Why do computers in general need to get any faster these days?
Medicine, physics, engineering, and AI all benefit from increasing computer power. There are probably numerous other fields that benefit secondarily, but those are probably the most important. Protein folding and cellular simulation will ultimately do more for medicine than anything in previous history, probably the same with physics and engineering. Nanotechnology will require massive computing power to design and test.
Computers would not get faster so quickly if it was only researchers who needed them: Supercomputing took decades to develop, but it is now growing as fast as desktop processors can be improved. The primary benefit of everyone buying the latest processor is really the benefit to people who actually need all the computing power. Additionally, if you think your parent's computer power is wasted, throw Folding@home on it.
I like the fractal universes idea. As for simulating the spatio-temporal Universe, the proposition is that it's not possible to simulate it with any device less complex than itself. It doesn't preclude a meta-Universe that can do so, but such a construct most likely wouldn't be observable from our point of view.
Two possibilities, one is to wait until we have more of the universe available for simulation and just focus on simulating the relatively simpler past. The second is to somehow escape the heat death of the universe and use the infinite time available to run any simulations we want in as much detail as we have computing memory. A perfect simulation of the entire universe may be possible, it really depends on whether we can encode the universe in a smaller format than it currently takes now, e.g. run the simulation on a losslessly compressed version. Even then we have quantum uncertainty to deal with, so in reality we would be simulating one of many possible universes similar to our own.
The meta-Universe you think of I have thought of, which I define as "The sum of Everything that was, is, will be, can be, and could be". A kind of meta-set, which as you mention would necessarily have to contain itself and its opposite along with everything else. Some call that by a three-letter word. From that point of view, there is no free will. But that the whole has no free will doesn't remove the possibility that its parts as differentiated within single spatio-temporal expressions of possibility could have free will.
Since such a universal set is so large, it certainly doesn't preclude god or gods from existing, in fact it essentially implies their existence for some universes. The question falls back to probability and the likelihood of finding one's self in a universe with an active, meaningful deity, essentially the same question we've dealt with in philosophy for thousands of years.
My understanding of free will essentially boils down to the simulation argument; If we can't simulate our universe in advance to know what we will do and the results of our actions, then we have free will. It is free in the sense that no matter what we know at any given moment, the choice of what we will actually do is unknown. Even if the universe is purely deterministic, our inability to know the end result is essentially what makes our choices free; free of dependence on the knowledge of final outcomes. If the choices are purely random, that just means that there is something about the universe that is free from its rules, having much the same effect. I don't think free will can be distilled to anything simpler, and anything greater ignores causality. For instance, people who believe in a spirit or soul assume that it gives them free will in an otherwise naturalistic universe, but ultimately where does a soul or spirit's action come from? It either follows from some metaphysical rules, or it is random. In either case, it is not the soul or spirit that is the source of free will, but the underlying disconnectedness between choices and knowledge of final outcomes.
In modal realism, free will is acted on by something like natural selection. Free will only exists in the universes where conscious entities have the ability to make decisions and not know the outcome, and by extension those decisions result in universes where free will continues to exist. That doesn't mean free will is a stable state, nuclear war or the discovery of omniscience could eliminate it, and it's apparent that free will arose through evolution.
One really interesting result of modal realism is that ultimately no consciousness ever truly dies; it persists in some universe, however unlikely it may be. This causes some philosophical problems, because what is the likelihood that existing forever will be a pleasant experience? One line of reasoning leads me to believe that unpleasant universes kill off their inhabitants quicker than pleasant ones, thus leading to a much greater percentage of pleasant univer
We have most of the answers right in front of us, we're just afraid to ask the right questions. Reality exists because we're aware of it. We are creating the Universe by perceiving it and by choosing which Universe we want to be in. But as someone else in this discussion has said, how are we so different that our status as observers makes us so special? Our consciousness ultimately derives from the same particles that make up what we observe, we simply have a higher order of organization and "synchronicity" so that our consciousness, for a while, is greater than simply the sum of what we are made of. The only conclusion that can be obtained from that realization is that elementary particles are conscious and have some measure of free will. That is what quantum probabilities measure: the possibility the particle will "choose" the different possible paths it can take within the laws of the spatio-temporal Universe. It is a very basic kind of consciousness, as the perception it has of the rest of the Universe is extremely limited: its own physical characteristics, and what other particles it can interact with in exchanges of energy. As particles start bonding together and organizing, we get to higher and higher degrees of order, organization, and what I call "synchronicity of purpose" where eventually as in higher primates it can actually work together to achieve a higher level of consciousness and awareness of its environment, but thus creating a different level of reality which is certainly more powerful but not necessarily more complete or "true" than the simple interactions the single particle can achieve.
I come at the problem from a much more mathematical point of view, but ultimately I think your description is correct. For any given universe, there are rules that govern the behavior of everything in it. The thing is, for any set of rules there are an infinite number of universes that can exist within those rules. The important thing seems to be the rules that choose how any given universe will change. For me, consciousness exists as the ability for a system to model itself, e.g. something constructed within a system that mimics the entire system, rules and all. It doesn't have to be an exact model (which leads to Zeno like paradoxes), just a working model. In other words, if some part of a system is self similar to itself, that is the beginning of consciousness. When the self similar model can be manipulated by the same rules (encoded in the system) that govern the system itself to explore other possibilities for the configuration of the system, consciousness is complete. There is an extension to consciousness which is self action, which is partly separate from the model and has the ability to change the system itself based on interactions with the model. Self awareness occurs when the model includes a generalization of the self action itself, e.g. it knows that it is a model of the system with the ability to change the system. You can reverse the definition and say that the entire system is conscious because of its ability to change itself, but in precise terms the consciousness is limited to the model. If the model is destroyed, the system remains but consciousness is gone.
I'm also a modal realist in a strict mathematical sense. I believe that everything expressible with mathematics exists just as much as the universe, and probably the universe exists because of its expressibility in some form of mathematics. I also think that there are probably higher mathematical models than we can conceive of in this universe that also exist, but we don't have the resources to actually construct those models. Still, it's an interesting question whether set theory is sufficient to describe any possible model at any level, or if there needs to be something else bigger than set theory (and category theory) to describe something. My guess is that there is, but those are the kinds of things we won't be able to imagine in this universe.
I still think the Universe isn't a simulation, because some
Well, it all makes sense, if you think of it. Whoever is running this MMO we call RL, can't possibly have the resources to simulate every single particle all the time. So until someone actually goes and observes the damn thing, there's no need to actually spawn/instantiate it.
What if they do have the resources? In other universes, being able to hold and manipulate the set of reals may be just as natural as working with stones or atoms. Push a few sets of reals together to make a space, squish some reals out of another space to make a function, map the function to the space and see what comes out.
Or say you're hacking away at a copper ore vein with your trusty cold iron pickaxe, like a good dwarf. Sometimes you get just a piece of copper ore, sometimes you also get 1-2 pieces of stone, sometimes you get a Shadowgem, or a Tigerseye or Malachite. Were they already there before you started to hack at the ore vein? Or did they exist only as a probability until someone actually gets that loot window?
It all depends on how the random number generator gets seeded and when it gets called. If you repeated the exact same sequence of events from the beginning of WoW coming online, including all character actions at the right time, then the same things would happen. Even if Blizzard had a true random number generator, that just reduces the problem back to the case of our universe: If everything in the universe was reset to some point in time, what would cause a different outcome? With quantum mechanics, there is no way for us to talk about a single state that the universe has, just a very large set of states of possible universes that match all our measurements. This means that our past and future is also a nondeterministic set of possible universes. The question of whether to believe that the universe makes an infinite number of random decisions at each step to pick one and only one result or that all the possibilities exist simultaneously is basically the question between some form of determinism versus modal realism. I find the latter much easier to accept on a mathematical and philosophical basis, because it gets rid of most of the unanswerable philosophical questions altogether, such as what existence is, why it exists, and how. Modal realism just means that everything that's possible exists. The question of what is possible is interesting to think about, but basically boils down to anything that we can imagine, and probably more.
If you can afford it and the police have a reasonable response time, an alarm monitoring company is probably the best solution. Cutting wires won't help because disconnecting the house from the monitoring company triggers the alarm on their end (and maybe at the house, too) It would be interesting to know what infrastructure the alarm company uses for their wires. I imagine they could lease lines from the phone company, but for video it would tend to be rather expensive. Maybe they borrow frequencies on the cable system or actually did go lay their own wiring in the neighborhood. If it's an upscale neighborhood where everyone has an alarm, I would guess the latter. Wire all the neighborhood into one box and then pay for the T1 or other connection back to the main office and split the costs. Is the video from the cameras high quality, or small and/or highly compressed?
Have you looked into running your generator during peak usage hours to reduce electricity costs? It would be interesting if natural gas ended up being cheaper than buying the electricity during that time. It sounds like the generator powers the entire house back through the existing wiring, which may mean that it's overloaded if you see voltage and frequency variations that high. Generally it should run right at 60 Hz unless it's under heavy load and can't keep up. I'm surprised that your UPS didn't like the voltage fluctuations since they're supposed to handle it as far as I know. They will definitely switch to running off the battery when the voltage or frequency goes out of whack, but should keep the battery charged from the wall power regardless.
The Door in the kitchen coming from the garage is controlled by a set of really strong magnets and and hooked through the security system too. Once it is locked, you need about as much force necessary to kick a regularly locked door in to open it.
Too bad most thieves cut the power before trying to break in. You do have your cameras on a UPS and the video sent to a remote location in realtime, right?
QNX has the advantage that I/O, like almost everything else in QNX, is done via inter-process message passing operations. The message passing system uses priority queues, and so requests to file systems and devices get handled in priority order. So resource managers (file systems, device drivers, etc.) don't have to explicitly handle priorities; it's done for them. Some resource managers, like disk handlers, process multiple requests at a time so they can reorder them to optimize access, but network devices and such are FIFO at the resource manager level and priority ordered at the message level.
Yeah, and then a priority inversion killed the Mars Rover (I think that's what it was). With standard scheduling, things would have just slowed down, but wouldn't have hung the whole system.
1. They better both pay equal amounts, or the neutrality is somewhat in question. The court should get to appoint an expert from a list of candidates submitted by both parties, to make it a little more impartial.
2. With MD5 hashes of everything and a redundant, untouched copy of the disk, it shouldn't matter to the plaintiffs who deletes the files. Hopefully, the respondent can get another computer expert to help out with their lawyer present and go through a list of files, including caches and swap files, that should be deleted as privileged data. Since this is an RIAA copyright case, every file except the ones in the P2P directories should be privileged, although I don't know how that argument would stand up in court. I assume that the plaintiffs could challenge the extent of deleted privileged data, but they should need a good reason to be looking at anything but the P2P software, it's directories and logs, the registry (assuming it's Windows), and reasonable parts of the OS logs (limited to the dates in question, or thereabouts).
3. Do they have the option of making a third copy of the disk, omitting the sectors that the privileged files resided on? That's far more secure and prevents the plaintiffs from attempting expensive data recovery on the drive with "deleted" files. What the ruling really should have specified was a list of MD5 hashes of each filesystem block, and then a hash of the entire list. That would ensure that when files were being deleted, only the blocks they were on would no longer match the hash, and the rest of the files could be verified as being from the original drive. Maybe they did do that, I don't know the details.
Smartfilter is an incompetent bunch of fucks too, with a heap of garbage for management software. We constantly have to get sites "recategorized" as you put it, because their censors are too lazy to do their job, and their software makes per-page exception lists impossible.
Real web browsers handle scripting languages with fewer problems than Office. So does the Java virtual machine. Office is just poorly written without any real security or privilege separation in mind.
Due to their extensibility, pdf and postscript are suspect in the eyes of the truly paranoid, let alone the complex modern formats.
Postscript is a Turing complete programming language, but I haven't heard of many exploits in ghostscript or PS printers. Clearly there's just something broken at Microsoft. It's possible to build secure interpreters for general scripting languages, and Microsoft just hasn't been able to accomplish it with Office. They managed to write.NET, but my guess is it's a vastly different (in terms of technical competence) team than the one working on Office.
McDonald's causes great hassle for Burger King as they refuse to release the recipe for the Big Mac's secret sauce. Sadly, this will only be available at McDonald's for the time being. There are no plans for cross-restaurant release.
In other news, the bread and burgers still come from wheat and cows (mostly). My point? The content is what people want, not the format. People couldn't care less what format their content is in as long as it works. Microsoft is trying to ensure that only the Microsoft formats that come with Windows will work, and monopolize the content. It would be like McDonalds buying all the cows in the world.
Much as I love Linux and free software, it is self-defeating and unrealistic to demand that Microsoft (and other companies) support Linux. Perhaps the much-vaunted free software community should produce its own solutions that are better then the closed-source competition? Instead of complaining about what other people do, take responsibility for your own needs and write the software you want.
As non-windows OS's become more popular (remember, there are billions of cell phones and millions of PDAs, it's not just Linux and Mac), content producers are going to need cross platform and open standards. Microsoft is not giving their customers the tools to reach all potential viewers, therefore Microsoft is failing in its job as a software maker.
Adobe(Macromedia) Flash has been around for a LONG time and I have yet to see anyone attempt to come up with a serious Open Source alternative.
As a user of open source, I generally detest flash. It's only good for making annoying, flashy, moving advertisements and web pages. I hate flash movies, because there's no reason not to just provide an AVI or MP4 stream. I hate flash menus, because if they're broken it's impossible to view the source and maybe have a chance of figuring out where they were trying to send you. Everyone hates flash ads. For interactive web pages, there's cross platform AJAX that you can actually see the source for, and Java for applications that just don't work in AJAX.
Is the router authorized to let you connect? I don't think so. It can connect, but it has no authority. It is like a neighbor's electric plug point. It can supply electricity, but that doesn't mean you can use it without neighbor's permission (plug point's permission is not enough).
It would be permissible if my neighbor always left his extension cord plugged into his house and the other end laying in my yard or coming in through my window. Before anyone even sends a DHCP request, they have to first receive the SSID broadcast from the wireless router. It is actively advertising its presence and willingness to accept and respond to unencrypted packets.
It's not like the bad guys have devil horns and the civilians carrying weapons have halos so you know what to shoot at.
The good guys aren't pumping bullets into unarmed civilians. These are my best attempt at some rules of engagement for armed citizens:
1. Don't shoot first.
2. Only shoot at an attacker if they have just obviously and directly injured an unarmed person.
3. If anyone shoots at you, only shoot back after you have identified yourself as a good guy and they keep shooting.
An attacker does not necessarily need to be shot to stop an attack, just convinced that it's futile to shoot anyone else. There will be an uneasy standoff until witnesses can identify who the real attacker is, but in general reaching a stalemate or physically stopping all the attackers should occur reasonably quickly, with a minimal number of shots fired. Feel free to find weaknesses in the rules.
If there are too many attackers, then the situation is basically hopeless, especially if they are well coordinated. An obvious example is a special ops team that can infiltrate a military installation in spite of numerical disadvantages. The biggest problem the armed civilians have is the identification of each other and attackers, but in an open area this usually becomes apparent when the unarmed civilians run and hide, the attackers keep shooting them, and the armed civilians are holding guns and looking for attackers. The worst case for the armed civilians is when the attackers return fire, reducing the available information about who the attackers and armed civilians are. That's why a heavy emphasis on rule number 1 is important. The more people who are armed and *not* shooting, the easier it is to identify the clique of armed citizens. The unarmed civilians will ultimately aid in the attacker identification process.
It almost makes me think mandatory police (or military) training would be a good idea.
Just give mplayer or mencoder the -ss option. I think this one needed about 150 seconds, longer than most of the other Sony DRM failures and probably the reason it doesn't work on some players. Normally it's only about 30 seconds into the movie that they put the bad sectors.
This copy protection prevents most people from renting/borrowing a DVD and making a copy of it. Until people download the latest software for cracking it. This is mostly targetting non-technical people who were given DVD Shrink by a friend.
If this is the typical "bad sectors at the beginning part with the lady holding a torch" DRM that Sony is so fond of, 'mencoder -ss 60' usually takes care of it. Just adjust the time to match the opening credits, and rip away. Mplayer accepts the same option if you just want to watch it.
I would find it difficult to characterize checksums or parity as derivatives. More likely they're either irrelevant (checksums) or just part of the overall reproduction (parity; see the videogame cases from the early 80's).
If a portion of the work can be recovered from the parity, doesn't that make it a derivative work? For instance, CDs using CIRC carriy 1 byte of parity for every 3 bytes of data, That means if the data itself could be split from the parity, there would be one perfect copy and one degraded copy (approximately 8 bit mono, versus the 16 bit stereo of the CD). If the parity was increased, the ratio of data bytes to parity bytes could be lowered further to produce increasingly accurate representations of the data. After there are more parity bytes than data bytes, a subset of the parity bytes can be used to recover the entire data. Another interesting possibility is secret sharing where data is split into multiple pieces of which some number are required to reassemble the original work, and without a sufficient number no information can be obtained. Would it be illegal to split a copyrighted work into three pieces, of which any two were required for reconstruction of the work, and then share the pieces individually? No portion of the actual work would actually exist in any of the pieces, just a mathematical relationship that allowed the work to be reconstructed from at least two pieces. See Shamir's secret sharing algorithm for instance.
In the past, when someone has actually bothered to raise the issue, the courts have found it to infringe. But usually, who cares? Since statutory damages are calculated per work, not per infringement, and since compensatory damages for the incidental copies would be overvalued at a nominal sum of $1, no one cares about them. The point is to get all the direct infringers you can, and from there to get the indirect infringers. This used to be of some interest when there were ISPs who were intermediaries and who had deep pockets, but they usually fall under the safe harbor now.
Every RIAA lawsuit I've heard of has targeted uploaders, perhaps because it's the only way they can find filesharers and still have a case, so it certainly seems that creating a copy and sending it over the network is considered infringement. The cases I've looked at claim both reproduction and distribution infringement. Looking at RTC v Netcom I found MAI SYSTEMS CORP. v. PEAK COMPUTER, INC., 991 F.2d 511, which ruled that simply loading software from permanent storage into RAM constituted the creation of a physical copy, which is apparently the same argument used to enforce EULAs.
You'll find that it is more nuanced than that. Again, the 512 safe harbor has really taken over here, but for an earlier case dealing with these issues, I strongly suggest you read Religious Technology Center v. Netcom. Of course, the reason we have the safe harbor is because not every court followed this particular logic. The uncertainty was not good for the Internet.
As far as I can tell, the difference is that service providers do not necessarily know the legal status of the content they allow users to store and transfer. On the other hand, someone running P2P software will usually know which files are being shared and their legal status, although this is not always the case. As the Napster case showed, even direct knowledge that a service is being used for illegal activities is enough to warrant contributory copyright infringement.
When Alice has a file on a server and Bob downloads it, what happens is that Bob cause his computer to request information from Alice's computer. Alice's computer responds and sends information to Bob. But it does not send a copy because hard drives do not travel over Ethernet. When Bob's computer receives that information, it writes it to various storage media, such as RAM, or hard disks. By doing that, it creates a new copy, namely, that RAM, or those hard disks.
If you really want to boil it down, when Bob requests information from Alice, Alice has to read the data from her hard disk (which has in addition to the original work, a derivative work in the form of error correcting codes), the hard disk stores a copy of the original work in its cache, the drive copies the information onto the data cable, the controller copies it into internal registers and then copies it onto the system bus when appropriate (either via DMA or when the CPU issues in IO read instruction). For DMA transfers, the information is copied directly into system RAM, for direct IO the CPU copies the information off the bus into internal registers and the cache, then copies it into internal RAM. To actually send the data to Bob, the CPU then copies the data from memory (requiring a read into the CPU and cache for each byte/word of the data) into a second copy in a network buffer (some OS's use scatter-gather to keep the data in one memory buffer and skip this step), then the CPU has to copy the data back over the IO bus to the network card/modem which requires more loads, caching, and stores. Finally, the network card or modem copies the data onto whatever network it's connected to.
In this relatively simple example, there are at least 8 copies made by Alice's computer. That doesn't even include the case where intermediate P2P software accesses the data from the disk, copies it and hashes it, and does whatever else before finally sending it to the network stack. It's inconsistent for the court to rule that infringement of reproduction rights does not occur when Alice makes her copies, but does when Bob makes his copy to his hard drive, because EULAs have been ruled valid on essentially the same claim; that Alice has to make internal copies of software she has purchased in order to use it.
And who caused this to happen, who bears the responsibility? That would be the downloader. Really, the only situation in which the downloader could convincingly argue that it wasn't his volitional act that caused the infringement would be if someone else had control of his computer and was making it do that.
If Bob asks Alice to violate a law, and she violates it, she is fully responsible for her actions unless she was coerced. Bob might have accessory charges or be charged with contributory infringement, but it wouldn't affect Alice's culpability.
Discussing whether or not to allow ED-209 to shoot at humans is silly. ED-209 can't tell the difference in the first place. If we allow him any targetting discretion at all, he's going to shoot at just about everything. That's the reality imposed by the technology and we're not on the verge of overcoming it.
That's not true. Current computer vision systems can easily distinguish between most people and most vehicles. People are tall with dangly appendages that swing around, vehicles are wide and low and don't have so many moving parts. You could literally hook a computer vision system up to an autocannon and stick it on a Darpa Grand Challenge vehicle to create a search and destroy killbot.
Is it just me or is a discussion of ethics laws for robots premature given the state of the art in artificial intelligence? If you want to teach a machine not to harm humans, it helps to first teach the machine the difference between a human and every other object it encounters.
The discussion is not premature. Land minds already kill humans, machines indiscriminately, and therefore would be in violation of this treaty. The interesting thing is that guided cluster bombs could violate the law if they were not built to only target heat sources large enough to be vehicles. I have no idea of the mis-targeting rate of guided cluster bomblets, but my guess is that the military doesn't mind if the soldiers standing next to the vehicles get targeted too...
Other automatic targeting systems will be developed soon. At some point the military will want to run borders autonomously instead of wasting troops to stand around all day. Autocannons are the logical choice, and those weapons will have to decide whether they are pointed at children, soldiers, animals, ambulances, or tanks. I don't think anyone wants to return to a no-mans-land style of warfare, but indiscriminate killing machines would lead to that.
I assume by "modern cyphers" you mean public key cryptography with large keys.
I was talking about symmetric ciphers like AES, Twofish, Cast5, and even 3des is still considered secure. Asymmetric (public/private key) ciphers rely on a symmetric cipher to do bulk encryption of data with a random key, and the random key is encrypted with the asymmetric cipher.
if I want to keep a secret, I'm going to assume those will be breakable in a few years or less using quantum computing or something else.
Quantum computing is not expected to be able to improve attacks on symmetric key ciphers to less than sqrt(n) work, and that only with massive amounts of memory to search the keyspace. In that case, AES with 256 bit keys is still as secure with all powerful quantum computers as AES 128 is today.
Additionally, there are no known quantum algorithms for quickly solving the discrete log problem (DLP) for elliptic curves, which means that elliptic curve cryptography will probably provide secure asymmetric ciphers once quantum computers are powerful enough to attack ciphers relying on standard DLP and factoring of integers.
Therefore, if my message must remain secret for more than a few years, I am going to do something to make it impossible to know when the correct key is guessed. Splitting the message itself is one way to do this.
You're relying on being able to keep separate copies of your data secure from coming together somehow. If you can do that, why can't you just keep one copy physically secure and forget the encryption? E.g. if you're relying on keeping a one time pad secret for several years, couldn't you just store your data in place of the one time pad? One of the basic assumptions of modern cryptography is that the entire ciphertext is always available to the attacker, along with complete knowledge of the ciphers and protocols used to encrypt the plaintext. The only things the attacker is assumed to lack are the keys themselves. Weaker assumptions lead to either weaker cryptosystems or complex ad-hoc mixtures of good and useless cryptography, making the entire system more difficult to analyze for flaws.
If you do really need perfect security, then skip the symmetric and asymmetric ciphers entirely and rely on secret sharing to distribute the pieces of your data to secure locations. The wikipedia link I posted has enough information to pick a scheme and implement it or find some software that will do it. Just be aware that anyone who needs your perfectly secure data is probably capable enough to follow you everywhere you go, and is already doing so if you have secrets that important, making multiple "secret" locations impossible. You would probably be safer to never write the secrets down in any form.
Security is always a trade-off. If you don't need perfect security and don't have spooks watching your every move, then modern ciphers are more than adequate for security and much more efficient and easy to use. Adding complexity to an already working system just complicates verification and makes it possible to leak information or make mistakes that can be used to break the system.
Slashdot Mirror
Why do computers in general need to get any faster these days?
Medicine, physics, engineering, and AI all benefit from increasing computer power. There are probably numerous other fields that benefit secondarily, but those are probably the most important. Protein folding and cellular simulation will ultimately do more for medicine than anything in previous history, probably the same with physics and engineering. Nanotechnology will require massive computing power to design and test.
Computers would not get faster so quickly if it was only researchers who needed them: Supercomputing took decades to develop, but it is now growing as fast as desktop processors can be improved. The primary benefit of everyone buying the latest processor is really the benefit to people who actually need all the computing power. Additionally, if you think your parent's computer power is wasted, throw Folding@home on it.
8 cores is enough to do real time raytracing of moderately complex scenes. See OpenRT for examples and the systems they've used.
I like the fractal universes idea. As for simulating the spatio-temporal Universe, the proposition is that it's not possible to simulate it with any device less complex than itself. It doesn't preclude a meta-Universe that can do so, but such a construct most likely wouldn't be observable from our point of view.
Two possibilities, one is to wait until we have more of the universe available for simulation and just focus on simulating the relatively simpler past. The second is to somehow escape the heat death of the universe and use the infinite time available to run any simulations we want in as much detail as we have computing memory. A perfect simulation of the entire universe may be possible, it really depends on whether we can encode the universe in a smaller format than it currently takes now, e.g. run the simulation on a losslessly compressed version. Even then we have quantum uncertainty to deal with, so in reality we would be simulating one of many possible universes similar to our own.
The meta-Universe you think of I have thought of, which I define as "The sum of Everything that was, is, will be, can be, and could be". A kind of meta-set, which as you mention would necessarily have to contain itself and its opposite along with everything else. Some call that by a three-letter word. From that point of view, there is no free will. But that the whole has no free will doesn't remove the possibility that its parts as differentiated within single spatio-temporal expressions of possibility could have free will.
Since such a universal set is so large, it certainly doesn't preclude god or gods from existing, in fact it essentially implies their existence for some universes. The question falls back to probability and the likelihood of finding one's self in a universe with an active, meaningful deity, essentially the same question we've dealt with in philosophy for thousands of years.
My understanding of free will essentially boils down to the simulation argument; If we can't simulate our universe in advance to know what we will do and the results of our actions, then we have free will. It is free in the sense that no matter what we know at any given moment, the choice of what we will actually do is unknown. Even if the universe is purely deterministic, our inability to know the end result is essentially what makes our choices free; free of dependence on the knowledge of final outcomes. If the choices are purely random, that just means that there is something about the universe that is free from its rules, having much the same effect. I don't think free will can be distilled to anything simpler, and anything greater ignores causality. For instance, people who believe in a spirit or soul assume that it gives them free will in an otherwise naturalistic universe, but ultimately where does a soul or spirit's action come from? It either follows from some metaphysical rules, or it is random. In either case, it is not the soul or spirit that is the source of free will, but the underlying disconnectedness between choices and knowledge of final outcomes.
In modal realism, free will is acted on by something like natural selection. Free will only exists in the universes where conscious entities have the ability to make decisions and not know the outcome, and by extension those decisions result in universes where free will continues to exist. That doesn't mean free will is a stable state, nuclear war or the discovery of omniscience could eliminate it, and it's apparent that free will arose through evolution.
One really interesting result of modal realism is that ultimately no consciousness ever truly dies; it persists in some universe, however unlikely it may be. This causes some philosophical problems, because what is the likelihood that existing forever will be a pleasant experience? One line of reasoning leads me to believe that unpleasant universes kill off their inhabitants quicker than pleasant ones, thus leading to a much greater percentage of pleasant univer
We have most of the answers right in front of us, we're just afraid to ask the right questions. Reality exists because we're aware of it. We are creating the Universe by perceiving it and by choosing which Universe we want to be in. But as someone else in this discussion has said, how are we so different that our status as observers makes us so special? Our consciousness ultimately derives from the same particles that make up what we observe, we simply have a higher order of organization and "synchronicity" so that our consciousness, for a while, is greater than simply the sum of what we are made of. The only conclusion that can be obtained from that realization is that elementary particles are conscious and have some measure of free will. That is what quantum probabilities measure: the possibility the particle will "choose" the different possible paths it can take within the laws of the spatio-temporal Universe. It is a very basic kind of consciousness, as the perception it has of the rest of the Universe is extremely limited: its own physical characteristics, and what other particles it can interact with in exchanges of energy. As particles start bonding together and organizing, we get to higher and higher degrees of order, organization, and what I call "synchronicity of purpose" where eventually as in higher primates it can actually work together to achieve a higher level of consciousness and awareness of its environment, but thus creating a different level of reality which is certainly more powerful but not necessarily more complete or "true" than the simple interactions the single particle can achieve.
I come at the problem from a much more mathematical point of view, but ultimately I think your description is correct. For any given universe, there are rules that govern the behavior of everything in it. The thing is, for any set of rules there are an infinite number of universes that can exist within those rules. The important thing seems to be the rules that choose how any given universe will change. For me, consciousness exists as the ability for a system to model itself, e.g. something constructed within a system that mimics the entire system, rules and all. It doesn't have to be an exact model (which leads to Zeno like paradoxes), just a working model. In other words, if some part of a system is self similar to itself, that is the beginning of consciousness. When the self similar model can be manipulated by the same rules (encoded in the system) that govern the system itself to explore other possibilities for the configuration of the system, consciousness is complete. There is an extension to consciousness which is self action, which is partly separate from the model and has the ability to change the system itself based on interactions with the model. Self awareness occurs when the model includes a generalization of the self action itself, e.g. it knows that it is a model of the system with the ability to change the system. You can reverse the definition and say that the entire system is conscious because of its ability to change itself, but in precise terms the consciousness is limited to the model. If the model is destroyed, the system remains but consciousness is gone.
I'm also a modal realist in a strict mathematical sense. I believe that everything expressible with mathematics exists just as much as the universe, and probably the universe exists because of its expressibility in some form of mathematics. I also think that there are probably higher mathematical models than we can conceive of in this universe that also exist, but we don't have the resources to actually construct those models. Still, it's an interesting question whether set theory is sufficient to describe any possible model at any level, or if there needs to be something else bigger than set theory (and category theory) to describe something. My guess is that there is, but those are the kinds of things we won't be able to imagine in this universe.
I still think the Universe isn't a simulation, because some
Well, it all makes sense, if you think of it. Whoever is running this MMO we call RL, can't possibly have the resources to simulate every single particle all the time. So until someone actually goes and observes the damn thing, there's no need to actually spawn/instantiate it.
What if they do have the resources? In other universes, being able to hold and manipulate the set of reals may be just as natural as working with stones or atoms. Push a few sets of reals together to make a space, squish some reals out of another space to make a function, map the function to the space and see what comes out.
Or say you're hacking away at a copper ore vein with your trusty cold iron pickaxe, like a good dwarf. Sometimes you get just a piece of copper ore, sometimes you also get 1-2 pieces of stone, sometimes you get a Shadowgem, or a Tigerseye or Malachite. Were they already there before you started to hack at the ore vein? Or did they exist only as a probability until someone actually gets that loot window?
It all depends on how the random number generator gets seeded and when it gets called. If you repeated the exact same sequence of events from the beginning of WoW coming online, including all character actions at the right time, then the same things would happen. Even if Blizzard had a true random number generator, that just reduces the problem back to the case of our universe: If everything in the universe was reset to some point in time, what would cause a different outcome? With quantum mechanics, there is no way for us to talk about a single state that the universe has, just a very large set of states of possible universes that match all our measurements. This means that our past and future is also a nondeterministic set of possible universes. The question of whether to believe that the universe makes an infinite number of random decisions at each step to pick one and only one result or that all the possibilities exist simultaneously is basically the question between some form of determinism versus modal realism. I find the latter much easier to accept on a mathematical and philosophical basis, because it gets rid of most of the unanswerable philosophical questions altogether, such as what existence is, why it exists, and how. Modal realism just means that everything that's possible exists. The question of what is possible is interesting to think about, but basically boils down to anything that we can imagine, and probably more.
If you can afford it and the police have a reasonable response time, an alarm monitoring company is probably the best solution. Cutting wires won't help because disconnecting the house from the monitoring company triggers the alarm on their end (and maybe at the house, too) It would be interesting to know what infrastructure the alarm company uses for their wires. I imagine they could lease lines from the phone company, but for video it would tend to be rather expensive. Maybe they borrow frequencies on the cable system or actually did go lay their own wiring in the neighborhood. If it's an upscale neighborhood where everyone has an alarm, I would guess the latter. Wire all the neighborhood into one box and then pay for the T1 or other connection back to the main office and split the costs. Is the video from the cameras high quality, or small and/or highly compressed?
Have you looked into running your generator during peak usage hours to reduce electricity costs? It would be interesting if natural gas ended up being cheaper than buying the electricity during that time. It sounds like the generator powers the entire house back through the existing wiring, which may mean that it's overloaded if you see voltage and frequency variations that high. Generally it should run right at 60 Hz unless it's under heavy load and can't keep up. I'm surprised that your UPS didn't like the voltage fluctuations since they're supposed to handle it as far as I know. They will definitely switch to running off the battery when the voltage or frequency goes out of whack, but should keep the battery charged from the wall power regardless.
The Door in the kitchen coming from the garage is controlled by a set of really strong magnets and and hooked through the security system too. Once it is locked, you need about as much force necessary to kick a regularly locked door in to open it.
Too bad most thieves cut the power before trying to break in. You do have your cameras on a UPS and the video sent to a remote location in realtime, right?
QNX has the advantage that I/O, like almost everything else in QNX, is done via inter-process message passing operations. The message passing system uses priority queues, and so requests to file systems and devices get handled in priority order. So resource managers (file systems, device drivers, etc.) don't have to explicitly handle priorities; it's done for them. Some resource managers, like disk handlers, process multiple requests at a time so they can reorder them to optimize access, but network devices and such are FIFO at the resource manager level and priority ordered at the message level.
Yeah, and then a priority inversion killed the Mars Rover (I think that's what it was). With standard scheduling, things would have just slowed down, but wouldn't have hung the whole system.
1. They better both pay equal amounts, or the neutrality is somewhat in question. The court should get to appoint an expert from a list of candidates submitted by both parties, to make it a little more impartial.
2. With MD5 hashes of everything and a redundant, untouched copy of the disk, it shouldn't matter to the plaintiffs who deletes the files. Hopefully, the respondent can get another computer expert to help out with their lawyer present and go through a list of files, including caches and swap files, that should be deleted as privileged data. Since this is an RIAA copyright case, every file except the ones in the P2P directories should be privileged, although I don't know how that argument would stand up in court. I assume that the plaintiffs could challenge the extent of deleted privileged data, but they should need a good reason to be looking at anything but the P2P software, it's directories and logs, the registry (assuming it's Windows), and reasonable parts of the OS logs (limited to the dates in question, or thereabouts).
3. Do they have the option of making a third copy of the disk, omitting the sectors that the privileged files resided on? That's far more secure and prevents the plaintiffs from attempting expensive data recovery on the drive with "deleted" files. What the ruling really should have specified was a list of MD5 hashes of each filesystem block, and then a hash of the entire list. That would ensure that when files were being deleted, only the blocks they were on would no longer match the hash, and the rest of the files could be verified as being from the original drive. Maybe they did do that, I don't know the details.
Smartfilter is an incompetent bunch of fucks too, with a heap of garbage for management software. We constantly have to get sites "recategorized" as you put it, because their censors are too lazy to do their job, and their software makes per-page exception lists impossible.
Does anyone else appreciate the extreme irony of going to protest in Cuba
Does anyone else appreciate the irony of having a U.S. military prison in Cuba? Wait, maybe that's not irony...
Real web browsers handle scripting languages with fewer problems than Office. So does the Java virtual machine. Office is just poorly written without any real security or privilege separation in mind.
.NET, but my guess is it's a vastly different (in terms of technical competence) team than the one working on Office.
Due to their extensibility, pdf and postscript are suspect in the eyes of the truly paranoid, let alone the complex modern formats.
Postscript is a Turing complete programming language, but I haven't heard of many exploits in ghostscript or PS printers. Clearly there's just something broken at Microsoft. It's possible to build secure interpreters for general scripting languages, and Microsoft just hasn't been able to accomplish it with Office. They managed to write
McDonald's causes great hassle for Burger King as they refuse to release the recipe for the Big Mac's secret sauce. Sadly, this will only be available at McDonald's for the time being. There are no plans for cross-restaurant release.
In other news, the bread and burgers still come from wheat and cows (mostly). My point? The content is what people want, not the format. People couldn't care less what format their content is in as long as it works. Microsoft is trying to ensure that only the Microsoft formats that come with Windows will work, and monopolize the content. It would be like McDonalds buying all the cows in the world.
Much as I love Linux and free software, it is self-defeating and unrealistic to demand that Microsoft (and other companies) support Linux. Perhaps the much-vaunted free software community should produce its own solutions that are better then the closed-source competition? Instead of complaining about what other people do, take responsibility for your own needs and write the software you want.
As non-windows OS's become more popular (remember, there are billions of cell phones and millions of PDAs, it's not just Linux and Mac), content producers are going to need cross platform and open standards. Microsoft is not giving their customers the tools to reach all potential viewers, therefore Microsoft is failing in its job as a software maker.
Adobe(Macromedia) Flash has been around for a LONG time and I have yet to see anyone attempt to come up with a serious Open Source alternative.
As a user of open source, I generally detest flash. It's only good for making annoying, flashy, moving advertisements and web pages. I hate flash movies, because there's no reason not to just provide an AVI or MP4 stream. I hate flash menus, because if they're broken it's impossible to view the source and maybe have a chance of figuring out where they were trying to send you. Everyone hates flash ads. For interactive web pages, there's cross platform AJAX that you can actually see the source for, and Java for applications that just don't work in AJAX.
Is the router authorized to let you connect? I don't think so. It can connect, but it has no authority. It is like a neighbor's electric plug point. It can supply electricity, but that doesn't mean you can use it without neighbor's permission (plug point's permission is not enough).
It would be permissible if my neighbor always left his extension cord plugged into his house and the other end laying in my yard or coming in through my window. Before anyone even sends a DHCP request, they have to first receive the SSID broadcast from the wireless router. It is actively advertising its presence and willingness to accept and respond to unencrypted packets.
It's not like the bad guys have devil horns and the civilians carrying weapons have halos so you know what to shoot at.
The good guys aren't pumping bullets into unarmed civilians. These are my best attempt at some rules of engagement for armed citizens:
1. Don't shoot first.
2. Only shoot at an attacker if they have just obviously and directly injured an unarmed person.
3. If anyone shoots at you, only shoot back after you have identified yourself as a good guy and they keep shooting.
An attacker does not necessarily need to be shot to stop an attack, just convinced that it's futile to shoot anyone else. There will be an uneasy standoff until witnesses can identify who the real attacker is, but in general reaching a stalemate or physically stopping all the attackers should occur reasonably quickly, with a minimal number of shots fired. Feel free to find weaknesses in the rules.
If there are too many attackers, then the situation is basically hopeless, especially if they are well coordinated. An obvious example is a special ops team that can infiltrate a military installation in spite of numerical disadvantages. The biggest problem the armed civilians have is the identification of each other and attackers, but in an open area this usually becomes apparent when the unarmed civilians run and hide, the attackers keep shooting them, and the armed civilians are holding guns and looking for attackers. The worst case for the armed civilians is when the attackers return fire, reducing the available information about who the attackers and armed civilians are. That's why a heavy emphasis on rule number 1 is important. The more people who are armed and *not* shooting, the easier it is to identify the clique of armed citizens. The unarmed civilians will ultimately aid in the attacker identification process.
It almost makes me think mandatory police (or military) training would be a good idea.
Hmm, didn't stop me. Just give mplayer or mencoder the -ss option. I think this one needed about 150 seconds, longer than most of the other Sony DRM failures and probably the reason it doesn't work on some players. Normally it's only about 30 seconds into the movie that they put the bad sectors.
This copy protection prevents most people from renting/borrowing a DVD and making a copy of it. Until people download the latest software for cracking it. This is mostly targetting non-technical people who were given DVD Shrink by a friend.
If this is the typical "bad sectors at the beginning part with the lady holding a torch" DRM that Sony is so fond of, 'mencoder -ss 60' usually takes care of it. Just adjust the time to match the opening credits, and rip away. Mplayer accepts the same option if you just want to watch it.
IINM, it IS the discs - with Sony's ARccOS protection.
Is that what all this is about? 'mencoder -ss 60' takes care of that. Skips all the nasty Sony logos, too.
I would find it difficult to characterize checksums or parity as derivatives. More likely they're either irrelevant (checksums) or just part of the overall reproduction (parity; see the videogame cases from the early 80's).
If a portion of the work can be recovered from the parity, doesn't that make it a derivative work? For instance, CDs using CIRC carriy 1 byte of parity for every 3 bytes of data, That means if the data itself could be split from the parity, there would be one perfect copy and one degraded copy (approximately 8 bit mono, versus the 16 bit stereo of the CD). If the parity was increased, the ratio of data bytes to parity bytes could be lowered further to produce increasingly accurate representations of the data. After there are more parity bytes than data bytes, a subset of the parity bytes can be used to recover the entire data. Another interesting possibility is secret sharing where data is split into multiple pieces of which some number are required to reassemble the original work, and without a sufficient number no information can be obtained. Would it be illegal to split a copyrighted work into three pieces, of which any two were required for reconstruction of the work, and then share the pieces individually? No portion of the actual work would actually exist in any of the pieces, just a mathematical relationship that allowed the work to be reconstructed from at least two pieces. See Shamir's secret sharing algorithm for instance.
In the past, when someone has actually bothered to raise the issue, the courts have found it to infringe. But usually, who cares? Since statutory damages are calculated per work, not per infringement, and since compensatory damages for the incidental copies would be overvalued at a nominal sum of $1, no one cares about them. The point is to get all the direct infringers you can, and from there to get the indirect infringers. This used to be of some interest when there were ISPs who were intermediaries and who had deep pockets, but they usually fall under the safe harbor now.
Every RIAA lawsuit I've heard of has targeted uploaders, perhaps because it's the only way they can find filesharers and still have a case, so it certainly seems that creating a copy and sending it over the network is considered infringement. The cases I've looked at claim both reproduction and distribution infringement. Looking at RTC v Netcom I found MAI SYSTEMS CORP. v. PEAK COMPUTER, INC., 991 F.2d 511, which ruled that simply loading software from permanent storage into RAM constituted the creation of a physical copy, which is apparently the same argument used to enforce EULAs.
You'll find that it is more nuanced than that. Again, the 512 safe harbor has really taken over here, but for an earlier case dealing with these issues, I strongly suggest you read Religious Technology Center v. Netcom. Of course, the reason we have the safe harbor is because not every court followed this particular logic. The uncertainty was not good for the Internet.
As far as I can tell, the difference is that service providers do not necessarily know the legal status of the content they allow users to store and transfer. On the other hand, someone running P2P software will usually know which files are being shared and their legal status, although this is not always the case. As the Napster case showed, even direct knowledge that a service is being used for illegal activities is enough to warrant contributory copyright infringement.
When Alice has a file on a server and Bob downloads it, what happens is that Bob cause his computer to request information from Alice's computer. Alice's computer responds and sends information to Bob. But it does not send a copy because hard drives do not travel over Ethernet. When Bob's computer receives that information, it writes it to various storage media, such as RAM, or hard disks. By doing that, it creates a new copy, namely, that RAM, or those hard disks.
If you really want to boil it down, when Bob requests information from Alice, Alice has to read the data from her hard disk (which has in addition to the original work, a derivative work in the form of error correcting codes), the hard disk stores a copy of the original work in its cache, the drive copies the information onto the data cable, the controller copies it into internal registers and then copies it onto the system bus when appropriate (either via DMA or when the CPU issues in IO read instruction). For DMA transfers, the information is copied directly into system RAM, for direct IO the CPU copies the information off the bus into internal registers and the cache, then copies it into internal RAM. To actually send the data to Bob, the CPU then copies the data from memory (requiring a read into the CPU and cache for each byte/word of the data) into a second copy in a network buffer (some OS's use scatter-gather to keep the data in one memory buffer and skip this step), then the CPU has to copy the data back over the IO bus to the network card/modem which requires more loads, caching, and stores. Finally, the network card or modem copies the data onto whatever network it's connected to.
In this relatively simple example, there are at least 8 copies made by Alice's computer. That doesn't even include the case where intermediate P2P software accesses the data from the disk, copies it and hashes it, and does whatever else before finally sending it to the network stack. It's inconsistent for the court to rule that infringement of reproduction rights does not occur when Alice makes her copies, but does when Bob makes his copy to his hard drive, because EULAs have been ruled valid on essentially the same claim; that Alice has to make internal copies of software she has purchased in order to use it.
And who caused this to happen, who bears the responsibility? That would be the downloader. Really, the only situation in which the downloader could convincingly argue that it wasn't his volitional act that caused the infringement would be if someone else had control of his computer and was making it do that.
If Bob asks Alice to violate a law, and she violates it, she is fully responsible for her actions unless she was coerced. Bob might have accessory charges or be charged with contributory infringement, but it wouldn't affect Alice's culpability.
Discussing whether or not to allow ED-209 to shoot at humans is silly. ED-209 can't tell the difference in the first place. If we allow him any targetting discretion at all, he's going to shoot at just about everything. That's the reality imposed by the technology and we're not on the verge of overcoming it.
That's not true. Current computer vision systems can easily distinguish between most people and most vehicles. People are tall with dangly appendages that swing around, vehicles are wide and low and don't have so many moving parts. You could literally hook a computer vision system up to an autocannon and stick it on a Darpa Grand Challenge vehicle to create a search and destroy killbot.
Is it just me or is a discussion of ethics laws for robots premature given the state of the art in artificial intelligence? If you want to teach a machine not to harm humans, it helps to first teach the machine the difference between a human and every other object it encounters.
The discussion is not premature. Land minds already kill humans, machines indiscriminately, and therefore would be in violation of this treaty. The interesting thing is that guided cluster bombs could violate the law if they were not built to only target heat sources large enough to be vehicles. I have no idea of the mis-targeting rate of guided cluster bomblets, but my guess is that the military doesn't mind if the soldiers standing next to the vehicles get targeted too...
Other automatic targeting systems will be developed soon. At some point the military will want to run borders autonomously instead of wasting troops to stand around all day. Autocannons are the logical choice, and those weapons will have to decide whether they are pointed at children, soldiers, animals, ambulances, or tanks. I don't think anyone wants to return to a no-mans-land style of warfare, but indiscriminate killing machines would lead to that.
I assume by "modern cyphers" you mean public key cryptography with large keys.
I was talking about symmetric ciphers like AES, Twofish, Cast5, and even 3des is still considered secure. Asymmetric (public/private key) ciphers rely on a symmetric cipher to do bulk encryption of data with a random key, and the random key is encrypted with the asymmetric cipher.
if I want to keep a secret, I'm going to assume those will be breakable in a few years or less using quantum computing or something else.
Quantum computing is not expected to be able to improve attacks on symmetric key ciphers to less than sqrt(n) work, and that only with massive amounts of memory to search the keyspace. In that case, AES with 256 bit keys is still as secure with all powerful quantum computers as AES 128 is today.
Additionally, there are no known quantum algorithms for quickly solving the discrete log problem (DLP) for elliptic curves, which means that elliptic curve cryptography will probably provide secure asymmetric ciphers once quantum computers are powerful enough to attack ciphers relying on standard DLP and factoring of integers.
Therefore, if my message must remain secret for more than a few years, I am going to do something to make it impossible to know when the correct key is guessed. Splitting the message itself is one way to do this.
You're relying on being able to keep separate copies of your data secure from coming together somehow. If you can do that, why can't you just keep one copy physically secure and forget the encryption? E.g. if you're relying on keeping a one time pad secret for several years, couldn't you just store your data in place of the one time pad? One of the basic assumptions of modern cryptography is that the entire ciphertext is always available to the attacker, along with complete knowledge of the ciphers and protocols used to encrypt the plaintext. The only things the attacker is assumed to lack are the keys themselves. Weaker assumptions lead to either weaker cryptosystems or complex ad-hoc mixtures of good and useless cryptography, making the entire system more difficult to analyze for flaws.
If you do really need perfect security, then skip the symmetric and asymmetric ciphers entirely and rely on secret sharing to distribute the pieces of your data to secure locations. The wikipedia link I posted has enough information to pick a scheme and implement it or find some software that will do it. Just be aware that anyone who needs your perfectly secure data is probably capable enough to follow you everywhere you go, and is already doing so if you have secrets that important, making multiple "secret" locations impossible. You would probably be safer to never write the secrets down in any form.
Security is always a trade-off. If you don't need perfect security and don't have spooks watching your every move, then modern ciphers are more than adequate for security and much more efficient and easy to use. Adding complexity to an already working system just complicates verification and makes it possible to leak information or make mistakes that can be used to break the system.