Prions appear to exist in a space somewhere between crystals (they "multiply" in a way mot dissimilar to crystals growing and do not make anything "new" the way a virus would) and the lower end of what is considered "true" life. This suggests to me that there is a continuum and that the terms "living" and "non-living" are not descriptive of anything fundamental. If that is correct, then I suspect we will find that prions were an important stage in getting to what we call "life". However, it is not obvious as to how you'd get from a free-floating RNA/DNA strand and a prion to a living cell, so there must be other stages in between, if that indeed was the sequence. I suspect that a study of prions could yield such an additional step, but only if the researchers are willing to accept such a stage could exist.
If badly-functioning financial institutions can be taken over by the Government, maybe it should be law that badly-functioning ISPs should be taken over by Slashdot readers. We can't do worse with technology than the Government does with money, and at least things like Mobile IP and NEMO (Network Mobility) might gain some visibility.
That's one form of Mobile IP, there appears to be some work on passing upstream routing optimizations so that the connection eventually switches from relayed to direct, presumably providing there is something that can act as a new home base in the new location. You're right, Mobile IP isn't much used, but I'll blame that less on the infrastructure and more on the lack of support or documentation in such systems as Windows, which - like it or not - determines what the hardware vendors care about. Bleagh.
First, there's never going to be a "good" DMCA, at least not in those terms. The copyright holders (not the artists, who generally get less from DMCA than they did prior to such laws) are trying to have their cake and eat it. Doesn't work.
Second, if you absolutely have to have such a law, or ANY law on technology, then it has to be written in collaboration with technologists who can help politicians understand what will and won't work, and what is and is not enforceable. You CANNOT EVER make a good law in a vacuum. Every single time politicians and a single special-interest side of the debate try to control everything, it falls apart. If you don't listen, you cannot learn. If you do not learn, you cannot hope to avoid the mistakes of the past.
Bear in mind Intermezzo and CODA were merged with the baseline, and suffered bit-rot. The kernel module for dynamic device numbering also died inside the mainstream, despite fairly heavy usage by the distros. Although it improves the odds of survival, it's not a guarantee of it. It's not much of a guarantee of quality, either - Lustre absolutely beats the carp out of GFS2 and Oracle's cluster file system. (Oracle don't even use the filesystem they donated to Linux, far as they're concerned, it's abandonware.) This doesn't mean ignore what is in the kernel, that should always be the first place to look, but never assume that it's either safe or optimal.
It's not the only "forgotten" architecture. The Transputer was nearly as revolutionary as the Connection Machine, and the AMULET range of CPUs were a rare example of asynchronous general-purpose processors. Processor-In-Memory architectures (which Cray played with for a bit) turned the entire Von Neumann architecture on its head. I'm sure there are other examples of unique, and forgotten, designs that warrant closer examination and which - if reimplemented using the same current limits of knowledge and technology as these high-end mainstream processors - would be as good or better. (Benchmarks against actual alternatives are good, but there should also be benchmarks against where those alternatives should be, given what we know and can do.)
MPICH isn't bad? Compared to what? LA-MPI should be more robust, OpenMPI is partially hand-turned assembly and damn-well should be faster. Commercial, specially-tuned MPIs should be en better. I see your bridge and raise you one Transputer grid. I prefer Occam's mobile processes and dynamic message passing. MPI can't even do collective operations AS a collective operation - it's invariably implemented as a "for" loop - which is exactly why I talk about the need for reliable multicast. You can do collective operations as a single multicast over any number of nodes. PVM should be nearly identical to MPI-1 for most things, so long as the system has a fixed number of nodes. Possibly better, as generic solutions tend to be less tuned than fixed solutions. Not sure how bulk sychronous processing compares. Linux' TIPC should be better for raw IPC-type messages than MPI-1 for the same reason.
...Apple used to use a Cray to design their new computers, whereas Seymoure Cray used an Apple to design his.
More compute power is nice, but only if the programs are making efficient use of it. MPI is not a particularly efficient method of message passing, and many implementations (such as MPICH) are horribly inefficient implementations. Operating systems aren't exactly well-designed for parallelism on this scale, with many benchtests putting TCP/IP-based communications ahead of shared memory on the same fripping node! TCP stacks are not exactly lightweight, and shared memory implies zero copy, so what's the problem?
Network topologies and network architectures are also far more important than raw CPU power, as that is the critical point in any high-performance computing operation. Dolphinics is quoting 2.5 microsecond latencies, Infiniband is about 8 microseconds, and frankly these are far far too slow for modern CPUs. That's before you take into account that most of the benchmarks are based on ping-pong tests (minimal stack usage, no data) and not real-world usage. I know of no network architecture that provides hardware native reliable multicast, for example, despite the fact that most problem-spaces are single-data, most networks already provide multicast, and software-based reliable multicast has existed for a long time. If you want to slash latencies, you've also got to look at hypercube or butterfly topologies, fat-tree is vulnerable to congestion and cascading failures - it also has the worst-possible number of hops to a destination of almost any network. Fat-tree is also about the only one people use.
There is a reason you're seeing Beowulf-like machines in the Top 500 - it's not because PCs are catching up to vector processors, it's because CPU count isn't the big bottleneck and superior designs will outperform merely larger designs. Even with the superior designs out there, though, I would consider them to be nowhere even remotely close to potential. They're superior only with respect to what's been there before, not with respect to where skillful and clueful engineers could take them. If these alternatives are so much better, then why is nobody using them? Firstly, most supercomputers go to the DoD and other Big Agencies, who have lots of money where their brains used to be. Secondly, nobody ever made headlines off having the world's most effective supercomputer. Thirdly, what vendor is going to supply Big Iron that will take longer to replace and won't generate the profit margins?
No great surprise, at least not after reading the recent BBC interview with the Flat Earth Society. Well, I guess I've actually been a lot more despondent about human intelligence long before that, but I guess that really proved things for me.
Look, YouTube isn't going to keep your home movies. They're too busy making copies of the porn people upload before it's deleted, along with all the tv programs and movies that they don't want to be seen pirating via peer-to-peer.
I don't care who makes the processor - let's face it, most chips in US computers are made in Asian countries anyway - all I need is for it to work well. I doubt the Chinese are doing anything radical (that's not generally their style), which is a pity because current chip designs are going down a dead-end and it'll take a radical shift to solve many of the issues to do with parallelism, increasing abstraction in programming languages, and increasing demand for highly robust software. Serious efforts into such radicalizing of technology can be seen with the IBM Cell design (which isn't going anywhere, at the moment) and could be seen in the Transmeta Crusoe and the Inmos Transputer, and the Manchester AMULET was ingenious enough, but pretty much everything else in the CPU world is based on stale ideas and stagnant approaches. Good for backwards compatibility at the binary level, lousy for long-term potential.
VIA is dropping some of its older chip series. IBM has open-sourced the patents for hardware in the past, and a few companies (as evidenced by opencores.org) have open-sourced the chips themselves. Now, not everything dropped could be opened like this, due to IP entanglements or IP that is still in highly active use and therefore important to the company, but that probably doesn't apply to everything. Has VIA debated at all the possibility of placing superceded but still interesting algorithms into open-source software or even open-source hardware?
(Either would seriously boost interest in open-source, and perhaps expand some mental horizons.)
...I'm less interested in pointing fingers. Besides, the US has a habit of shooting at fingers with hellfire missiles. Instead of "naming names", it would seem better to have a close to global tracking and monitoring of pollution in general, to show WHERE different types of pollution are a problem (regardless of source). You could then add in solar-powered UAVs to collect air samples at random points, where the isotope ratios are calculated and the pollutant sources (not necessarily the factories, just the sources) are derived. The factories can be inferred from plotting the pollution clouds, if anyone is genuinely concerned, but frankly I'd have thought that cleaning fuels and raw materials would have a bigger impact, as there are likely far fewer sources than factories, factories see cleaning as expensive, but higher grade fuels and materials are worth more to their producers. Ergo, cleaning at source will be seen as making money, cleaning at point of use will be seen as spending money, even though the end result (in terms of pollution, money-flow, profits, etc) should be absolutely identical.
Industrialists are, by and large, not very bright and highly prejudiced towards green-stuff feel-good factors. Which means that something that is good won't be accepted no matter how good it actually is, unless it is presented as something that'll feel good to their accountants. Being honest isn't worth a damn thing, but it isn't necessary to be honest to be accurate. This is why politics is a scam. Politicians don't sell you what you want, they sell you what they want dressed up to look like it's something you want. But you're quite capable of giving as good as you get.
Honest environmentalists go nowhere, although they usually get some recognition AFTER the disaster they predicted has swept through. Why? Because their phrasing makes it sound like people have to put in hard work and money for something that isn't 100% predictable anyway. Completely the wrong move. Think like Dogbert, not Dilbert, on this one. Dilbert always gets ignored, Dogbert always gets things done. The difference is not in what they're doing, but in the psychology. Dilbert assumes people are basically bright, compassionate and thoughtful. Dogbert assumes people are manipulative, deceitful, corrupt and 100% gullible. Environmentalists need to listen to Dogbert. Dilbert is correct, but will never go anywhere. In mythological terms, he represents the Wise Fool - he knows a lot but his attempts to explain make him sound like a complete fool.
Saving money has never worked, any better than saving the planet, but if the first part of the "food chain" decides cleanliness is next to richness, it gets imposed on everyone else regardless. They have no choice but to go green. They won't even be aware they've done so. Things'll cost more, but as gas prices have demonstrated, customers ignore that until the last possible moment, and then blame it on anyone they happen to dislike at the time. Use that self-inflicted blindness to make consumers green, and the world will be cleaner within a year without the consumers ever noticing what you're doing. If they say anything, it'll be to flame the environmentalists for doom-saying about pollution and greenhouse gasses, same as they did with Y2K after several trillion dollars were spent in fixing flaws across the world.
(And, yes, for those who care, Y2K did strike older electronic credit-card readers, older banking systems, and many home and office products - including many of Microsoft's. If they'd done nothing, the world might well have ended. Instead, the fixes were imposed on an unwilling and ignorant population in such a way that they remained unwilling and ignorant. And that is the SOLE reason you are still breathing today.)
What Y2K demonstrated was that the masses are dumb, but that really doesn't matter. You can fix what does matter without ever concerning yourself with the widespread ignorance in the world. In this case, you can fix mines, quarries, power stations, oil, coal, and all kinds of other resources, with the help of a handful of executives who can make a mint off the deal. Do that, and national follies will be of no importance whatsoever.
What about the myth that private sector TV stations (like Discovery) are more honest and open than Government-funded stations (like the BBC)? On that note, I would like to challenge anyone to build a machine that can steam-roller the named corporations purely from household kitchen and powered by an 8" elastic band. Oh, and it must also carry a raw egg.
You notice that I don't necessarily agree with changing the durations of sentences, merely the context of them. Actually, given that most prisoners will have some combination of hard-headedness, mental illness and social ineptitude, perhaps divide the same sentence into those three components according to an assessment of the nature of the criminal. The idea is not to let people out early because of a low risk, but to ensure that there is a balance between punitive measures and corrective measures. Instead of letting murderers out early, perhaps have other criminals kept in longer. Petty crimes have a low punitive action associated, but if there high reoffense rates, you clearly need much much more corrective and theraputic action. If there's not enough time for that in a conventional sentence, increase it until there is, but ensure the extra IS used for corrective and theraputic action, and reduce the purely punitive measures if appropriate.
You will also note that when I talk of corrective and theraputic measures, I'm not talking about in the community or voluntary schemes. No. These would be relatively compact prison units that are highly specialized to provide medical and/or psychotheraputic facilities. Basically, prison mental hospitals with better facilities and more emphasis on fixing issues than merely stabilizing and quietening troublesome individuals.
It's important to be careful with such an idea - it can't go into cult-like programming techniques (which is the direction Clockwork Orange went in) or into wholesale brain modification (many early techniques in brain surgery involved sticking coat hangers into brains and fishing round until the problem appeared to stop). It has to be responsible and ethical. That's not always going to work, some problems are unfixable with current knowledge and some criminals don't want fixing, so they'd go back to the standard prisons. No big deal - the level of incarceration never changes, the duration and fundamental nature of the sentence doesn't change, all that changes is what options the prisoner has to modify the way they are.
You are correct, and that's why I'd put a slightly different angle on it. Rehabilitation of someone who is going to get released minimizes the risk of re-offense. If they re-offend with rehabilitation, they'd have re-offended without it - and probably worse, as they'd have used that same time to acquire new skills and knowledge from other criminals on how to hide evidence better and be less trackable.
Criminals who would NOT normally be released - ever - due to their dangerous nature should still be rehabilitated, trained/educated in-house, and developed as far as possible, on the theory that (a) they're then less likely to train others in how to commit crimes better, and (b) they may become productive enough that they actually offset the cost of incarceration and training. In the latter case, you have a self-supporting virtual colony of prisoners within the prison system who can actually afford to improve their conditions and the conditions of their fellows off their own backs and not the taxpayers'. They'd still be constrained and confined, but they'd have something meaningful in their lives to look forward to, some prospect of betterment, even though they'd never be free. The lack of any real prospect of anything ever changing, no matter what they do, no matter what they achieve, no matter how reformed they become, must surely place an unimaginable strain on prisoners.
(Why care? Well, aside from the humanity of it, when being shot to death becomes an appealing alternative, prisoners have no incentive to play nice and the prison population becomes an extremely dangerous powder keg. One mistake and you've an army of kamakazis would would see not the remotest possibility of ever being hurt more no matter how many guards get disemboweled. With a spoon. Herbert's first Dune novel also goes into detail on why harsh prison systems breed great psychopaths but lousy citizens.)
Arguably, exploiting Hans' abilities is good for the free software community and, by extension, the country. It doesn't help Nina's family, but then nothing else would either, so it's arguably better to say that it's no worse for them than the best possible case. Is it good for Hans? Maybe, as the sudden lack of mental activity is a major contributing factor behind people becoming insane in prison. Such work would maximize the chances of Hans Reiser being able to keep himself functional. Now, perhaps there should be a punitive measure to it - oh, perhaps the prison could REQUIRE him to work X number of hours per day on the system, or face loss of privileges, along with some sort of reward (eg: demonstrating the gain of maturity and acquiring of social responsibility via honouring and respecting the views of others in the community, rather than slagging them off, as a factor to be considered when he becomes eligible for parole).
I don't see how the system could lose by doing that, but I'm aware that the public by-and-large tend to be far harsher than I am on punishment systems and that's why I was looking for a way to achieve the same kind of mental stimulation and same sort of demonstrable contrition without giving the appearance of a good-old-boys club seeking luxuries for one of their own. That's not what I'm after. However, if you think that computer access would not give such an impression, then yes, the greater role would be the better role.
"He deserved death? I daresay. I daresay many who live deserve death, and many who die deserve life. Can you give them that also? Then do not be so quick to judge." (My memory is awful at the moment, so I can't be 100% sure of some of the phrasing but that was the general gist.)
Translated into modern terms, justice is simply legalized revenge. It serves no constructive purpose and simply adds to the misery of the world. It feels satisfying at the time, sure. But it's a cathexis - a temporary feeling of relief that soon fades and must be regenerated repeatedly to retain the same feeling. Which ultimately means that those seeking relief must - usually covertly - "retaliate" against other people, innocent people, and keep doing so until they deal with their problem or die.
Of course, if they'd dealt with their problem at the start, they'd have no need of the original cathexis. It would not give them any feeling of justice, because that would have been taken care of in their eyes already. (The criminal is never going to repeat, the criminal is being removed from society, the criminal is deprived of luxuries and benefits, thus is punitively dealt with, satisfying the demands of REAL justice without the added component of a revenge cycle.)
It is by NOT dealing with issues, but avoiding them, that is the real reason such a large percentage of Americans are in prison - more per capita than any nation bar China. Avoidance mistaken for justice is the gravest injustice of all. For everyone.
Neutrinos normally interact very very rarely - well, except for the extremely high-energy neutrinos, which get absorbed at or near the surface. One reason neutrino detectors are deep underground is that they want the low-energy neutrinos from the sun, not the really high-energy stuff. However, if radioactive decay is caused (even in part) by neutrino interactions, then radioactive decay must primarily involve easily-absorbed neutrinos, which means the really high-energy stuff.
That's the principle I'm largely relying on. Now, the other part - setting up a diffraction grating and using diffraction patterns to demonstrate variation in decay rates - that's tougher. What's the wavelength of a neutrino? Does anyone know? And does anyone know how to build a grating that fine? (The grid must be roughly the same scale as the thing you're diffracting. I'm not sure what the spacing is between sheets of graphite, or how large the linear motor layer is for skimming electrons over in a superconductor, but those might actually be too big to diffract something like a high-energy neutrino.)
In other words, this would not be a simple experiment. BUT, if it could be done, it might make a massive difference in accelerating decay. Once you accelerate decay locally, regular neutron emissions would accelerate the decay of the rest.
He is able to write, so technically he can help the community - say by documenting Reiser4, or writing down some of his ideas. True, it's not as good as if he was in the outside world, but it's better than nothing. Personally, I think criminals who have verifiable mental issues would be better in a hospital (with equal confinement and punitive measures, but focussed on curative action). Those who have committed crimes they are unlikely to repeat, possibly including Hans, might be better off in a smaller, more secure, facility intended for rehabilitation. Purely punitive systems should really be restricted to those who are unwilling to change except under duress. And, frankly, I don't think there are many such people. There was a good blog discussion about that on the BBC website recently, with a lot of hostility from prison guards, prison governers and social workers to Victorian-style systems except as a last-resort, and not much more patience for the panoptican idea (a prison where a central warden can see into every cell directly from a central station.-
I agree that it's very important to put the text online. The best way to minimize the risk of tampering is if the originators publish a public key, and a digital signature for each photograph uploaded. Any signature that can't be verified by the official key is an unofficial signature. Any picture for which there is no signature or where the signature doesn't match up to the picture is an indication of tampering (even if benign). It only minimizes the risk, but it goes a long way towards helping.
Next, there are many documents already online (the early version of the New Testament broken into fragments after World War 2 and kept in several nations, for example, which might be a palimpset - the photographs indicate it but the researchers refused to answer any questions regarding the document) and many that I'd like to see online. However, these resources are scattered and hard to use in any kind of comparitive or structured way. Better coordination by the projects and better navigation between them would be greatly beneficial.
The Mandaeans (surviving direct descendents of those Nazarenes who followed John the Baptist, according to both their traditions and genetic studies) have published some of their texts online, but it's confusing and again very hard to compare with this other literature, as they've provided straight ASCII text only. I can't find quality documents on the early beliefs of the Falasha (the claimed descendents of King Solomon who migrated to Ethiopia during Solomon's reign). Although neither are technically "Christian", I am a firm believer in understanding some of the context the Old Testament and New Testament writers may have assumed we'd know but which has not been preserved over the course of history. We'll never know everything. The Old Testament makes references to books that do not appear to have survived and to historic events probably never recorded. The more we can understand how they saw things, the more we can be sure which translation, where ambiguities exist and there are lots, were intended. We can also do better at distinguishing between cultural beliefs, Rabinical law and "revealed" law - these are not clear and many religious arguments have centered around which should be taken as which.
From a historical perspective, better connections between texts can help improve our understanding of how the religion changed over time and - back to that context thing - why. Analysis of the texts appear to indicate that some modern books are the synthesis of seperate, distinct books, but without a better way of mining the texts, you will NEVER be able to see how, when or why, or be able to reliably check other documents for similar merging or editing. Some documents have also been subject to forks (yes, that's not a problem unique to open source software) within the Judeo-Christian amalgam but also between Judeo-Christian beliefs and other religions. Unless you can mine the data efficiently and between all these distinct text, how can anyone be expected to know which is the vanilla branch, why the fork happened, or which adheres closest to the "requirement specification" the text was supposed to serve?
(If you're a believer in revealed religion, then the requirement is that it meets the needs and goals of the revealing power. If you are not, or are but believe that text was based on social beliefs only, then the requirement is whether the text met the goals of those social beliefs.)
As things stand, almost all the information could could possibly want exists. Somewhere. But unless you have a dedicated army of translators, photographers, researchers and theologans, plus the computing power of Google and the Top500 combined, I do not see how anyone could sensibly claim to be able to make actual use of what's there or reliably provide clear limits on what missing information could have provided or know which texts were in-period "fakes" (where "genuine" refers to the requirements and the source, not to the accuracy of a specific belief).
Neutrinos cannot be absorbed at a different energy than the anti-neutrino would be released at had the decay worked in that direction instead. This alters the problem slightly. If you could find a way of shifting the energy off the high-energy neutrinos, so that there were far more neutrinos of the ideal energy, you increase the number of the desired neutrinos and therefore increase the rate of decay. The other question is HOW the neutrinos interact. I'm not convinced there's a good answer to that, given that it's still very new to think of radioactive decay being variable. The point being, if you know how neutrinos need to interact with nuclei/quarks/sub-quark particles/whatever, you've a better chance of rigging the odds in your favour. It might be something relatively trivial (a neutrino can only be absorbed if such-and-such a state exists) or it might be something much tougher (you need a direct strike on the third higgs particle on the left, but if and only if there is a fat lady singing somewhere on the planet).
You could probably experiment with the idea of hitting a specific target by using a diffraction grid and a segmented neutrino detector. If neutrinos can only be absorbed at very specific points, diffracting them should produce changes in probability that are a function of the density, and the segments will record a varying absorbtion rate accordingly. If it's a state thing, then the change in state should (I would think) be propogated between close enough detectors, so the absorbtion rate will NOT vary the way you would predict from a simple interference pattern. It'll still vary, but the variation could be expected to be skewed. (Simply take the sum of the square of the absolute differences between actual and predicted for each detector. Over time, the variance should remain roughly constant on an unskewed system, but should be asymtotic to that same constant for a skewed system.)
Slashdot Mirror
Prions appear to exist in a space somewhere between crystals (they "multiply" in a way mot dissimilar to crystals growing and do not make anything "new" the way a virus would) and the lower end of what is considered "true" life. This suggests to me that there is a continuum and that the terms "living" and "non-living" are not descriptive of anything fundamental. If that is correct, then I suspect we will find that prions were an important stage in getting to what we call "life". However, it is not obvious as to how you'd get from a free-floating RNA/DNA strand and a prion to a living cell, so there must be other stages in between, if that indeed was the sequence. I suspect that a study of prions could yield such an additional step, but only if the researchers are willing to accept such a stage could exist.
If badly-functioning financial institutions can be taken over by the Government, maybe it should be law that badly-functioning ISPs should be taken over by Slashdot readers. We can't do worse with technology than the Government does with money, and at least things like Mobile IP and NEMO (Network Mobility) might gain some visibility.
That's one form of Mobile IP, there appears to be some work on passing upstream routing optimizations so that the connection eventually switches from relayed to direct, presumably providing there is something that can act as a new home base in the new location. You're right, Mobile IP isn't much used, but I'll blame that less on the infrastructure and more on the lack of support or documentation in such systems as Windows, which - like it or not - determines what the hardware vendors care about. Bleagh.
First, there's never going to be a "good" DMCA, at least not in those terms. The copyright holders (not the artists, who generally get less from DMCA than they did prior to such laws) are trying to have their cake and eat it. Doesn't work.
Second, if you absolutely have to have such a law, or ANY law on technology, then it has to be written in collaboration with technologists who can help politicians understand what will and won't work, and what is and is not enforceable. You CANNOT EVER make a good law in a vacuum. Every single time politicians and a single special-interest side of the debate try to control everything, it falls apart. If you don't listen, you cannot learn. If you do not learn, you cannot hope to avoid the mistakes of the past.
Bear in mind Intermezzo and CODA were merged with the baseline, and suffered bit-rot. The kernel module for dynamic device numbering also died inside the mainstream, despite fairly heavy usage by the distros. Although it improves the odds of survival, it's not a guarantee of it. It's not much of a guarantee of quality, either - Lustre absolutely beats the carp out of GFS2 and Oracle's cluster file system. (Oracle don't even use the filesystem they donated to Linux, far as they're concerned, it's abandonware.) This doesn't mean ignore what is in the kernel, that should always be the first place to look, but never assume that it's either safe or optimal.
It's not the only "forgotten" architecture. The Transputer was nearly as revolutionary as the Connection Machine, and the AMULET range of CPUs were a rare example of asynchronous general-purpose processors. Processor-In-Memory architectures (which Cray played with for a bit) turned the entire Von Neumann architecture on its head. I'm sure there are other examples of unique, and forgotten, designs that warrant closer examination and which - if reimplemented using the same current limits of knowledge and technology as these high-end mainstream processors - would be as good or better. (Benchmarks against actual alternatives are good, but there should also be benchmarks against where those alternatives should be, given what we know and can do.)
MPICH isn't bad? Compared to what? LA-MPI should be more robust, OpenMPI is partially hand-turned assembly and damn-well should be faster. Commercial, specially-tuned MPIs should be en better. I see your bridge and raise you one Transputer grid. I prefer Occam's mobile processes and dynamic message passing. MPI can't even do collective operations AS a collective operation - it's invariably implemented as a "for" loop - which is exactly why I talk about the need for reliable multicast. You can do collective operations as a single multicast over any number of nodes. PVM should be nearly identical to MPI-1 for most things, so long as the system has a fixed number of nodes. Possibly better, as generic solutions tend to be less tuned than fixed solutions. Not sure how bulk sychronous processing compares. Linux' TIPC should be better for raw IPC-type messages than MPI-1 for the same reason.
...Apple used to use a Cray to design their new computers, whereas Seymoure Cray used an Apple to design his.
More compute power is nice, but only if the programs are making efficient use of it. MPI is not a particularly efficient method of message passing, and many implementations (such as MPICH) are horribly inefficient implementations. Operating systems aren't exactly well-designed for parallelism on this scale, with many benchtests putting TCP/IP-based communications ahead of shared memory on the same fripping node! TCP stacks are not exactly lightweight, and shared memory implies zero copy, so what's the problem?
Network topologies and network architectures are also far more important than raw CPU power, as that is the critical point in any high-performance computing operation. Dolphinics is quoting 2.5 microsecond latencies, Infiniband is about 8 microseconds, and frankly these are far far too slow for modern CPUs. That's before you take into account that most of the benchmarks are based on ping-pong tests (minimal stack usage, no data) and not real-world usage. I know of no network architecture that provides hardware native reliable multicast, for example, despite the fact that most problem-spaces are single-data, most networks already provide multicast, and software-based reliable multicast has existed for a long time. If you want to slash latencies, you've also got to look at hypercube or butterfly topologies, fat-tree is vulnerable to congestion and cascading failures - it also has the worst-possible number of hops to a destination of almost any network. Fat-tree is also about the only one people use.
There is a reason you're seeing Beowulf-like machines in the Top 500 - it's not because PCs are catching up to vector processors, it's because CPU count isn't the big bottleneck and superior designs will outperform merely larger designs. Even with the superior designs out there, though, I would consider them to be nowhere even remotely close to potential. They're superior only with respect to what's been there before, not with respect to where skillful and clueful engineers could take them. If these alternatives are so much better, then why is nobody using them? Firstly, most supercomputers go to the DoD and other Big Agencies, who have lots of money where their brains used to be. Secondly, nobody ever made headlines off having the world's most effective supercomputer. Thirdly, what vendor is going to supply Big Iron that will take longer to replace and won't generate the profit margins?
(Me? Cynical?)
No great surprise, at least not after reading the recent BBC interview with the Flat Earth Society. Well, I guess I've actually been a lot more despondent about human intelligence long before that, but I guess that really proved things for me.
Look, YouTube isn't going to keep your home movies. They're too busy making copies of the porn people upload before it's deleted, along with all the tv programs and movies that they don't want to be seen pirating via peer-to-peer.
They're a search company and obviously used the first EULA their search engine found.
I don't care who makes the processor - let's face it, most chips in US computers are made in Asian countries anyway - all I need is for it to work well. I doubt the Chinese are doing anything radical (that's not generally their style), which is a pity because current chip designs are going down a dead-end and it'll take a radical shift to solve many of the issues to do with parallelism, increasing abstraction in programming languages, and increasing demand for highly robust software. Serious efforts into such radicalizing of technology can be seen with the IBM Cell design (which isn't going anywhere, at the moment) and could be seen in the Transmeta Crusoe and the Inmos Transputer, and the Manchester AMULET was ingenious enough, but pretty much everything else in the CPU world is based on stale ideas and stagnant approaches. Good for backwards compatibility at the binary level, lousy for long-term potential.
VIA is dropping some of its older chip series. IBM has open-sourced the patents for hardware in the past, and a few companies (as evidenced by opencores.org) have open-sourced the chips themselves. Now, not everything dropped could be opened like this, due to IP entanglements or IP that is still in highly active use and therefore important to the company, but that probably doesn't apply to everything. Has VIA debated at all the possibility of placing superceded but still interesting algorithms into open-source software or even open-source hardware?
(Either would seriously boost interest in open-source, and perhaps expand some mental horizons.)
...I'm less interested in pointing fingers. Besides, the US has a habit of shooting at fingers with hellfire missiles. Instead of "naming names", it would seem better to have a close to global tracking and monitoring of pollution in general, to show WHERE different types of pollution are a problem (regardless of source). You could then add in solar-powered UAVs to collect air samples at random points, where the isotope ratios are calculated and the pollutant sources (not necessarily the factories, just the sources) are derived. The factories can be inferred from plotting the pollution clouds, if anyone is genuinely concerned, but frankly I'd have thought that cleaning fuels and raw materials would have a bigger impact, as there are likely far fewer sources than factories, factories see cleaning as expensive, but higher grade fuels and materials are worth more to their producers. Ergo, cleaning at source will be seen as making money, cleaning at point of use will be seen as spending money, even though the end result (in terms of pollution, money-flow, profits, etc) should be absolutely identical.
Industrialists are, by and large, not very bright and highly prejudiced towards green-stuff feel-good factors. Which means that something that is good won't be accepted no matter how good it actually is, unless it is presented as something that'll feel good to their accountants. Being honest isn't worth a damn thing, but it isn't necessary to be honest to be accurate. This is why politics is a scam. Politicians don't sell you what you want, they sell you what they want dressed up to look like it's something you want. But you're quite capable of giving as good as you get.
Honest environmentalists go nowhere, although they usually get some recognition AFTER the disaster they predicted has swept through. Why? Because their phrasing makes it sound like people have to put in hard work and money for something that isn't 100% predictable anyway. Completely the wrong move. Think like Dogbert, not Dilbert, on this one. Dilbert always gets ignored, Dogbert always gets things done. The difference is not in what they're doing, but in the psychology. Dilbert assumes people are basically bright, compassionate and thoughtful. Dogbert assumes people are manipulative, deceitful, corrupt and 100% gullible. Environmentalists need to listen to Dogbert. Dilbert is correct, but will never go anywhere. In mythological terms, he represents the Wise Fool - he knows a lot but his attempts to explain make him sound like a complete fool.
Saving money has never worked, any better than saving the planet, but if the first part of the "food chain" decides cleanliness is next to richness, it gets imposed on everyone else regardless. They have no choice but to go green. They won't even be aware they've done so. Things'll cost more, but as gas prices have demonstrated, customers ignore that until the last possible moment, and then blame it on anyone they happen to dislike at the time. Use that self-inflicted blindness to make consumers green, and the world will be cleaner within a year without the consumers ever noticing what you're doing. If they say anything, it'll be to flame the environmentalists for doom-saying about pollution and greenhouse gasses, same as they did with Y2K after several trillion dollars were spent in fixing flaws across the world.
(And, yes, for those who care, Y2K did strike older electronic credit-card readers, older banking systems, and many home and office products - including many of Microsoft's. If they'd done nothing, the world might well have ended. Instead, the fixes were imposed on an unwilling and ignorant population in such a way that they remained unwilling and ignorant. And that is the SOLE reason you are still breathing today.)
What Y2K demonstrated was that the masses are dumb, but that really doesn't matter. You can fix what does matter without ever concerning yourself with the widespread ignorance in the world. In this case, you can fix mines, quarries, power stations, oil, coal, and all kinds of other resources, with the help of a handful of executives who can make a mint off the deal. Do that, and national follies will be of no importance whatsoever.
What about the myth that private sector TV stations (like Discovery) are more honest and open than Government-funded stations (like the BBC)? On that note, I would like to challenge anyone to build a machine that can steam-roller the named corporations purely from household kitchen and powered by an 8" elastic band. Oh, and it must also carry a raw egg.
You notice that I don't necessarily agree with changing the durations of sentences, merely the context of them. Actually, given that most prisoners will have some combination of hard-headedness, mental illness and social ineptitude, perhaps divide the same sentence into those three components according to an assessment of the nature of the criminal. The idea is not to let people out early because of a low risk, but to ensure that there is a balance between punitive measures and corrective measures. Instead of letting murderers out early, perhaps have other criminals kept in longer. Petty crimes have a low punitive action associated, but if there high reoffense rates, you clearly need much much more corrective and theraputic action. If there's not enough time for that in a conventional sentence, increase it until there is, but ensure the extra IS used for corrective and theraputic action, and reduce the purely punitive measures if appropriate.
You will also note that when I talk of corrective and theraputic measures, I'm not talking about in the community or voluntary schemes. No. These would be relatively compact prison units that are highly specialized to provide medical and/or psychotheraputic facilities. Basically, prison mental hospitals with better facilities and more emphasis on fixing issues than merely stabilizing and quietening troublesome individuals.
It's important to be careful with such an idea - it can't go into cult-like programming techniques (which is the direction Clockwork Orange went in) or into wholesale brain modification (many early techniques in brain surgery involved sticking coat hangers into brains and fishing round until the problem appeared to stop). It has to be responsible and ethical. That's not always going to work, some problems are unfixable with current knowledge and some criminals don't want fixing, so they'd go back to the standard prisons. No big deal - the level of incarceration never changes, the duration and fundamental nature of the sentence doesn't change, all that changes is what options the prisoner has to modify the way they are.
You are correct, and that's why I'd put a slightly different angle on it. Rehabilitation of someone who is going to get released minimizes the risk of re-offense. If they re-offend with rehabilitation, they'd have re-offended without it - and probably worse, as they'd have used that same time to acquire new skills and knowledge from other criminals on how to hide evidence better and be less trackable.
Criminals who would NOT normally be released - ever - due to their dangerous nature should still be rehabilitated, trained/educated in-house, and developed as far as possible, on the theory that (a) they're then less likely to train others in how to commit crimes better, and (b) they may become productive enough that they actually offset the cost of incarceration and training. In the latter case, you have a self-supporting virtual colony of prisoners within the prison system who can actually afford to improve their conditions and the conditions of their fellows off their own backs and not the taxpayers'. They'd still be constrained and confined, but they'd have something meaningful in their lives to look forward to, some prospect of betterment, even though they'd never be free. The lack of any real prospect of anything ever changing, no matter what they do, no matter what they achieve, no matter how reformed they become, must surely place an unimaginable strain on prisoners.
(Why care? Well, aside from the humanity of it, when being shot to death becomes an appealing alternative, prisoners have no incentive to play nice and the prison population becomes an extremely dangerous powder keg. One mistake and you've an army of kamakazis would would see not the remotest possibility of ever being hurt more no matter how many guards get disemboweled. With a spoon. Herbert's first Dune novel also goes into detail on why harsh prison systems breed great psychopaths but lousy citizens.)
Arguably, exploiting Hans' abilities is good for the free software community and, by extension, the country. It doesn't help Nina's family, but then nothing else would either, so it's arguably better to say that it's no worse for them than the best possible case. Is it good for Hans? Maybe, as the sudden lack of mental activity is a major contributing factor behind people becoming insane in prison. Such work would maximize the chances of Hans Reiser being able to keep himself functional. Now, perhaps there should be a punitive measure to it - oh, perhaps the prison could REQUIRE him to work X number of hours per day on the system, or face loss of privileges, along with some sort of reward (eg: demonstrating the gain of maturity and acquiring of social responsibility via honouring and respecting the views of others in the community, rather than slagging them off, as a factor to be considered when he becomes eligible for parole).
I don't see how the system could lose by doing that, but I'm aware that the public by-and-large tend to be far harsher than I am on punishment systems and that's why I was looking for a way to achieve the same kind of mental stimulation and same sort of demonstrable contrition without giving the appearance of a good-old-boys club seeking luxuries for one of their own. That's not what I'm after. However, if you think that computer access would not give such an impression, then yes, the greater role would be the better role.
"He deserved death? I daresay. I daresay many who live deserve death, and many who die deserve life. Can you give them that also? Then do not be so quick to judge." (My memory is awful at the moment, so I can't be 100% sure of some of the phrasing but that was the general gist.)
Translated into modern terms, justice is simply legalized revenge. It serves no constructive purpose and simply adds to the misery of the world. It feels satisfying at the time, sure. But it's a cathexis - a temporary feeling of relief that soon fades and must be regenerated repeatedly to retain the same feeling. Which ultimately means that those seeking relief must - usually covertly - "retaliate" against other people, innocent people, and keep doing so until they deal with their problem or die.
Of course, if they'd dealt with their problem at the start, they'd have no need of the original cathexis. It would not give them any feeling of justice, because that would have been taken care of in their eyes already. (The criminal is never going to repeat, the criminal is being removed from society, the criminal is deprived of luxuries and benefits, thus is punitively dealt with, satisfying the demands of REAL justice without the added component of a revenge cycle.)
It is by NOT dealing with issues, but avoiding them, that is the real reason such a large percentage of Americans are in prison - more per capita than any nation bar China. Avoidance mistaken for justice is the gravest injustice of all. For everyone.
Neutrinos normally interact very very rarely - well, except for the extremely high-energy neutrinos, which get absorbed at or near the surface. One reason neutrino detectors are deep underground is that they want the low-energy neutrinos from the sun, not the really high-energy stuff. However, if radioactive decay is caused (even in part) by neutrino interactions, then radioactive decay must primarily involve easily-absorbed neutrinos, which means the really high-energy stuff.
That's the principle I'm largely relying on. Now, the other part - setting up a diffraction grating and using diffraction patterns to demonstrate variation in decay rates - that's tougher. What's the wavelength of a neutrino? Does anyone know? And does anyone know how to build a grating that fine? (The grid must be roughly the same scale as the thing you're diffracting. I'm not sure what the spacing is between sheets of graphite, or how large the linear motor layer is for skimming electrons over in a superconductor, but those might actually be too big to diffract something like a high-energy neutrino.)
In other words, this would not be a simple experiment. BUT, if it could be done, it might make a massive difference in accelerating decay. Once you accelerate decay locally, regular neutron emissions would accelerate the decay of the rest.
He is able to write, so technically he can help the community - say by documenting Reiser4, or writing down some of his ideas. True, it's not as good as if he was in the outside world, but it's better than nothing. Personally, I think criminals who have verifiable mental issues would be better in a hospital (with equal confinement and punitive measures, but focussed on curative action). Those who have committed crimes they are unlikely to repeat, possibly including Hans, might be better off in a smaller, more secure, facility intended for rehabilitation. Purely punitive systems should really be restricted to those who are unwilling to change except under duress. And, frankly, I don't think there are many such people. There was a good blog discussion about that on the BBC website recently, with a lot of hostility from prison guards, prison governers and social workers to Victorian-style systems except as a last-resort, and not much more patience for the panoptican idea (a prison where a central warden can see into every cell directly from a central station.-
I agree that it's very important to put the text online. The best way to minimize the risk of tampering is if the originators publish a public key, and a digital signature for each photograph uploaded. Any signature that can't be verified by the official key is an unofficial signature. Any picture for which there is no signature or where the signature doesn't match up to the picture is an indication of tampering (even if benign). It only minimizes the risk, but it goes a long way towards helping.
Next, there are many documents already online (the early version of the New Testament broken into fragments after World War 2 and kept in several nations, for example, which might be a palimpset - the photographs indicate it but the researchers refused to answer any questions regarding the document) and many that I'd like to see online. However, these resources are scattered and hard to use in any kind of comparitive or structured way. Better coordination by the projects and better navigation between them would be greatly beneficial.
The Mandaeans (surviving direct descendents of those Nazarenes who followed John the Baptist, according to both their traditions and genetic studies) have published some of their texts online, but it's confusing and again very hard to compare with this other literature, as they've provided straight ASCII text only. I can't find quality documents on the early beliefs of the Falasha (the claimed descendents of King Solomon who migrated to Ethiopia during Solomon's reign). Although neither are technically "Christian", I am a firm believer in understanding some of the context the Old Testament and New Testament writers may have assumed we'd know but which has not been preserved over the course of history. We'll never know everything. The Old Testament makes references to books that do not appear to have survived and to historic events probably never recorded. The more we can understand how they saw things, the more we can be sure which translation, where ambiguities exist and there are lots, were intended. We can also do better at distinguishing between cultural beliefs, Rabinical law and "revealed" law - these are not clear and many religious arguments have centered around which should be taken as which.
From a historical perspective, better connections between texts can help improve our understanding of how the religion changed over time and - back to that context thing - why. Analysis of the texts appear to indicate that some modern books are the synthesis of seperate, distinct books, but without a better way of mining the texts, you will NEVER be able to see how, when or why, or be able to reliably check other documents for similar merging or editing. Some documents have also been subject to forks (yes, that's not a problem unique to open source software) within the Judeo-Christian amalgam but also between Judeo-Christian beliefs and other religions. Unless you can mine the data efficiently and between all these distinct text, how can anyone be expected to know which is the vanilla branch, why the fork happened, or which adheres closest to the "requirement specification" the text was supposed to serve?
(If you're a believer in revealed religion, then the requirement is that it meets the needs and goals of the revealing power. If you are not, or are but believe that text was based on social beliefs only, then the requirement is whether the text met the goals of those social beliefs.)
As things stand, almost all the information could could possibly want exists. Somewhere. But unless you have a dedicated army of translators, photographers, researchers and theologans, plus the computing power of Google and the Top500 combined, I do not see how anyone could sensibly claim to be able to make actual use of what's there or reliably provide clear limits on what missing information could have provided or know which texts were in-period "fakes" (where "genuine" refers to the requirements and the source, not to the accuracy of a specific belief).
Neutrinos cannot be absorbed at a different energy than the anti-neutrino would be released at had the decay worked in that direction instead. This alters the problem slightly. If you could find a way of shifting the energy off the high-energy neutrinos, so that there were far more neutrinos of the ideal energy, you increase the number of the desired neutrinos and therefore increase the rate of decay. The other question is HOW the neutrinos interact. I'm not convinced there's a good answer to that, given that it's still very new to think of radioactive decay being variable. The point being, if you know how neutrinos need to interact with nuclei/quarks/sub-quark particles/whatever, you've a better chance of rigging the odds in your favour. It might be something relatively trivial (a neutrino can only be absorbed if such-and-such a state exists) or it might be something much tougher (you need a direct strike on the third higgs particle on the left, but if and only if there is a fat lady singing somewhere on the planet).
You could probably experiment with the idea of hitting a specific target by using a diffraction grid and a segmented neutrino detector. If neutrinos can only be absorbed at very specific points, diffracting them should produce changes in probability that are a function of the density, and the segments will record a varying absorbtion rate accordingly. If it's a state thing, then the change in state should (I would think) be propogated between close enough detectors, so the absorbtion rate will NOT vary the way you would predict from a simple interference pattern. It'll still vary, but the variation could be expected to be skewed. (Simply take the sum of the square of the absolute differences between actual and predicted for each detector. Over time, the variance should remain roughly constant on an unskewed system, but should be asymtotic to that same constant for a skewed system.)
I tried dating Sapphire, but she put me in a closed time loop.