I was speaking on principle. Let's see... a near perfect analog system would consist of 1) somebody playing an acoustic instrument, and 2) a human ear. Better SNR than 16-bit quantization, and the only noise in the system is going to be due to bloodflow in the ear (ok, maybe dogs and traffic if you're not in a sound room). Anything else processing the sound in between, digital or analog, is just going to add noise.
Tell me, in the scenario you listed above, what would the point be of putting any digital component into the system? The main reasons for digitizing an audio signal are:
1) Storing it for later playback 2) Transmitting it greater distances than it will carry naturally 3) Applying special effects to it
If you're not going to do any of these things to the signal, a digital system is totally pointless anyway, so there's no point in comparing them.
Nyquist's theorem is useful for data transmission. If it is applied to audio, then there's really no reason that a classical music CD should sound worse than actually being there, even when played on the best hardware. But then we'd be getting into psychoacoustics.
Indeed.
There are plenty of other weak points in the channel, though. Microphones, amplifiers, speakers... These all add their own bits of distortion into the signal.
Correct me on this if I'm wrong (but do it nicely). If the original sound source has two signal components of 60kHz and 65kHz, there will be a tertiary tone of 5kHz as a result of the other two being superimposed. Is that 5kHz tone sampled successfully with a 44.1kHz sampling rate?
Consider the opposite question. 60-65kHz is outside the range of any consumer-quality analog recording equipment. (Maybe some professional quality equipment can do it, but then, you can trivially expand the range of a digital system by increasing the sampling rate and sample size, too.)
So, neither the analog nor the digital system will pick up the fundamental frequencies. That means, then, that either both the digital and the analog system will pick up the 5kHz beat signal, or neither of them will.
Comparing W2K to a Linux distribution isn't a good comparison, however. Linux distributions contain all kinds of application software that not even Microsoft would dare to call part of the OS.
On a related note, it does make me curious how many of those W2K bugs would disappear if they dropped some of their blatently application software *cough*IE*cough* from their distribution.
It should also be noted that with the current hardware situation, with CPU's being much faster than memory, it is vital that there be as few cache misses as possible. If there's a cache miss, the processor must re-read the data from main memory, which can stall the processor for many cycles, thus reducing performance.
If you overdo it and use too many threads, you obviously have to spend more time switching between them. Every time you switch threads, you invalidate the cache, causing cache misses left and right. As I stated above, this is a Bad Thing.
The day when RAM can keep pace with the CPU core, then it won't matter how many threads you use. But until that day, you need to be careful not to use too many threads.
CISC has become more like RISC. RISC has become more like CISC.
Really, the reason that these polarized viewpoints exist in the first place is because each of them have advantages. Isn't it only natural that they might be made even better by incorporating some of the advantages of the other viewpoint?
Slashdot Mirror
Sounds like a good enough excuse to me. :)
Of course. A 'system' consists of both hardware and software, and both have to work properly. Some systems even go a bit farther and try to compensate for 'wetware' failures -- that is, if the human who is controlling the system gives an order that the system decides must be a mistake, it will disregard it.
Fault tolerance is actually a much simpler concept than realtime, but just as hard to implement. :P
A fault tolerant system is a system in which the designers have anticipated some of the ways in which a system can fail, and designed the system to keep on functioning, at least at some level, in spite. An example of this would be redundant hardware -- if, for example, one power supply fails, the system has a backup power supply that will take over.
Fault tolerant systems are not necessarily realtime -- for example, a lot of servers have some degree of fault tolerance built in. Real time systems are not necessarily fault tolerant either, but in practice, since realtime systems are often used in places where a failure would be catastrophic (ie. the control system of an airplane) it would be a very wise idea to design some fault tolerance into them.
Since some people are asking what a realtime system/OS is, here goes:
A realtime system is a system in which the response time of the system to an external event is predictable and bounded. That is, if a user presses a button, a realtime system will guarantee that the system takes whatever action is associated with that button press in a given time, or less.
This does not necessarily mean that the system is fast or responsive. For example, the design spec could say that if the user presses the button, the system must respond in 5 seconds or less. As long as the system always responds within 5 seconds, it is still a realtime system, even though most people wouldn't think of it as such.
If the system fails to respond within the specified time frame, it is considered a design failure. As such, it is used mainly in things like industrial or vehicle control systems, where failing to respond within the specified time can have catastrophic consequences.
A realtime OS is simply an operating system designed for use in realtime systems. Because of strict constraints on response times, some harsh tradeoffs must be made -- ones that users of desktop and server operating systems would find hard to fathom.
For example, a realtime filesystem is very difficult, if not impossible, to design, because some filesystem operations can have a potentially unbounded completion time. Usually, any processor cache is disabled, because, although the cache greatly decreases the average execution time, it can actually _increase_ the worst-case execution time, and it makes it very difficult to determine how long a piece of code is going to take to execute.
The Heisenberg Uncertainty Principle basically puts an upper limit on the accuracy with which you can measure the position and the velocity of a given particle. So, you can measure the position very accurately, but then your velocity measurement goes all to hell, and vice versa. For a transporter-like device to function, it stands to reason that you would need to measure both very accurately.
But, who knows? We are talking about the future, after all, and there's a lot we don't know about physics, especially at the particle level. I'm certainly willing to suspend disbelief about pieces of future technology, even if modern science says that they shouldn't work. They are, after all, based on possible _future_ scientific knowledge, not modern.
What does get my goat, though, is the gross misunderstanding of the theory of evolution that the Star Trek writers show every time they write an episode having to do with evolution. (eg. the episode of TNG where everyone 'de-evolves', and the episode of Voyager where Janeway and Paris have their 'rate of evolution' increased, and evolve into lizard-things)
According to the theory of evolution, species evolve. Individuals don't. Darwin's book, after all, was called 'The Origin of Species'. Come on, guys, that should be a hint.
Ironically, I've even seen Creationists that seem to be able to grasp this concept. I wonder why the creators of Star Trek can't seem to.
http://www.guug.de/
http://yahoo.ca/
Since these are both valid domain names, I suspect that even if the above scheme was ever postulated, it was treated with the neglect that it deserves.
kirby.sucks
:)
hoover.sucks
filter-queen.sucks
and, of course:
electrolux.sucks
...because we all know, nothing sucks like an Electrolux!
On a different note, why not register really.sucks, then sell subdomains?
Yes but will it include support for large files (> 2 GB) on 32 bit machines?
Yes! The support has been there for a couple of months now.
Fifteen subjects, or 23%, reported that they play female characters because they are "more aesthetically pleasing" than male characters.
That goes for real life too. The female form is a much more pleasing shape to look at than the male form. (Don't think I'm biased because I'm male, ladies. Anyone who's ever studied art knows this.)
Aside from the uncertainty involved in saying that you 'know' something related to aesthetics...
Anyway, I'm curious to know what the female perspective on this is. Which form do the ladies think is more aesthetically pleasing?
As a previous poster mentioned, there is almost certainly a cultural component to this preference. I'd be curious to know if there's a biological component as well.
It is therefore of the utmost importance to create a usable, clear window manager if Linux is ever to gain any significant acceptance among users who are not systems administrators. Its other strongpoints (stability, configurability) are of no significance in achieving this if a well-thought out interface isn't realized. Soon.
Well, it looks like you have your work cut out for you. Stop wasting time posting here, and get cracking at designing said window manager!
What I did complain about was that a lot of `theme-developers' seem to mistake excessive artiness for `fresh approach to designing user interfaces'. Clever image-manipulation does not make a good working environment. A good working environment consists of clear, logical metaphors and a layout that doesn't distract the user from the task he or she is undertaking.
:)
Perhaps true. I know that I certainly don't like most of the overdone wm themes. But, obviously, some people do. They are created, and they are downloaded. If you don't like them, you don't have to care about them. Simple as that.
I would encourage you, however, to take a look at sawmill.themes.org. The sawmill users seem to be of a slightly different bent than enlightenment users, and there are about as many simple, minimalistic themes as there are gaudy, overdone themes.
The theming capabilities of window managers (and other programs) allow me to create the look and feel that _I_ want, rather than the look and feel that somebody else thinks I want. Why, pray tell, is this a bad thing?
The issue is identical to the free speech issue. If you have the freedom to say what you want, then it follows that everybody else does, too, even if they may say things that you don't like. If you don't like that, then there's always a number of totalitarian states around that you can emigrate to.
Most new interfaces (like Propagan(d?)a
Propaganda is an interface? Gosh, last time I looked, it was just a collection of abstract background tiles.
What is needed right now is for someone to come up with the one brilliant idea that will change the whole paradigm.
And I take it that you're volunteering for the task?
Dumping yet another set of overworked GIMP-manipulations on the web just isn't the way to go.
You seem to have this strange idea that Propaganda was intended to be a revolutionary concept in user-interface design. Are you perhaps taking the 'JFK' motif of the website a little too literally?
Propaganda is a collection of eye-candy backgrounds. If you like it, then use it. If you don't like it, then don't use it. Whining about it will only waste time that could be better spent designing that paradigm-changing user interface that you volunteered to create.
My question is: what does this have to do with anything? There is no DFT performed in the signal path of a consumer digital audio system. About the only place I can see a DFT being used in a piece of consumer audio equipment is to implement a fancy spectrum analyser on the front panel.
BTW, what exactly do you mean by 50% of the signal? It doesn't seem to have a lot of meaning.
Well yeah, that's assuming your DAC is doing a nice job of reconstructing the sine wave. If you sample any waveform at 120 samples/sec with a fundamental of 60Hz and that only has harmonic overtones (120,180, etc) the output will always be the same. As far as I know, most DAC's don't do this very well at their frequency limit, in which case the output would be more "connect-the-dots" and you'd have a triangle waveform with some slightly rounded-off tips.
Certainly true. I should have prefixed that with 'theoretically'. In practice, no device performs very well near the limits of its range.
Most audio pros work with the digital audio at 96k and then downsample right before the CD master. On the other hand, my SBLive insists on outputting 48khz, which makes that card useless for digital transfers to/from DAT that wind up on CD. I just do everything at 44.1k (on another soundcard) and I'm happy.
I think you misunderstood what I was saying. The oversampling I was talking about occurs in the A/D converter, and the end user never sees it. As it turns out, it's easier to sample at a rate so high that aliasing simply won't occur, then use a digital filter for antialiasing, then convert the sample rate down to the user requested sample rate than it is to create an analog filter with the sharp cutoff required for the antialiasing filter.
For example, to get that 96kHz sample rate, the A/D chip might sample at 250kHz or 500kHz, then use an FIR lowpass filter (which can be made with a very sharp cutoff, and, as a bonus, can be made so that they don't mess with the phase of the signal like analog filters do) for antialiasing. Then they might downsample the signal to 96kHz.
All this happens on the A/D converter chip. (Obviously, it has to have a DSP processor -- essentially a special-purpose CPU -- integrated if it's going to do this.) The downsample to 44.1kHz just before the CD mastering stage is an entirely separate step.
Actually, any recording method, be it analog or digital, will lose some of the information in the original signal, and distort some of the remaining information. I think the thing is that LP fans actually like the sound of the distortion introduced into the signal.
Considering that compressing a digital audio signal with an MP3 encoder will discard about 90% of the information in the original signal while only having a slight impact on sound quality, you can see that the minute amount of information lost in the quantization process shouldn't amount to a hill of beans.
OK, this is interesting. Basically, you've stated that the difference isn't in the recording technology itself, but rather in the pre- and post-recording stages which surround the recording itself.
With 16-bit audio, the best S/N ratio you can hope for is 96db, am I correct? I haven't seen the math behind this, I've just read this in several places as being the theoretical S/N ratio for 16-bit digital audio.
I'm not going to go through the derivation of this -- check out a DSP textbook if you want to know that -- but the formula for the SQNR (Signal to Quantization Noise Ratio) for a digital system is:
SQNR = 6.02B + 1.76 dB
where B is the number of bits per sample.
Therefore, the theoretical maximum SNR for a 16 bit digital system is 98.08 dB. For a 24 bit system, it's 146.24 dB, and for a 32 bit system, it's 194.4 dB.
There are actually upper limits on the SNR of an analog system, too, that result from the effects of thermal noise, but I don't know what they are, off-hand. Also, you never truly have infinite bandwidth.
The thing is, I find that comparing devices that can only exist in theory to be rather pointless.
As the matter of fact, it has become commonplace to oversample the heck out of a signal in the front end of the A/D converter (ie. sample at several times the Nyquist rate), do the anitaliasing filtering with a digital filter (which have much nicer characteristics than analog filters), then downsample the signal to 44.1kHz. This gets around some of the stickier problems associated with sampling.
Interestingly enough, with your 60Hz sine wave example, you got the situation backwards. If you sample a 60Hz triangle wave at 120Hz, playing it back will result in a 60Hz sine wave. The reason? A 60Hz sine wave consists of only a single harmonic -- that is, if you took the Fourier transform of it, you'd end up with a single spike at 60Hz. The 60Hz triangle wave, on the other hand, consists of a fundamental harmonic at 60Hz, and a diminishing series of harmonics at higher frequencies. Since all of these higher harmonics would be lost in the sampling process, you'd end up with only the fundamental harmonic -- a 60Hz sine wave.
Neat, huh?
However, those who are willing to put up with the extra work, and spend a lot of money for good equipment, end up with a sound you don't find on CD.
I beg to differ. I suspect that you only _think_ that you have better sound quality.
Do you remember back in the late 80's/early 90's when people swore that colouring the edge of a CD with a green marker improved the sound quality? I think the same thing is at work here.
The ones who told you it was perfect reconstruction were not scientists, they were corporate marketers attempting to replace the LP in popular media with the CD. Sure worked, didn't it? Even got many people believing the mathematically, provably wrong claims of no distortion.
Kindly refrain from constructing straw men. I don't recall anyone ever saying that there was no distortion in practical applications -- just that the distortion was much less than was possible with comparable analog systems. If you think that this isn't the case, then kindly provide some proof.
BTW, the claim of no distortion is provably correct, but only assuming a bunch of ideal conditions that never occur in the real world.
Did I say consumer-level? I was speaking on principle. Let's see... a near perfect analog system would consist of 1) somebody playing an acoustic instrument, and 2) a human ear. Better SNR than 16-bit quantization, and the only noise in the system is going to be due to bloodflow in the ear (ok, maybe dogs and traffic if you're not in a sound room). Anything else processing the sound in between, digital or analog, is just going to add noise.
Tell me, in the scenario you listed above, what would the point be of putting any digital component into the system? The main reasons for digitizing an audio signal are:
1) Storing it for later playback
2) Transmitting it greater distances than it will carry naturally
3) Applying special effects to it
If you're not going to do any of these things to the signal, a digital system is totally pointless anyway, so there's no point in comparing them.
Nyquist's theorem is useful for data transmission. If it is applied to audio, then there's really no reason that a classical music CD should sound worse than actually being there, even when played on the best hardware. But then we'd be getting into psychoacoustics.
Indeed.
There are plenty of other weak points in the channel, though. Microphones, amplifiers, speakers... These all add their own bits of distortion into the signal.
Correct me on this if I'm wrong (but do it nicely). If the original sound source has two signal components of 60kHz and 65kHz, there will be a tertiary tone of 5kHz as a result of the other two being superimposed. Is that 5kHz tone sampled successfully with a 44.1kHz sampling rate?
Consider the opposite question. 60-65kHz is outside the range of any consumer-quality analog recording equipment. (Maybe some professional quality equipment can do it, but then, you can trivially expand the range of a digital system by increasing the sampling rate and sample size, too.)
So, neither the analog nor the digital system will pick up the fundamental frequencies. That means, then, that either both the digital and the analog system will pick up the 5kHz beat signal, or neither of them will.
Comparing W2K to a Linux distribution isn't a good comparison, however. Linux distributions contain all kinds of application software that not even Microsoft would dare to call part of the OS.
On a related note, it does make me curious how many of those W2K bugs would disappear if they dropped some of their blatently application software *cough*IE*cough* from their distribution.
Well said.
It should also be noted that with the current hardware situation, with CPU's being much faster than memory, it is vital that there be as few cache misses as possible. If there's a cache miss, the processor must re-read the data from main memory, which can stall the processor for many cycles, thus reducing performance.
If you overdo it and use too many threads, you obviously have to spend more time switching between them. Every time you switch threads, you invalidate the cache, causing cache misses left and right. As I stated above, this is a Bad Thing.
The day when RAM can keep pace with the CPU core, then it won't matter how many threads you use. But until that day, you need to be careful not to use too many threads.
So, you're dead, then? Nice to know that you can still use computers after you've died. :D
Not really.
Consider another holy war: RISC vs. CISC
What has happened in the meantime?
CISC has become more like RISC. RISC has become more like CISC.
Really, the reason that these polarized viewpoints exist in the first place is because each of them have advantages. Isn't it only natural that they might be made even better by incorporating some of the advantages of the other viewpoint?
You've failed to convince me that there's any connection between these two statements.
As the original author suggested, the connection to the faded programming fads of the past seems much stronger.