What about production costs for the music? The standard practice is to charge these against an advance payment to the artist -- still a loss to the publisher if they don't sell enough copies to cover the advance.
Actually, the main argument is not what your rights are under the law right now -- that is clear. The problem arises when the RIAA tries to use their data to argue to lawmakers that the current laws aren't working for their intended purpose, and hence should be revised to help the poor little musicians get back on their feet.
There is one severe problem with your reasoning here: you only account for variable costs, i.e. the marginal amount spent per item produced; you completely ignore fixed costs, which must be amortized over all items sold. For your examples, most of the cost is variable cost, so the impact of a small number of thefts on fixed cost can essentially be ignored. In the case of music or video recordings or software, variable costs are miniscule, but up-front fixed costs are huge. So the marginal cost to the producer of a CD in a box or a pirated copy is almost the same, but the former can act as a vehicle to amortize fixed costs, while the latter cannot. How does your model take this into account?
If you can't afford to buy a copy of some (probably pretty bad) music, I think you are morally obligated to respect the creators' wish not to obtain it via other means. In this case, it has nothing to do with obeying politicians; basic copyright law just happens to be the most reasonable idea. After all, it's not like you need this music to save your life (this issue with patented medications is a bit trickier; some countries, like India, choose to simply ignore those patents). Anyway, the main reason so many people illegally download music is convenience -- not because they can't afford it.
No, the whole idea of a proof is that it is a statement about something which can be verified as correct using a very simple algorithm. The problem is that the specification, to which you are proving your implementation conforms, may itself have bugs.
What good does a couple hundred physical registers get you when you can have hundreds of thousands of effective registers with l1 / l2 caches?
First of all, caches still tend to have higher latencies than registers, e.g. 3 additional cycles. Secondly, modern x86 processors do have hundreds of physical registers, they simply are not architectural. It may, however, be the case that since these renaming registers are distributed throughout the chip, they can be accessed more efficiently than a monolthic register file. Finally, x86's success is despite, not necessarily because of its lack of architectural registers. 8 not-quite-general purpose registers mean you do need extra instructions to swap data in and out of the register file. Unfortunately, the graph-coloring register allocation algorithm is not exactly efficient on such a constrained register file (I've heard Intel's compilers use a somewhat modified, considerably more complicated algorithm). Finally, consider the Opteron: it introduces 8 more architectural registers, resulting in considerable performance gains in 64-bit mode.
Have you actually looked at the work done at Microsoft Research? It's actually quite good, seeing as how they've been able to hire some of the most best researchers in programming languages, graphics, databases, etc. The thing is, these guys are mostly only concerned with cutting-edge research, not product development. The other thing is that many of these people are working in fairly well-established fields, rather than new integrative efforts. To them, the fact that they're employed by Microsoft is only a minor detail of funding source. Microsoft still makes out fairly well from this arrangement, since they maintain all rights over this research, and if it piqued the interest of of research-minded people in Microsoft *Development*, it could perhaps make its was into an interesting product.
First of all, "Indian" ain't a language, mofo. Secondly, any person in India who Americans would have enough desire to speak to that they could afford paying a real-time phone translator is probably already quite proficient at English (seeing as how it's one of their official national languages, notably used in academia).
In my mind, the reasons people like to have their own car are: (1) always available to go anywhere (uh, unless it's broken), (2) less hassle due to the insurance infrastructure, and (3) they keep all their junk in there. People are filthy animals.
That's exactly how it works right now. You license out your technology to the manufacturer (or reseller, or whatever) in exchange for money. Like the fraction of a PC running Windows that goes to Microsoft. Or the fraction of the price of your $20 5-wire USB cables that goes to Intel, et al.
Most standard telephone networks do NOT communicate over the Internet. Telephone companies have separate networks for phone calls only, which use different protocols and associated hardware than IP in order to guarantee high reliability. Furthermore, they have the appropriate legal agreements with their peers to make it a big incentive that even when your call has to pass through a different company's network, you will still receive the desired level of reliability, delay and audio quality.
Slashdot Mirror
What about production costs for the music? The standard practice is to charge these against an advance payment to the artist -- still a loss to the publisher if they don't sell enough copies to cover the advance.
Uh, yes they did.
Too bad this article is about an Itanium cluster, not x86.
GCC? On Itanium? Optimized quite well? Whatever. Check out Trimaran for the HP/Illinois/NYU compilers which basically inspired Itanium.
Actually, the main argument is not what your rights are under the law right now -- that is clear. The problem arises when the RIAA tries to use their data to argue to lawmakers that the current laws aren't working for their intended purpose, and hence should be revised to help the poor little musicians get back on their feet.
There is one severe problem with your reasoning here: you only account for variable costs, i.e. the marginal amount spent per item produced; you completely ignore fixed costs, which must be amortized over all items sold. For your examples, most of the cost is variable cost, so the impact of a small number of thefts on fixed cost can essentially be ignored. In the case of music or video recordings or software, variable costs are miniscule, but up-front fixed costs are huge. So the marginal cost to the producer of a CD in a box or a pirated copy is almost the same, but the former can act as a vehicle to amortize fixed costs, while the latter cannot. How does your model take this into account?
If you can't afford to buy a copy of some (probably pretty bad) music, I think you are morally obligated to respect the creators' wish not to obtain it via other means. In this case, it has nothing to do with obeying politicians; basic copyright law just happens to be the most reasonable idea. After all, it's not like you need this music to save your life (this issue with patented medications is a bit trickier; some countries, like India, choose to simply ignore those patents). Anyway, the main reason so many people illegally download music is convenience -- not because they can't afford it.
No wonder I didn't understand. And how can someone who can devote so much time to merely playing video games have "not enough time" for anything?
VMS ain't Unix, dude.
How is a Dell laptop (i.e. crap -- I have one) any better than a Compaq?
That's awesome.
What's with the Arbitrary capitalization and "Free Trade between Europe" -- shouldn't that Be "free trade within Europe?"
No, the whole idea of a proof is that it is a statement about something which can be verified as correct using a very simple algorithm. The problem is that the specification, to which you are proving your implementation conforms, may itself have bugs.
First of all, caches still tend to have higher latencies than registers, e.g. 3 additional cycles. Secondly, modern x86 processors do have hundreds of physical registers, they simply are not architectural. It may, however, be the case that since these renaming registers are distributed throughout the chip, they can be accessed more efficiently than a monolthic register file. Finally, x86's success is despite, not necessarily because of its lack of architectural registers. 8 not-quite-general purpose registers mean you do need extra instructions to swap data in and out of the register file. Unfortunately, the graph-coloring register allocation algorithm is not exactly efficient on such a constrained register file (I've heard Intel's compilers use a somewhat modified, considerably more complicated algorithm). Finally, consider the Opteron: it introduces 8 more architectural registers, resulting in considerable performance gains in 64-bit mode.
I'm sorry to hear about your experiences. Was this a community college, perchance?
Have you actually looked at the work done at Microsoft Research? It's actually quite good, seeing as how they've been able to hire some of the most best researchers in programming languages, graphics, databases, etc. The thing is, these guys are mostly only concerned with cutting-edge research, not product development. The other thing is that many of these people are working in fairly well-established fields, rather than new integrative efforts. To them, the fact that they're employed by Microsoft is only a minor detail of funding source. Microsoft still makes out fairly well from this arrangement, since they maintain all rights over this research, and if it piqued the interest of of research-minded people in Microsoft *Development*, it could perhaps make its was into an interesting product.
Them's injuns to you, dot-head.
First of all, "Indian" ain't a language, mofo. Secondly, any person in India who Americans would have enough desire to speak to that they could afford paying a real-time phone translator is probably already quite proficient at English (seeing as how it's one of their official national languages, notably used in academia).
Same deal at Illinois.
This is Slashdot... you sure it wasn't ECE?
In my mind, the reasons people like to have their own car are: (1) always available to go anywhere (uh, unless it's broken), (2) less hassle due to the insurance infrastructure, and (3) they keep all their junk in there. People are filthy animals.
Working with broken windows invariably leads to open sores.
That's exactly how it works right now. You license out your technology to the manufacturer (or reseller, or whatever) in exchange for money. Like the fraction of a PC running Windows that goes to Microsoft. Or the fraction of the price of your $20 5-wire USB cables that goes to Intel, et al.
Most standard telephone networks do NOT communicate over the Internet. Telephone companies have separate networks for phone calls only, which use different protocols and associated hardware than IP in order to guarantee high reliability. Furthermore, they have the appropriate legal agreements with their peers to make it a big incentive that even when your call has to pass through a different company's network, you will still receive the desired level of reliability, delay and audio quality.
ahem, "voila" comes from French, mademoiselle.