Actually, I believe the transatlantic cable has indeed been cut by accident at least once in the past decade. Even so, we're talking very different incident rates here. Anyways, as for who is responsible, half the planet has motives and the other half have motives for making the first half look guilty of something. This makes "Murder on the Orient Express" look like a tale of lilly-white innocents. It could be any of them. It could even be all of them. The earliest we're likely to know for sure is in 50 years time when the UK declassifies this year's goings-on - as part of SIGINT, they will know what is happening, one way or another.
Google uses a basic citation index, but as far as I know doesn't consider references, multiple generations of citation, references or citations, citations of references, duplication of citations/references (mirrors should not weigh as much as originals), credibility of sources (not sure how you'd measure that one) or proximity to known good results (the user could flag good results, which could then be mined by a search engine to improve the search terms). This method, basically an adaptation of how academics look things up, is tried-and-tested but may still not be useful on the web where relationships between ideas can be tenuous or obscure.
Most search engines also (these days) spell-check terms, look up singular/plural forms and perform other trivial operations. Having it use a thesaurus for alternative meanings might be a good idea, too. This would produce more results, yes, but if you then applied stricter filtering on those results, and stricter weightings for sorting what's left, you should get about the same number of results in total with a higher percentage that are actually useful.
I can't help but remember a different, much older, search engine, though. CAS Online. This stored abstracts on chemistry papers. Virtually all of them that had ever been published. You performed not one search but many, where a search could be new or on the results from one or more previous searches, using set logic and a very primitive SQL-like query language. Because you could build on previous results, you always ended up with only a few results, almost all of which were highly relevent to what you wanted. However, as with attempts to use regular expressions in web searches, what is used (and useful) elsewhere may not be so useful or practical for Joe Average when searching the web.
Otherwise, how will you conduct evacuations, correct containment procedures, etc? Emergency service personnel are massively underpaid and under-equipt, sometimes under-trained as well, and usually suffering from mental disorders or addictions, making them more than a little vulnerable. Anyone who has been to a security briefing knows these are the very people you're advised to watch out for as the greatest potential security risks. So, either massive population centres are in extreme danger from emergency services not being suitably aware, OR massive population centres are in extreme danger from emergency services being aware.
Seems to me that the two cases would have equal consequences and equal risk levels, and that no other individual could possibly modify those values significantly, reducing the security through obscurity to someone's job security through obscurity. Tell me, why should I care about this person's job more than I care about any potential risk to my wellbeing?
...was the discovery of the large planetoid larger than Pluto and also outside of Pluto's orbit that was discovered by an American team and then rediscovered by a European team based on information they'd obtained from the first lot. I imagine it's commonplace amongst astronomers, due to the timescales involved in verifying findings and the difficulty of proving plagarism when dealing with objects visible from half the Earth's surface for extremely long periods of time. It's also common in mathematics - Sir Isaac Newton stole copiously from Huygens, Descartes, Hooke, and anyone else stupid enough to let him. Or perhaps not stupid - the only person to resist Newton's claim of ownership did die rather soon after.
Technology is another area with a dubious history. Edison was rather notorious for "inventing" other people's inventions, which is a slight variant form of plagarism. Countries, as well as individuals, have been suspected (or proven guilty) of conducting industrial espionage in order to beat someone else to the goal of being first.
In other words, it happens. A lot. The acclaim and fortune that goes with being first is too alluring for some to refuse. Some don't bother to steal, they just make it up. Some in the hope they can get the "right" results later, others in the hope that nobody notices until they're rich and elsewhere. I'd place the professor of cloning from South Korea in the first category, simply because he could have left when suspicions were first raised, but didn't. I think he genuinely thought he could make a real breakthrough first and that everyone would then forgive him for past misdeeds. On the other hand, the cold fusion guys from Utah were good enough chemists to know that you can't perform fusion through elecrolosys. Cold fusion might be possible, but if all you needed was an anode and cathode, the first potato clock ever made would have ended up rather more than baked.
It would be good if there was some sort of independent international auditing body that examined initial claims and then revisited that claim after so many years, again after the claimant's death, and also at the 50 year and 100 year marks (as those are when papers held as secret by Governments are usually declassified automatically), where that body had power to reassign credit and possibly award some percent of past earnings to newly-recognized discoverers/inventors. It still wouldn't stop fraud, but some redress is better than a one-line entry in a textbook nobody will ever read.
Existing LANs are faster than existing WANs. Ethernet runs at 10 Gbps and can be channel-bonded up to 100 Gbps. Infiniband runs within the same range. What's more, long-distance Ethernet and Infiniband are possible today, making it far more economic and far more practical to replace WANs with one gigantic LAN than the other way round. There simply isn't any sense in running a T3 or T4 down residential streets, when Japan already provides gigabit ethernet links to each home.
I guess you are correct, but then we get back to the problem of why this option doesn't currently exist for consumers. Hang on, I think my brain is going to explode.
The problems with traditional microkernels lie in the heaviness of the module-to-module communication and in the number of context switches. An exokernel is pretty much entirely in one context, and exopc seemed to have very efficient communication, so that design looked extremely good. (Although a fair comparison isn't possible, a crude one would be to compare Exopc + the Cheetha web server with Linux + Tux, both serving static content. See how well they scale, when stress-tested.)
Exokernels aren't the only microkernels of interest, though. There have been efforts to produce mobile nanokernels, on the theory that drivers are generally smaller than data, so in a cluster, moving the code to the data should be more efficient on resources. The opposite extreme has been to produce kernels that span multiple systems, producing a single virtual machine. Here, kernelspace and userspace are segmented and the latency between machines is simply another sort of context switch delay, yet the overall performance is greater than a loosely-coupled cluster could ever produce.
Microkernels have a lot of potential, a lot of problems have been solved, there are still problems that need to be solved better. eg: if a driver crashes, there needs to be a transaction log that permits the hardware to be returned to a valid state if at all possible, or rebooted then rolled into the last valid state. This isn't just a software problem, it's a hardware problem as well. Being able to safely reboot individual components on a motherboard in total isolation requires more than fancy coding and software switches. You need a lot more smoothing circuits and capacitors to ensure that a reboot has electrically no impact - not so much as a flicker - on anything else.
Where microkernels would truly be "at home" would be in machines that support processor-in-memory architecture. Absurdly common function calls, instead of going to the CPU, having instructions and data fetched, and then being executed, along a long path from the OS' entry point to some outer segment of code, can be embedded in the RAM itself. Zero overhead, or damn near. It violates the principle of single-entry, single-exit, but if you don't need such a design, then why waste the cycles to support it?
I agree that any person should be capable of understanding and auditing any and all stages in an election. For paper ballots, I'd not go with pencils and scrap paper, though. You really want archive-grade ink or a medieval iron-oxide ink, and archive-grade paper. They'll last easily for more than a century, which also means they will be resistant to tampering. The use of carbon paper to make a duplicate which must traverse an independent path under the watch of whoever is in opposition to those running the voting station would improve reliability. A missing ballot box has much less impact if a carbon copy of the contents has already been delivered.
Electronic voting would be hard, you are correct, and it would require all - or essentially all - adults to have a solid understanding of higher maths. Like I said, some improvements in education would be required. Most people hate maths because it has either been badly presented or the lecturer was crap. I've taught 11-year-olds and 12-year-olds, and I can tell you that there was not one amongst them who was scared of maths. Neither boys nor girls. To them, it was a game, something fun, something that may have been in a school but was still entertainment. Now look at typical teachers in subsequent grades. Dour, dull as ditch water, unimaginative, cynical, determined to find discipline problems where none would otherwise exist, pressuring students to do well in exams and to hell with understanding the subject, and utterly power-mad. Sure, there are exceptions, but were they actually any better at teaching? It doesn't matter how different a teacher is, if they can't do better than slugs on acid. But if that issue was fixed? Then, hell, yeah, I'd be willing to bet that the vast majority of Americans could audit an encrypted, authenticated and verified e-voting system. Maybe not a 100% still, but more than could audit the existing system.
Customers are almost certain never to get IP Multicast, but (probably) not for technological reasons. It's easy to bill per stream, for unicast streams, but harder for multicast. And, let's face it, there are certain segments of the entertainment industry - not just the *AA's - that have a vested interest in providing heavily metered audio/video streams. Multicasting has the potential to slash revenue by an order or two of magnitude. It's also easier to guague interest (for advertising reasons) for unicast connections than for multicast. And since unicast demands more on the CPU and on the pipe, machine manufacturers and ISPs have financial incentives to encourage customers to use the least-efficient delivery format possible.
If the customers are the only ones who could gain, and everyone else would lose, then who is going to be insane enough to switch on multicast routing to the home?
Part of the problem is that people confuse the concept (e-voting) with the implementation (sending unsecured e-mails, using Diebold ATM machines, etc). The two are distinct. The concept can - in principle - be implemented as well as, or better than, alternatives. But only the implementation can be evaluated. The concept is nebulous and has no specific meaning or quality.
Another part of the problem is that most existing implementations are - frankly - crap. They offer minimal security, have frequently been reported as having errors such as non-zero counts, have poor reliability, provide minimal accountability and often provide no means of verification. This is wholly unacceptable. Nobody would accept that from a cash register in a supermarket, never mind a system that is mission-critical in a democracy.
Hand-counts can be reliable. For the longest time, the British system was entirely done by hand-counting, with very small error rates for a population of 60 million. The American system includes machine counts, statistical sampling, and other mechanisms for speeding up the returns, with different States using different methods. It is also worrying that the first returns are announced prior to the polls closing on the west coast, which will inevitably introduce bias and strategic voting. The British system isn't perfect, and has recently developed all kinds of flaws and fraudulant practices, but it can be used as a yardstick of what a democracy should minimally achieve.
Of course, a democracy has other dependencies. It's only meaningful if enough of the population votes for the votes to truly represent the population. The electoral college has the potential for distorting the consensus of the people and probably has. There is no ballot option to reject all candidates and re-open nominations. Media saturation and candidate funding warp awareness. The educational system isn't up to the standards needed to ensure the population have the breadth or depth of knowledge to understand the complexities of a nation or avoid the wiles of a skilled talker. If these flaws remain, then even a perfect voting system can never represent what the public actually want or need, which is what a democracy is about. Being heard has no meaning if you never learned how to talk.
To me, the question shouldn't merely be how we reliably count votes, but should also include how we reliably cast them. There may be no better solution than the one we have, I accept that, but I won't accept that this is known until it actually is.
If you're using MPLS on a WAN, I'm scared. MPLS is for extranets, the WAN is concealed. MPLS on a WAN is therefore a contradiction in terms. Either you're dealing with a WAN or you are dealing with an extranet. You are never dealing with both at the same time. You are almost never dealing with MPLS at the enterprise-level, because that level of detail is normally hidden. You have an entry point onto the extranet, but how that extranet is formed is transparent.
Of course, none of this matters if you're not using extranets. Any Enterprise-level network that exists in one physical location or connects to other locations over the public Internet or connects to other locations via leased physical lines has no need of MPLS or any other virtual circuit technology. Metropoliton networks, such as GMING, tend to use leased physical lines and point-to-point protocols such as ATM. As do most of the smaller-scale backbones (such as national DSL providers) who bought up dark fibre in bulk. If you're using an ATM network, you don't waste time with IP-based virtual circuits.
I've admined my share of b2b, national and international Enterprise networks. I co-founded the IPv6 backbone in the UK. I've run large-scale X.25, IP and ATM networks. Before that, I designed my own LAN and WAN wire protocols. I'm no Vint Cerf, but let's be realistic here. Anyone can make a claim, but it takes more than that to make a difference.
As for your multicasting commentless comment, the announcement of the dissolution of the MBONE in the mid 90s is quite sufficient as a reply. It's you against the world, and the world declared many years ago that you lost. That war is over and the multicasters won. Give it up. Accept defeat gracefully. The only serious resistance left are the ISPs for SOHOs and individuals and the increasing number of multicasting applications in Windows mean they will submit soon enough. Even mobile phone companies, such as Nokia provide their services by multicast. They have surrendered to the only rational networking technology for distributed services. The Infiniband Consortium and the Open Fabrics Consortium depend heavily on the multicast capabilities of modern technologies. Resistance is futile, you will be IGMPv3'ed.
As did I. Of course, neither of us is in a position to ever really know what happened, all we can ever go by is the limited (and likely editorialized) information made available, which is why it's so important to add all the qualifiers. If there was an error made in this thread, it was on my part for not putting in enough qualifiers in the first place.
Bits of the pedestrian gave way. The driver was cited by police as doing 20+ MPH over the speed limit, although the news story cites (unnamed) independent experts as saying it would have been closer to 40 MPH over the speed limit. (The speed limit was 55 MPH). Far as I can tell from the news article, the driver is not claiming that there was shared responsibility for the accident or that the pedestrian did anything wrong, merely that the pedestrian caused damage to his Audi. If that really is the whole story, then the attitude is no different from the RIAA's (bringing this thread back on topic).
Well, it depends on what exactly was optimized, but this is usually an iterative process where the next step is determined by herustics. So, yes, genetic algorithms would be fine for this. You keep going until whatever variable(s) you're optimizing hit a local minimum (the point at which herustics give up and decide all options will de-optimize the solution). You can either stop there or try another starting point to see if it produces a "better" result. If you keep going, then you've some stopping condition (eg: N successive runs in which no better result was obtained, or you have calculated more than some percent of the minima that system of equations would potentially allow for). If you've not calculated all potential minima, then ultimately "recognition" of correct results will be down to a gut instinct.
This assumes that the system can't actually be solved or reduced/simplified to one that can without losing too much accuracy. If you can solve it, then all you have is some basic matrix algebra on an 8000x8000 array. Non-trivial, sure, but 4Gb of RAM and a good gaming machine (you want fast maths) would be adequate to crunch such data. Alternatively, an analogue computer would be ideal for a problem like this, as you'd have far greater precision and far greater parallelization. It would also take far more space and cost far more, but the world economy could do with a boost about now.
Funny you should talk about killing. There was a news story recently about a driver who hit and killed a pedestrian, then sued the family for damages to his car. I think we should check to see if he's on the RIAA's advisory panel.
Maybe he does. I've not seen 3Com for a while and Bay went belly-up. If he was the chief designer for those two, it would explain what happened to them.
Any ISP that uses RIPv2, OSPFv3 or ISIS on their internal network - or to connect to other networks - uses multicast for the routing protocol. Core providers broadcasting multicast services that are visible to AmericaFree.tv are listed in table format. It's longer than two entries. Translations of BGP entries to providers are . Local ISPs that actually provide multicast to the home are a rarity, but as of 2002, there are a handful.
I never said anything about nuclear war, I specified fault tolerence. There's a big difference. Several tens of megatonnes of difference. Providers offering virtual circuits for extranets sometimes use MPLS, but not all. Service guarantees are over time and can be provided by packet marking and QoS. Isolation likewise. When providing full mesh services, which is what many corporate ISPs are moving towards, something like MPLS is a disadvantage. Containment and QoS is only safe at endpoints, and running anything heavier than you need on a full mesh (which itself is insanely hard on routers) is asking for trouble.
You only need an address for each group address subscribed to by a downstream node. Since you have access to port numbers, you can place as many streams on a single address as you like (up to 65535), although obviously you lose some benefit from the multicasting if you overload too many streams onto a single group address. Well, unless you use source-specific multicast (SSM), in which case so long as the content is differentiated by source address, you can stuff everything onto a single group if you really want.
IPv6-over-IPv6 seems to work ok. Some of the earliest routing protocols provided firewalling and NATting within the routing protocol itself (Telebit's router provided superb NAT and Firewall capabilities as an integrated facility). Permanent addresses lead to fragmented heirarchies and exploding routing tables, which is a major problem with IPv4.
Slashdot Mirror
Eel urn.
Some comments simply deserve a special "otherwise unachievable" rating. The parent post would be one of them.
Actually, I believe the transatlantic cable has indeed been cut by accident at least once in the past decade. Even so, we're talking very different incident rates here. Anyways, as for who is responsible, half the planet has motives and the other half have motives for making the first half look guilty of something. This makes "Murder on the Orient Express" look like a tale of lilly-white innocents. It could be any of them. It could even be all of them. The earliest we're likely to know for sure is in 50 years time when the UK declassifies this year's goings-on - as part of SIGINT, they will know what is happening, one way or another.
That, and the Egyptian captains were issued whetstones by the US navy for anchor-sharpening practice.
Ironyblind just means you can't see Irn Bru soda.
Google uses a basic citation index, but as far as I know doesn't consider references, multiple generations of citation, references or citations, citations of references, duplication of citations/references (mirrors should not weigh as much as originals), credibility of sources (not sure how you'd measure that one) or proximity to known good results (the user could flag good results, which could then be mined by a search engine to improve the search terms). This method, basically an adaptation of how academics look things up, is tried-and-tested but may still not be useful on the web where relationships between ideas can be tenuous or obscure.
Most search engines also (these days) spell-check terms, look up singular/plural forms and perform other trivial operations. Having it use a thesaurus for alternative meanings might be a good idea, too. This would produce more results, yes, but if you then applied stricter filtering on those results, and stricter weightings for sorting what's left, you should get about the same number of results in total with a higher percentage that are actually useful.
I can't help but remember a different, much older, search engine, though. CAS Online. This stored abstracts on chemistry papers. Virtually all of them that had ever been published. You performed not one search but many, where a search could be new or on the results from one or more previous searches, using set logic and a very primitive SQL-like query language. Because you could build on previous results, you always ended up with only a few results, almost all of which were highly relevent to what you wanted. However, as with attempts to use regular expressions in web searches, what is used (and useful) elsewhere may not be so useful or practical for Joe Average when searching the web.
Seems to me that the two cases would have equal consequences and equal risk levels, and that no other individual could possibly modify those values significantly, reducing the security through obscurity to someone's job security through obscurity. Tell me, why should I care about this person's job more than I care about any potential risk to my wellbeing?
Technology is another area with a dubious history. Edison was rather notorious for "inventing" other people's inventions, which is a slight variant form of plagarism. Countries, as well as individuals, have been suspected (or proven guilty) of conducting industrial espionage in order to beat someone else to the goal of being first.
In other words, it happens. A lot. The acclaim and fortune that goes with being first is too alluring for some to refuse. Some don't bother to steal, they just make it up. Some in the hope they can get the "right" results later, others in the hope that nobody notices until they're rich and elsewhere. I'd place the professor of cloning from South Korea in the first category, simply because he could have left when suspicions were first raised, but didn't. I think he genuinely thought he could make a real breakthrough first and that everyone would then forgive him for past misdeeds. On the other hand, the cold fusion guys from Utah were good enough chemists to know that you can't perform fusion through elecrolosys. Cold fusion might be possible, but if all you needed was an anode and cathode, the first potato clock ever made would have ended up rather more than baked.
It would be good if there was some sort of independent international auditing body that examined initial claims and then revisited that claim after so many years, again after the claimant's death, and also at the 50 year and 100 year marks (as those are when papers held as secret by Governments are usually declassified automatically), where that body had power to reassign credit and possibly award some percent of past earnings to newly-recognized discoverers/inventors. It still wouldn't stop fraud, but some redress is better than a one-line entry in a textbook nobody will ever read.
Existing LANs are faster than existing WANs. Ethernet runs at 10 Gbps and can be channel-bonded up to 100 Gbps. Infiniband runs within the same range. What's more, long-distance Ethernet and Infiniband are possible today, making it far more economic and far more practical to replace WANs with one gigantic LAN than the other way round. There simply isn't any sense in running a T3 or T4 down residential streets, when Japan already provides gigabit ethernet links to each home.
I guess you are correct, but then we get back to the problem of why this option doesn't currently exist for consumers. Hang on, I think my brain is going to explode.
Exokernels aren't the only microkernels of interest, though. There have been efforts to produce mobile nanokernels, on the theory that drivers are generally smaller than data, so in a cluster, moving the code to the data should be more efficient on resources. The opposite extreme has been to produce kernels that span multiple systems, producing a single virtual machine. Here, kernelspace and userspace are segmented and the latency between machines is simply another sort of context switch delay, yet the overall performance is greater than a loosely-coupled cluster could ever produce.
Microkernels have a lot of potential, a lot of problems have been solved, there are still problems that need to be solved better. eg: if a driver crashes, there needs to be a transaction log that permits the hardware to be returned to a valid state if at all possible, or rebooted then rolled into the last valid state. This isn't just a software problem, it's a hardware problem as well. Being able to safely reboot individual components on a motherboard in total isolation requires more than fancy coding and software switches. You need a lot more smoothing circuits and capacitors to ensure that a reboot has electrically no impact - not so much as a flicker - on anything else.
Where microkernels would truly be "at home" would be in machines that support processor-in-memory architecture. Absurdly common function calls, instead of going to the CPU, having instructions and data fetched, and then being executed, along a long path from the OS' entry point to some outer segment of code, can be embedded in the RAM itself. Zero overhead, or damn near. It violates the principle of single-entry, single-exit, but if you don't need such a design, then why waste the cycles to support it?
Electronic voting would be hard, you are correct, and it would require all - or essentially all - adults to have a solid understanding of higher maths. Like I said, some improvements in education would be required. Most people hate maths because it has either been badly presented or the lecturer was crap. I've taught 11-year-olds and 12-year-olds, and I can tell you that there was not one amongst them who was scared of maths. Neither boys nor girls. To them, it was a game, something fun, something that may have been in a school but was still entertainment. Now look at typical teachers in subsequent grades. Dour, dull as ditch water, unimaginative, cynical, determined to find discipline problems where none would otherwise exist, pressuring students to do well in exams and to hell with understanding the subject, and utterly power-mad. Sure, there are exceptions, but were they actually any better at teaching? It doesn't matter how different a teacher is, if they can't do better than slugs on acid. But if that issue was fixed? Then, hell, yeah, I'd be willing to bet that the vast majority of Americans could audit an encrypted, authenticated and verified e-voting system. Maybe not a 100% still, but more than could audit the existing system.
If the customers are the only ones who could gain, and everyone else would lose, then who is going to be insane enough to switch on multicast routing to the home?
Another part of the problem is that most existing implementations are - frankly - crap. They offer minimal security, have frequently been reported as having errors such as non-zero counts, have poor reliability, provide minimal accountability and often provide no means of verification. This is wholly unacceptable. Nobody would accept that from a cash register in a supermarket, never mind a system that is mission-critical in a democracy.
Hand-counts can be reliable. For the longest time, the British system was entirely done by hand-counting, with very small error rates for a population of 60 million. The American system includes machine counts, statistical sampling, and other mechanisms for speeding up the returns, with different States using different methods. It is also worrying that the first returns are announced prior to the polls closing on the west coast, which will inevitably introduce bias and strategic voting. The British system isn't perfect, and has recently developed all kinds of flaws and fraudulant practices, but it can be used as a yardstick of what a democracy should minimally achieve.
Of course, a democracy has other dependencies. It's only meaningful if enough of the population votes for the votes to truly represent the population. The electoral college has the potential for distorting the consensus of the people and probably has. There is no ballot option to reject all candidates and re-open nominations. Media saturation and candidate funding warp awareness. The educational system isn't up to the standards needed to ensure the population have the breadth or depth of knowledge to understand the complexities of a nation or avoid the wiles of a skilled talker. If these flaws remain, then even a perfect voting system can never represent what the public actually want or need, which is what a democracy is about. Being heard has no meaning if you never learned how to talk.
To me, the question shouldn't merely be how we reliably count votes, but should also include how we reliably cast them. There may be no better solution than the one we have, I accept that, but I won't accept that this is known until it actually is.
Of course, none of this matters if you're not using extranets. Any Enterprise-level network that exists in one physical location or connects to other locations over the public Internet or connects to other locations via leased physical lines has no need of MPLS or any other virtual circuit technology. Metropoliton networks, such as GMING, tend to use leased physical lines and point-to-point protocols such as ATM. As do most of the smaller-scale backbones (such as national DSL providers) who bought up dark fibre in bulk. If you're using an ATM network, you don't waste time with IP-based virtual circuits.
I've admined my share of b2b, national and international Enterprise networks. I co-founded the IPv6 backbone in the UK. I've run large-scale X.25, IP and ATM networks. Before that, I designed my own LAN and WAN wire protocols. I'm no Vint Cerf, but let's be realistic here. Anyone can make a claim, but it takes more than that to make a difference.
As for your multicasting commentless comment, the announcement of the dissolution of the MBONE in the mid 90s is quite sufficient as a reply. It's you against the world, and the world declared many years ago that you lost. That war is over and the multicasters won. Give it up. Accept defeat gracefully. The only serious resistance left are the ISPs for SOHOs and individuals and the increasing number of multicasting applications in Windows mean they will submit soon enough. Even mobile phone companies, such as Nokia provide their services by multicast. They have surrendered to the only rational networking technology for distributed services. The Infiniband Consortium and the Open Fabrics Consortium depend heavily on the multicast capabilities of modern technologies. Resistance is futile, you will be IGMPv3'ed.
As did I. Of course, neither of us is in a position to ever really know what happened, all we can ever go by is the limited (and likely editorialized) information made available, which is why it's so important to add all the qualifiers. If there was an error made in this thread, it was on my part for not putting in enough qualifiers in the first place.
Bits of the pedestrian gave way. The driver was cited by police as doing 20+ MPH over the speed limit, although the news story cites (unnamed) independent experts as saying it would have been closer to 40 MPH over the speed limit. (The speed limit was 55 MPH). Far as I can tell from the news article, the driver is not claiming that there was shared responsibility for the accident or that the pedestrian did anything wrong, merely that the pedestrian caused damage to his Audi. If that really is the whole story, then the attitude is no different from the RIAA's (bringing this thread back on topic).
This assumes that the system can't actually be solved or reduced/simplified to one that can without losing too much accuracy. If you can solve it, then all you have is some basic matrix algebra on an 8000x8000 array. Non-trivial, sure, but 4Gb of RAM and a good gaming machine (you want fast maths) would be adequate to crunch such data. Alternatively, an analogue computer would be ideal for a problem like this, as you'd have far greater precision and far greater parallelization. It would also take far more space and cost far more, but the world economy could do with a boost about now.
Funny you should talk about killing. There was a news story recently about a driver who hit and killed a pedestrian, then sued the family for damages to his car. I think we should check to see if he's on the RIAA's advisory panel.
Maybe he does. I've not seen 3Com for a while and Bay went belly-up. If he was the chief designer for those two, it would explain what happened to them.
The translation list is here.
I never said anything about nuclear war, I specified fault tolerence. There's a big difference. Several tens of megatonnes of difference. Providers offering virtual circuits for extranets sometimes use MPLS, but not all. Service guarantees are over time and can be provided by packet marking and QoS. Isolation likewise. When providing full mesh services, which is what many corporate ISPs are moving towards, something like MPLS is a disadvantage. Containment and QoS is only safe at endpoints, and running anything heavier than you need on a full mesh (which itself is insanely hard on routers) is asking for trouble.
You only need an address for each group address subscribed to by a downstream node. Since you have access to port numbers, you can place as many streams on a single address as you like (up to 65535), although obviously you lose some benefit from the multicasting if you overload too many streams onto a single group address. Well, unless you use source-specific multicast (SSM), in which case so long as the content is differentiated by source address, you can stuff everything onto a single group if you really want.
I think you may have missed the -1 CowboyNeil option.
IPv6-over-IPv6 seems to work ok. Some of the earliest routing protocols provided firewalling and NATting within the routing protocol itself (Telebit's router provided superb NAT and Firewall capabilities as an integrated facility). Permanent addresses lead to fragmented heirarchies and exploding routing tables, which is a major problem with IPv4.