I sidestepped it because I don't consider it relavent. The arguments of the pro-lifers aren't couched in law, they are couched in morality. The law says if you let someone die right in front of you, when you could have saved their life, you aren't in the wrong. Morality says otherwise.
That's a contrived example. Given the science behind what we're talking about, we're not talking about sacrificing your own fetus, but somebody else's. That little bit makes all the differences.
Some would feign horror at the thought of "sacrificing" someone else's life to better your own, but that would be naive. We do this on a daily basis, just to maintain our comfortable life. We, the citizens of the United States could easily give up much of our standard of living to save millions of lives around the world. We do not, and should not, because we are only human, and by nature self-interested. The bare truth is that you've already sacrificed human life just to live as comfortably as you do. We sacrifice human life for many reasons, and the presevation of those we love is a more justifiable one that most.
Since when do you care about the death of another child? Since when does anyone? I ordered a PowerMac the other day. I didn't really need one, but they looked cool. I could have taken that money and given it to a starving kid in Bangladesh (my native country). The average Bengali works 5 years to accumulate that much money, and you can bet it would have saved at least one life. With my lack of concern, I basically allowed somebody to die. At a logical level, it is no different than if I had let someone get hit by a bus, without trying to warn them or push them out of the way. It might not have the same emotional impact, but at an abstract level, it is no different. This is a poignant, yet somehow painless truth.
What am I getting at? That we're all evil for letting children starve? No! We cannot live our lives shackled to the destinies of others. We cannot torture ourselves for the good we could do but do not. We are only human, we're not built to care for other people in that way. You claim to care about human life, but would you sacrifice your comfort to preserve it? Would you sacrifice a nice home for your family, a nice school for your children, your gas-guzzling, environment-polluting car, or even your trifling conveniences and luxuries? Would you sacrifice any of these things? There are some people who do, and while we admire such people at an intellectual level, almost none of us are willing to follow their example.
It is for this reason that I find the "pro life" argument disturbing. Here are people who allegedly care about life, yet, they spend an enormous amount of effort (and a not-insignificant amount of money), fighting for "abstract people", while letting "concrete people" die every day. I can't see how anybody can rationalize that. It makes no sense at a concrete level, or at an abstract level.
Loans are rather different from grants, as anybody with student loans will tell you. The claim that higher education is a racket is moronic. Its the superiority of American higher-education that keeps us the dominant technological power. Yes, it costs the government money, but that money pays huge dividends.
Now, you might think the IT field is not subsidized, but consider its origins. A very large fraction of IT technology was developed under government grants. The internet is the classical example, but even if you consider stuff that was developed at commercial research labs, you have to realize a lot of those developments were at least partially the result of government contracts. In the end, an engineer complaining about government subsidies is like a farmer complaining about government subsidies!
The right doesn't want to admit it was wrong about global warming and the left doesn't want to admit it was wrong about nukes.
So true! The right doesn't believe there is a problem, and the left thinks everything is a problem and won't allow any of the solutions. Everybody is waiting for a "perfect" solution, and even if there is one of these, we're not going to find one in the next century!
You do realize that engineering, as a profession, is more subsidized than paleontology? Paleontologists are usually supported by universities, which are at least partially supported by tuition. Meanwhile, entire branches of the engineering profession (eg: aerospace), exist either due to the DoD budget (for military stuff), or due to airline bailouts (for civilian stuff).
japanese anime also seems to do just fine in the US, fans seem to prefer the original japanese dialogue with subtitles.
First, Japanese anime is quite a niche. Second, its heavily edited in its American versions, and only the niche of the niche bother to see the original.
I don't see how your comment relates to mine? Where did Intel come into play? My point was that a dual-processor Athlon 64 system has an advantage over a dual-core Athlon 64 system because of the dual memory busses. On processors like the G5 and P4, which use a shared memory bus, dual-processor systems don't really have an advantage over dual-core ones.
This is the classical sort of thinking that perpetuates poverty. Africa is a continent of over 1 billion people. If they spent all their effort trying to provide people with food, water, and shelter, they'd never get anywhere. Its unsustainable development, and something that the international development community is quite aware of. The "give a man a fish" saying seems trite, but it really does fit. China is an excellent example. Right now, there is actually hope for the people of China that in the future they might not be as poor as they are now. If China had exerted all its resources trying to take care of its billion people, instead of building up their manufacturing and financial capacity, there would not be any such hope.
In terms of the Athlon dual-core, I'm not positive on this, but I think it's integrated memory controller maxes throughput at 6.4GBps, and the I/O at 8GBps.
Yep. The dual-core G5s have an edge on balanced read-write bandwidth, while the Athlon has an edge on total bandwidth.
The Athlon has a 128-bit memory bus and a 16-bit system bus, both of which are unidirectional (either in or out, not simultaneous).
The Hypertransport bus is 2 16-bit unidirectional links running at 2 GHz (hence the 8GB/sec of I/O bandwidth).
The Apple bus architecture eliminates the latency associated with flipping the direction of the bus as well as the latency associated with the processor and system controller negotiating who will be using the bus next.
It's not quite so simple as that. With balanced reads and writes (the common case), the 2.3 and the 2.5 can saturate the DDR2 bus. So for balanced code, you're looking at a non-trivial 30% bandwidth increase. There is a latency increase for DDR2, but there is also an (alleged) latency decrease in the new G5 chipset. The previous G5 chipset had pretty bad latency compared to other chipsets, so improvements in the chipset likely offsets some of the latency increase. The cache is the biggie --- for workstation apps, its going to help quite a bit.
Also, I wouldn't be surprised if the new machines don't cost Apple appreciably less than the old ones. The new motherboard has 32 PCI-E lanes, something you only find in expensive NForce4 Pro boards with the extra PCI-E bridge chip. The new machines also have a substantially better graphics card with more RAM. It's still probably a cost savings for Apple, but not as much as you might think.
Define "loud". I've got one, and the only way I can hear a squeal is if I put my ear right against the vent slits on top. Even a couple of inches away, its inaudible over my computer (which uses a couple of 20db fans and a fanless graphics card). Actually, the loudest thing in the room is the overhead fluorescent light!
The 7800GT option will probably be shipping when Apple can get ahold of enough cards to ship in machines. As usual, the 7800GT is still rather new and not showing up in large volumes yet. Whether the DDR2 can take faster DIMMs is largely irrelevent, since the G5's FSB wouldn't be fast enough to be able to use them. Yes, it is dual channel.
The new dual core will perform better, given the larger caches, on-chip cache-to-cache bus, and faster memory. If it was an Athlon 64, it could have performed worse, because you'd be going from dual independent memory controllers to a shared memory controller, but the G5's have a shared memory bus anyway, even when there are two seperate physical processors.
Seriously. $1800 isn't a lot of money. In many engineering fields (mechanical, aerospace, maybe chemical and electrical), a full suite of software for a single engineer can cost well over $20,000. Something like CATIA's product-life-management suite starts at $12,000, and can cost over $30,000 per seat over 5 years when maintainence and support are factored in. Heck, even something like Matlab will cost you $2000 for the initial license, then another $4000-$7000 for all the plugins you need for your particular field.
What took them so long is the whole "getting it working" part. Popular Science has the luxury of reporting on ideas. Engineers actually have to actually figure out how to manufacture it efficiently and cheaply before putting it into production.
16bit is enough to represent a dynamic range of 90dB, which is far more than what you'll hear in the overcompressed crap that populates the charts today.
Most people who care about music aren't listening to the "overcompressed crap" that populates the charts today:) A commenly cited value (in academia) for the dynamic range of the human ear is 120 dB So 90 dB is definitely quite a bit short of "enough". Moreover, you seem to minimize the importance of higher sampling frequencies. Yes, their primary effect is to make the design of low-pass filters (both digital and analog) easier. However, the low-pass filter is a very significant and complex part of most analog designs, and giving the designer the headroom to use a more gently sloping low-pass filter can have a very big impact on the quality of the signal in the audible band.
You know, if it was as easy as you make it sound, we would have done it by now. Not only are there problems with the technology (a large one being heating of the skin due to aerodynamic friction), but just by the nature of the physics, it'll always cost you several times more fuel to fly at high mach numbers than at low ones. You don't even need to be an aerodynamicist to understand it. Drag goes up with the square of velocity, you figure out what that does to fuel consumption. Existing turbofan engines are extremely efficient, yet airlines still can't turn a profit. You think the solution is to make airplanes that are even less efficient?
Hey, let's not lump Java's failings in with GC in general. Java's memory management isn't exactly state-of-the-art. There are better GC algorithms, and more importantly, compiler optimizations that don't cause apps to allocate like crazy. Java compilers don't (yet) do any of these optimizations. The performance qualities of GC become much less relevant when you've got a compiler that doesn't do standard things like unboxing arrays of objects!
I always find it interesting to see how many C/C++ programmers think manual memory management is "fast". I think anybody spouting off about such topics should be required to have least *implemented* a fast malloc()/free(). It's a non-trivial task, and when you see that the hot-path is still dozens of instructions long, it really puts the cost of malloc() into perspective.
Slashdot Mirror
At what point does replacing a ten grand a year employee with a 100 grand machine become impractical?
When the machine does the work of ten people, or when its expected to last more than 10 years, or some combination thereof?
I sidestepped it because I don't consider it relavent. The arguments of the pro-lifers aren't couched in law, they are couched in morality. The law says if you let someone die right in front of you, when you could have saved their life, you aren't in the wrong. Morality says otherwise.
That's a contrived example. Given the science behind what we're talking about, we're not talking about sacrificing your own fetus, but somebody else's. That little bit makes all the differences.
Some would feign horror at the thought of "sacrificing" someone else's life to better your own, but that would be naive. We do this on a daily basis, just to maintain our comfortable life. We, the citizens of the United States could easily give up much of our standard of living to save millions of lives around the world. We do not, and should not, because we are only human, and by nature self-interested. The bare truth is that you've already sacrificed human life just to live as comfortably as you do. We sacrifice human life for many reasons, and the presevation of those we love is a more justifiable one that most.
Since when do you care about the death of another child? Since when does anyone? I ordered a PowerMac the other day. I didn't really need one, but they looked cool. I could have taken that money and given it to a starving kid in Bangladesh (my native country). The average Bengali works 5 years to accumulate that much money, and you can bet it would have saved at least one life. With my lack of concern, I basically allowed somebody to die. At a logical level, it is no different than if I had let someone get hit by a bus, without trying to warn them or push them out of the way. It might not have the same emotional impact, but at an abstract level, it is no different. This is a poignant, yet somehow painless truth.
What am I getting at? That we're all evil for letting children starve? No! We cannot live our lives shackled to the destinies of others. We cannot torture ourselves for the good we could do but do not. We are only human, we're not built to care for other people in that way. You claim to care about human life, but would you sacrifice your comfort to preserve it? Would you sacrifice a nice home for your family, a nice school for your children, your gas-guzzling, environment-polluting car, or even your trifling conveniences and luxuries? Would you sacrifice any of these things? There are some people who do, and while we admire such people at an intellectual level, almost none of us are willing to follow their example.
It is for this reason that I find the "pro life" argument disturbing. Here are people who allegedly care about life, yet, they spend an enormous amount of effort (and a not-insignificant amount of money), fighting for "abstract people", while letting "concrete people" die every day. I can't see how anybody can rationalize that. It makes no sense at a concrete level, or at an abstract level.
Loans are rather different from grants, as anybody with student loans will tell you. The claim that higher education is a racket is moronic. Its the superiority of American higher-education that keeps us the dominant technological power. Yes, it costs the government money, but that money pays huge dividends.
Now, you might think the IT field is not subsidized, but consider its origins. A very large fraction of IT technology was developed under government grants. The internet is the classical example, but even if you consider stuff that was developed at commercial research labs, you have to realize a lot of those developments were at least partially the result of government contracts. In the end, an engineer complaining about government subsidies is like a farmer complaining about government subsidies!
The right doesn't want to admit it was wrong about global warming and the left doesn't want to admit it was wrong about nukes.
So true! The right doesn't believe there is a problem, and the left thinks everything is a problem and won't allow any of the solutions. Everybody is waiting for a "perfect" solution, and even if there is one of these, we're not going to find one in the next century!
You do realize that engineering, as a profession, is more subsidized than paleontology? Paleontologists are usually supported by universities, which are at least partially supported by tuition. Meanwhile, entire branches of the engineering profession (eg: aerospace), exist either due to the DoD budget (for military stuff), or due to airline bailouts (for civilian stuff).
japanese anime also seems to do just fine in the US, fans seem to prefer the original japanese dialogue with subtitles.
First, Japanese anime is quite a niche. Second, its heavily edited in its American versions, and only the niche of the niche bother to see the original.
I don't see how your comment relates to mine? Where did Intel come into play? My point was that a dual-processor Athlon 64 system has an advantage over a dual-core Athlon 64 system because of the dual memory busses. On processors like the G5 and P4, which use a shared memory bus, dual-processor systems don't really have an advantage over dual-core ones.
This is the classical sort of thinking that perpetuates poverty. Africa is a continent of over 1 billion people. If they spent all their effort trying to provide people with food, water, and shelter, they'd never get anywhere. Its unsustainable development, and something that the international development community is quite aware of. The "give a man a fish" saying seems trite, but it really does fit. China is an excellent example. Right now, there is actually hope for the people of China that in the future they might not be as poor as they are now. If China had exerted all its resources trying to take care of its billion people, instead of building up their manufacturing and financial capacity, there would not be any such hope.
the video has 4GBps, and the PCI-x has 4GBps.
Video is 8GB/sec.
In terms of the Athlon dual-core, I'm not positive on this, but I think it's integrated memory controller maxes throughput at 6.4GBps, and the I/O at 8GBps.
Yep. The dual-core G5s have an edge on balanced read-write bandwidth, while the Athlon has an edge on total bandwidth.
The Athlon has a 128-bit memory bus and a 16-bit system bus, both of which are unidirectional (either in or out, not simultaneous).
The Hypertransport bus is 2 16-bit unidirectional links running at 2 GHz (hence the 8GB/sec of I/O bandwidth).
The Apple bus architecture eliminates the latency associated with flipping the direction of the bus as well as the latency associated with the processor and system controller negotiating who will be using the bus next.
Yep.
It's not quite so simple as that. With balanced reads and writes (the common case), the 2.3 and the 2.5 can saturate the DDR2 bus. So for balanced code, you're looking at a non-trivial 30% bandwidth increase. There is a latency increase for DDR2, but there is also an (alleged) latency decrease in the new G5 chipset. The previous G5 chipset had pretty bad latency compared to other chipsets, so improvements in the chipset likely offsets some of the latency increase. The cache is the biggie --- for workstation apps, its going to help quite a bit.
Also, I wouldn't be surprised if the new machines don't cost Apple appreciably less than the old ones. The new motherboard has 32 PCI-E lanes, something you only find in expensive NForce4 Pro boards with the extra PCI-E bridge chip. The new machines also have a substantially better graphics card with more RAM. It's still probably a cost savings for Apple, but not as much as you might think.
Define "loud". I've got one, and the only way I can hear a squeal is if I put my ear right against the vent slits on top. Even a couple of inches away, its inaudible over my computer (which uses a couple of 20db fans and a fanless graphics card). Actually, the loudest thing in the room is the overhead fluorescent light!
The 7800GT option will probably be shipping when Apple can get ahold of enough cards to ship in machines. As usual, the 7800GT is still rather new and not showing up in large volumes yet. Whether the DDR2 can take faster DIMMs is largely irrelevent, since the G5's FSB wouldn't be fast enough to be able to use them. Yes, it is dual channel.
The new dual core will perform better, given the larger caches, on-chip cache-to-cache bus, and faster memory. If it was an Athlon 64, it could have performed worse, because you'd be going from dual independent memory controllers to a shared memory controller, but the G5's have a shared memory bus anyway, even when there are two seperate physical processors.
What's ethically dubious about it? They tell you upfront what they're going to do with the code!
Heh. Neither Qt nor GNOME even existed when BeOS was developed. The BeOS interface was modeled on the Mac's.
Seriously. $1800 isn't a lot of money. In many engineering fields (mechanical, aerospace, maybe chemical and electrical), a full suite of software for a single engineer can cost well over $20,000. Something like CATIA's product-life-management suite starts at $12,000, and can cost over $30,000 per seat over 5 years when maintainence and support are factored in. Heck, even something like Matlab will cost you $2000 for the initial license, then another $4000-$7000 for all the plugins you need for your particular field.
What took them so long is the whole "getting it working" part. Popular Science has the luxury of reporting on ideas. Engineers actually have to actually figure out how to manufacture it efficiently and cheaply before putting it into production.
"Bovine Scatalogical Pathology"? Given your comment, it seems you suffer from "Foot in Mouth Disease".
16bit is enough to represent a dynamic range of 90dB, which is far more than what you'll hear in the overcompressed crap that populates the charts today.
:) A commenly cited value (in academia) for the dynamic range of the human ear is 120 dB So 90 dB is definitely quite a bit short of "enough". Moreover, you seem to minimize the importance of higher sampling frequencies. Yes, their primary effect is to make the design of low-pass filters (both digital and analog) easier. However, the low-pass filter is a very significant and complex part of most analog designs, and giving the designer the headroom to use a more gently sloping low-pass filter can have a very big impact on the quality of the signal in the audible band.
Most people who care about music aren't listening to the "overcompressed crap" that populates the charts today
To be fair, good upsampling already is done. libsamplerate is supposed to have a particularly good algorithm, though I haven't tried it myself.
You know, if it was as easy as you make it sound, we would have done it by now. Not only are there problems with the technology (a large one being heating of the skin due to aerodynamic friction), but just by the nature of the physics, it'll always cost you several times more fuel to fly at high mach numbers than at low ones. You don't even need to be an aerodynamicist to understand it. Drag goes up with the square of velocity, you figure out what that does to fuel consumption. Existing turbofan engines are extremely efficient, yet airlines still can't turn a profit. You think the solution is to make airplanes that are even less efficient?
Hey, let's not lump Java's failings in with GC in general. Java's memory management isn't exactly state-of-the-art. There are better GC algorithms, and more importantly, compiler optimizations that don't cause apps to allocate like crazy. Java compilers don't (yet) do any of these optimizations. The performance qualities of GC become much less relevant when you've got a compiler that doesn't do standard things like unboxing arrays of objects!
I always find it interesting to see how many C/C++ programmers think manual memory management is "fast". I think anybody spouting off about such topics should be required to have least *implemented* a fast malloc()/free(). It's a non-trivial task, and when you see that the hot-path is still dozens of instructions long, it really puts the cost of malloc() into perspective.