A man-in-the-middle attack is possible, using your suggested method, which is why quantum entanglement (in which there is no intermediate location) is a better solution. However, it is by no means certain, as quantum cryptography relies on the assumption that messages cannot be blocked - that the MitM can't prevent the original receiver from also getting the message and therefore getting the information that the key had been intercepted.
Another option is to package the values up into blocks and use a cryptographically authenticating block chaining method. That way, if someone attempts to build their own chain, the chaining method will fail to authenticate correctly. That way, even if a MitM found a way to block reception of the original OTP, the recipient would still know that the key had been intercepted.
...if you take all of the source needed to build a useful distribution for the general user, it would likely eclipse 10,000 bugs. My guess would be nearer 100,000 for the kernel alone, and 10,000,000 for the system. (This is including your kernel, your basic utilities, X, a general-purpose window manager, office utilities, system management tools, etc.)
Could this be reduced to the point where the general user would be satisfied? Well, I believe it to be already superior to many other popular general-purpose systems. There's some lack in third-party applications, X is still horrible for abstract graphics and distros aren't merging in drivers that are necessary (eg: madwifi), but those are deficiencies in imagination, not bugs in the code.
Nonetheless, there are bugs in the software, and yes these can be fixed. If there's a code audit, a random number of bugs will be found on each sweep, so the time taken to fix all of them will be indeterminate. The correctness of any given fix will also be indeterminate. If we propose an average rate of 1 bugfix every 3 days (this includes time for tracking down even the most obscure of bugs, testing the fix, and having it filter through the usual chain of developers), it would take about twenty thousand programmers a decade to reduce the number of bugs to the absolute minimum - I'm guessing that this would be in the low hundreds. (As the code is being developed, new bugs will be being added, so there will come a time when the bugs being added by the core developers will exceed the bugs being developed by this hypothetical army.)
A more rigorous method would be to use formal software methods to validate each function is doing what is intended. You could then be sure no new bugs were being added by the bug-stompers. However, because ANY change would totally trash the calculations being done, you would need to fix entire functions at a time, regardless of how many bugs there were. I'd say that it would take the same sized mob about the same length of time, but with greater certainty as to what was fixed and what was left to be fixed.
By this logic, Microsoft or IBM could have fixed their flagship Operating Systems to the point of being pretty nearly invulnerable to defects of any kind. They've had the manpower, the time and the money to perform the kind of total code review suggested. They have clearly not done so, suggesting that the marketplace is largely unaffected by the number of defects in software, but is far more influenced by application availability. Where applications are equal, then the fight is decided by brand.
Is there anything that can independently confirm this suspicion? Yes. The BBC Microcomputer had branding and software. The Acorn Business Computer had neither, even though it was technically made by the same people. OS/2 and NT4 were built from the same underlying codebase, but NT had the applications. Guess which won. The history of computing is littered with such examples.
This is why I am always skeptical of the argument that large numbers of bugs are inevitable. I believe it is closer to the truth that users don't give a damn (except when they lose data, and then they blame the computer, the dog, the telephone - anything but the software). I also believe that computer companies know that the users don't care, and therefore lower their standards to meet those expectations.
At present, Coverity's software seems to indicate that Open Source is down to the low hundreds of bugs per million lines of code, for those classes of bug their software can detect - which is some unknown fraction of the number of possible bug types that could exist. (I'd guess 10%, which would mean there's about a thousand bugs per million lines of code.) I see no reason why, within a decade, we could not be seeing reports of Open Source having fewer than a hundred bugs per trillion lines of code - other than nobody wanting to invest in that much clean-up when almost none of the mainstream users
Quantum Cryptography employs a one-time pad. One-time pads cannot be broken without the pad, because every possible decoded state is equally valid. Breaking cryptography relies on being able to know when you have the right key, but if all keys are equally good, you have no means of knowing. This is made worse by the fact that a true OTP involves a key of equal length to the message. So you not only have no means of knowing which key is correct, you have virtually no hope of sweeping through that part of the keyspace that actually has the right key.
The one big vulnerability with OTPs is that you've now got to send the key securely. Since it is equal in size to the message and is only valid for one message, it is equally hard to send the key securely as it is to send the message securely. Because the pad is pure randomness, it is possible (using existing methods) to send the pad by public key encryption, as it is non-trivial for someone intercepting the message to know how to decrypt it, as it's hard to know when you've broken the encryption. One piece of randomness looks much like another.
Generally, though, people take shortcuts. Instead of using a full-sized one-time pad, a much smaller, repeatedly-used pad is used instead, with some form of pseudo-random mangling to churn things up so that it acts in a very similar manner to a one-time pad. This is generally how stream ciphers work.
Quantum Cryptography - if used sensibly - would involve transmitting a gigantic OTP. Far bigger than the one you need. You then drop all of the bytes that are intercepted. The only bytes used in the pad are the ones the intercepting person does NOT have, so you know the pad is free of holes.
A "better" solution would be to not transmit the key at all, but somehow exploit photon teleportation to deliver the key in a secure manner. However, if you could do that, you wouldn't need encryption in the first place.
There's no doubt NetBSD's port for VAX hardware is by far the best *nix for those machines. In fact, given all the times I saw VMS 5.5 crash from resource management issues, I'd be willing to believe you if you said it was the best OS for VAX hardware.
My point with Linux was definitely not to claim Linux was better for that platform - it isn't. It was half intended to be vaguely humerous and half intended to provoke any Linux user reading it into wondering just what else is out there in the way of extensions and capabilities that they don't know about. It's too easy to assume that the mainstream kernel (or the one that comes with the distro) is all there is, when really it's only the leading edge of the very first part of the beginning of what's available.
(I wouldn't presume to do that for NetBSD, but only because I don't know of any extensions for it. If I did, and I had enough familiarity with that kernel, I'd probably be looking to stretch a few mental muscles there as well.)
...everything would be done in hardware. When using software, you expend energy moving each operation from memory to the processor, in addition to data. Hardware implementations involve moving just the data and only from the last processing element to the next, with extremely limited logic. This is all power-saving on a massive scale.
In practice, nobody writes complex programs purely in hardware any more. It's much more power-efficient, it is usually faster, but there's a price to pay. It takes more space and it's often impossible to update. In consequence, the optimal solution is a compromise - not realy "pure" hardware solutions, but still "mostly" hardware with a bunch of programmable registers and gates.
Even then, this won't do everything you'd want. You still need some software. But it should be kept to an absolute minimum, for a low-power device. Software is cheap on space, but costly on power. It is also dog-slow, which means that you need faster (and therefore more power-hungry) CPUs just to get the same performance.
Ok, given that, what needs to be done to wring as much performance as possible, for as little power? Offload engines for CPU-intensive work (such as cryptography) is an obvious must. You also really want offload engines for anything that is used frequently enough that even a small saving would make a huge impact. A good memory manager would be useful, by that reasoning.
You also want to use every trick in the book to squeeze resources as far as possible. If the CPU is only doing something 50% of the time, then halve the clock rate. Don't waste power on NO-OPs. You don't need to have a fully-powered network card to detect the presence of a signal - use the signal's own power for detection purposes. I'm not sure if arial types are mandated, but if not, three loop arials at 90' to each other (with only the arial with the strongest signal being in use) would reduce your power requirements for transmit significantly.
Ok, so you've used every trick in the book (and written three further volumes besides) to save power through hardware where at all humanly possible. What else can you do?
Reduce the workload the software needs to do. By freezing images and restoring them, you eliminate startup requirements on time and power. By compiling with profiling, then compiling using the profiling to improve optimization, you can reduce space requirements by a fair amount. Eliminate costly programs (such as X) and use KGI or Framebuffer instead to eliminate a lot of overhead on space, power and internal bandwidth.
So there's a mix of software and hardware solutions, but it is NEVER more software, and it is ESSENTIAL that the software and hardware is squeezed as much as possible.
Ouch! LinuxBIOS takes 5 seconds to boot a basic system, and it wouldn't take much to save a memory image of a booted GUI, which could then be pulled in through suspend2. True, this would probably require another chip, and space on these systems is always extremely tight, but if you can slash a 60 second boot time to under 10, there'll be a bigger market, even if the device is larger.
If your GUI can be crammed into the kernel, say by seriously mutilating KGI or framebuffers to provide everything you need directly, then have your desktop as an init replacement, you could even get away without a pre-booted image.
The next step after that would be to have the kernel run in non-volatile RAM, so that if the machine is turned off and back on, the system doesn't need to reboot at all. NVR is slower than volatile, so it would depend on whether the last few seconds of boot time were more obnoxious than the degradation in performance.
The power problem should be easy for them to fix - you simply enable CPU frequency scaling, then scale the speed of the CPU so that it never runs faster than the slowest speed that isn't painful for the applications you're running. So the more you run, the faster the device runs (up to maximum speed) so that performance remains fairly constant but doesn't waste power idling for the next task. My guess is that this would improve the lifespan of the battery by a fair amount.
If wireless is a power killer, then it might be wise to look at the wireless device for opportunities. For example, there's no point blasting out full power when the remote device is right next to you, so don't transmit more power than needed to get a working signal to the WAP. Although the received power wouldn't be enough to drive the network chips, it would be enough to drive a transistor to let the device know that a nearby signal is present (although it wouldn't be good enough to detect any signal within the normal range of the wireless device). There may also be more energy-efficient wireless NIC chips out there, which could save on power.
I imagine power is also why cryptography isn't really supported. There are plenty of crypto chips (Motorola's S1 looks good) and those will take less power than doing crypto in software. All it would take is adding the necessary hooks to OpenSSL's crypto engine support. Problem solved.
So if I'm so smart, why aren't I building these and making myself rich? Well, I don't have a few million dollars to burn. Nor do I have a red paperclip. (That guy HAS to be Ferrengi!)
What it means is that if a rootkit alters the internal IP tables for a Microsoft address, most virus checkers won't pick up on it (the Hosts file will be untouched) and it will be impossible for the user to override the problem in order to get to Microsoft's website to download the necessary patches.
...the fact that they decided it would offend non-Christians. This makes as much sense as banning an Odinist Longship-burning ceremony on the grounds that it might offend non-Baptists. Hmmm. There again, that might explain some things.
But it shouldn't be appointed, as it is in Europe or America. Quangos merely tell you what you've appointed them to say, which tells you nothing. Since people are generally the most intelligent between the ages of 24 and 30, it might be better if scientific advisors in each of the relevant fields were drawn at random from the pool of PhDs who have spent between two to five years in post-doctoral research, for a one-year post as advisor. Then they're at their peak of intellect, they're experienced in how to do the necessary work, and (by being both random and short-term) they owe no favours and can be independent in their advice.
Technocrats and intellectual elites are a Bad Idea, but utilizing the intellectual resources as far as possible for the benefit of the country is a Good Idea. My suggestion - almost a jury pool of advisors - is probably not the best option anyone can come up with, but it would seem better than the fiasco we have at present.
....can barely spell the word "environment". I wouldn't trust them to look after a flowerpot, let alone the nation. How would I get someone out into the country? Dunno about Britney, but I walk if I want to go cross-country. Ok, I can only do about 30 or 40 miles in a day and I'm not going anywhere much after that. (King Harold managed about 50 or 60, whilst wearing about a hundred pounds of armour, for three consecutive days, fighting a long drawn-out battle on either side. That's stamina.)
For the purpose of conveying a single person, with near-zero baggage, to and from work, a car is stupid and an SUV is moronic. Mass transit - if correctly designed - would be the "ideal" solution, but I'd certainly regard motorbikes that are efficient as vastly preferable to SUVs and hatchbacks.
(In fact, it would be great if, instead of having one HOV lane and the rest of the road to whatever's out there, they had two HOV, one bike lane, and if there's anything left, cars/trucks get squished into that. If people were adopting sound transport strategies, they wouldn't need a single-occupancy lane except for very rare situations and emergencies.)
Individual rights exist only at the expense of communal rights, so the only sound way of living is to balance these two as best you can. If individual rights trump everything, then no communal rights can exist at all and there will be no community and no individual can exist for long in utter, irrevocable isolation. If communal rights trump everything, then the individual ceases to exist and the community will stagnate, rot and die.
Common sense dictates that the middle way is the only long-term way, that extremes (of any kind) are degenerate and will fail. This fact was not discovered by a "tree hugger" (not totally sure what that means, as REAL environmentalists are way too busy to hug trees, and would consider it damaging to the micro ecosphere of the tree anyway) but was known for sitting around in the shade. It's an interesting worldview that politicos would likely profit from.
Common sense also dictates that venom should be reserved for the true offenders (and I'm more than happy to consider the current EPA as one of the worst offenders of all time). There seems little point and less sense in condemning those who may even agree with you, if there was any kind of meaningful communication. I tend to agree with a lot of different perspectives, because I work hard at finding out what is meant and why. (I also tend to be in a bunch of flamewars, because I also speak my mind whether it is popular or not, and my views tend not to be popular with large segments of society.)
I said that it was a matter of cleaner for the same work that mattered. If you want to get from A to B and can do so using less gas by using a motorbike than an SUV, then that motorbike is clearly cleaner for the same amount of work. Duh.
I have no problems with motorbikes. Well, I'm not completely convinced 2-stroke engines are capable of delivering the same power for the same fuel as a 4-stroke engine, if optimally designed, and I'm definitely not convinced either would be superior to a well-designed rotary engine, but that's for a highly optimized design. For right now, I'd be willing to believe that motorbikes have superior performance to a car.
I do not drive an SUV, I do not regard SUVs as suitable for any purpose whatsoever, and I'd regard any environmentalist who owned an SUV as brain-dead.
Yes, cars are too complex these days. This has nothing to do with efficiency, however. There is nothing efficient about a car that cannot be maintained. That should be obvious, since if it cannot be maintained, it cannot be kept at an optimum for the conditions the vehicle is subject to. A hyper-complex car can, at best, only be any good under absolutely average test conditions and nothing else.
The add-ons do nothing for efficiency within themselves, either. They aren't designed to. They're designed to add "value" to a car (read: money in the hands of the manufacturer) but really don't benefit anyone. Cruise control? Well, anything that can control Cruise can't be bad, but I digress. There are numerous reports of cruise controls in cars malfunctioning and activating in a manner that turned the car into an uncontrollable missile - even though there are supposed to be failsafes to prevent just that. The failsafes don't exist. There are also reports of car electronics sparking even when the car is switched off, resulting in houses being burned to the ground and a number of fatalities.
But what have these things added? Most fuel is used in stop-go traffic, not highway cruising. Such additions look "cool", have added nothing and have caused (fortunately only a few, but still far too many) fatalities.
Am I in favour of complexity? No. I am in favour of efficiency. Often, efficiency can be gained through simplicity. Not always - sometimes you need to add enough complexity to be able to do something well. To misquote one famous scientist, things should be as simple as possible, but no simpler.
Am I an environmentalist? I don't like the term, because I don't see things as being necessarily in conflict, and many of my beliefs that are friendly towards the environment are so because they are as simple as possible but no simpler. Those things that are good at achieving results with minimal effort must necessarily also achieve results with the minimum of impact. I like the idea of preserving as many species as possible, but that's just as much because diverse environments are stable AND offer the greatest possiblity of providing me with whatever I might need.
I despise those who believe things because it is their religion. That earns no respect from me. I DO respect those who believe things because they have established something to be true, and I definitely respect those who work hard on pushing their understanding to the limits. That is without regard to the beliefs, including whether I hold them myself. I reserve my greatest disgust and contempt, though, for those who turn meaningless labels (such as "environmentalism") into religious objects of worship or fanatical objects of hate. Such people have no feelings towards that which they profess to worship or hate - how could they, when they don't even know what they are? So not only are they claiming prophetic powers of received wisdom, they are also claiming the divine right of prejudice. Contempt is far better than such people deserve, but in the spirit of not wasting needlessly, I will waste nothing more on them.
First, you can use less energy by (as you say) doing less. I've already argued that this is a Bad Idea and also Needless, so I won't debate it further here, except to say that it it STILL a Bad Idea and STILL Needless.
Second, you can use less energy by being more efficient. This is my preferred option for many areas. Many devices have efficiencies in the 5-10% range, with the other 90-95% of the energy being completely wasted. To me, this leaves considerable scope for improvement. You get the same amount done, but with far less waste. If it turns out the waste is unimportant, you can then ramp up and do far more with the same input as you currently have. This doesn't involve broken windows, this involves preferring insulation to burning money.
Third, you could always adjust the needs. Hauling pig iron around the country at high speed, then letting it rust in some storage area, is wasteful of energy in two ways. You waste energy accelerating a large mass by more than is useful to you, then you waste energy reducing the iron oxide back to iron. If, on the other hand, you moved "non-perishable" raw materials much more slowly, with better containment, it would require much less energy to do so at all points along the way.
(If you've a full pipeline of canal barges, it makes no difference that they travel at a few miles a day, as you can add/remove exactly the same amount to the pipeline over the same time as you could by truck and your stockpile would be no different, the only difference is that you need no storage yards, you use almost no energy and it's vastly easier to protect a barge against the elements.)
Your "broken window" theory also assumes that you can directly compare R&D of efficient techniques (which add value to society and to business) with the activities of any non-contributing industry. Circulating money in a manner that does not add value will actually decrease value. (Money suffers from a form of entropy called inflation.) But the entire game changes when you invest in something that provides direct returns through improved scalability and reduced requirements.
Better scalability, better pipelining, lower energy requirements - sounds like a shopping list of things computer geeks are always asking Intel for, but really these are the things that will make or break ANY industry. It is in failing in these areas that doomed many British industries - the products themselves were generally superior to their American counterparts, but the processes used were too primitive, too neolithic, to survive any serious pressure.
Ah, when politicians (who know little about the subject) try to set standards, they usually do a very bad job of it. They assume that you can pencil in some magic number and everything will work out fine. That's easy to test (which is why they like it) but really achieves very little.
(The example I gave - 10% reduction in pollution, but with a 50% reduction in work done - would be a good example of an illusary improvement. We may have achieved the magic number, but only by almost doubling the overall pollution.)
Some things could be done. Car millage is currently based on a tiny random sample, and I believe it's only an arithmetic average (mean value), not a minimum or even the average by other measures (median and mode) and is a fixed value according to the car's class. If this is the path to be followed, a minimum (or variance) and a median should be specified as well. Because cars are NOT uniformly travelling at 55mph along perfectly smooth roads, you want to know the efficiency under a wide range of conditions, and you want to define what is within acceptable limits under all of those conditions.
This is NOT the work of politicians - this is an area for experts in mechanical engineering, chemistry, physics and perhaps even civil engineering (since we'd want to know how the variables we're operating in change over time). However committees rarely design anything useful, so they should be required to actually produce proofs of concept that can achieve what they claim is achievable, within the constraints of what they claim is required to achieve it.
Politicians should then be handed a document outlining what is typically achieved (which should be the minimum standards) and what could be realistically achieved (which should be the standards at which manufacturers would be rewarded somehow). The politicians should be limited to signing off on the work of those who know what they're doing.
Practical examples and demonstration implementations should make for laws that corporations would WANT to follow, because if their competitors do so before them, their competitors will take away their business.
...I still do not see how slowing emissions could slow the economy, as the necessary R&D to produce cleaner AND more efficient technology would necessarily pour billions into all sectors of industry and academia. (Cleaner alone says nothing, as it is the ratio of pollution to work done that matters, if you're going to get the same amount of work done overall.)
You are correct that this is not political. Either you're investing in better management of resources and waste, or you are not. There is nothing inherently left-wing or right-wing about superior management technques. Sure, certain groups emphasise certain aspects of the issue, but the groups cover the whole political spectrum. All that happens is that each group emphasises the issues that are important to them. This is a Good Thing, overall, but can create the illusion of politics where none exist.
...if the consent is genuine. Not everyone physically over 18 is automatically intellectually and emotionally also over 18, but these are almost impossible for Joe/Jane Average to quantify in person, and totally impossible under isolating conditions such as a photograph or a movie.
That's also assuming that there's no duress involved (Linda Lovelace claimed to have worked at gunpoint, and the UN reports that slavery is on the increase in the US) and that the alternatives are viable (minimum wage is less than half of what's needed for a single person to actually live). One of the problems is that the education needed to get a job worthy of the name is expensive, even though well-paid jobs pay the Government through extra income tax.
It's impossible for anyone to be 100% sure that consent has been given, but if you are confident that it has (within whatever threshold you set for yourself) and that it is likewise genuine, then I can see nothing wrong.
I would add one other proviso - if fMRI scans subsequently prove that there is permanent harm to those sections of the brain involving judgement, then you would also need to give yourself consent to cause that harm, in full knowledge. At this time, no such harm has been proven - all studies have been statistical in nature and no study of actual mechanisms have been bothered with. (Both sides seem to prefer clearly defined regions of doubt and uncertainty, suggesting that most studies are actually being carried out by philosophical white mice.)
Note that this last bit requires a demonstration of actual harm, BUT ALSO allows all individuals concerned the right to override this through personal choice.
Freedom of choice is only meaningful if it allows you to choose something someone else might not like. But it is also only meaningful if you are free to choose, and have a working knowledge of what the choices are.
...it depends on how you define proof. Traditional scientific proof isn't the same as a rigorous mathematical proof. They have differing requirements and differing standards.
However, there ARE other kinds of proof, that is also mathematical in nature, but aren't generally covered in these discussions - one being the existance/non-existance proof. This kind of proof establishes if an answer exists (or does not exist) to a specific question, without attempting to actually produce any examples. It is a useful technique when you only want to know if a counter-example must exist, even if there is no actual way of discovering it.
Let us take the theory of evolution. Evolution requires a series of random mutations to accumulate, eventually diversifying species. This has a number of obvious requirements, the first of which is that for all species that exist or have ever existed, there must exist (or have existed) a one and only one other species, where the genetic variation within either does not overlap with the genetic variation of the other, and where a sequence of mutations can be defined such that no step in the sequence involves changes in excess of those that naturally occur between generations and all steps produce viable genetic code for an organism with no significant disadvantage over others within the species, and where independent evidence exists that the two genetic sequences are directly related to each other.
Those are some very tough requirements and it's not at all clear you could solve these requirements for ANY species that currently exist. However, the problem is not one of solving the requirements. This is an existance proof, not a solution. So long as we can show that a solution must exist, we don't actually need to know what the solution is.
A second kind of proof is proof by induction. A proof by induction requires that we prove that a specific step occurs and that a given initial condition occurs. Everything else then follows. Again, we use the above argument. We start by saying that a species forks from another species, according to evolution. Every step along that fork must be valid, or it would have died off. We can use the initial species as our first step, then prove that a step will exist such that it mutates the organism without impairing it. From this, it would be possible to formulate (for any given branch) a proof by induction of the entire branch. Once you have done that for an arbritary branch, you can use that as a proof of the general mechanism and prove by induction that the entire tree must also be valid (from a given starting point).
Relationships between branches is also important, as this eliminates the possibility of parallel evolution, isolates or independent creation by any other means. It's also one of the hardest problems. There is no evidence from mDNA earlier that "mitochondrial Eve", even though there were plenty of humans existing before then. What you'd want is a species where individual lifetimes were vastly longer AND where the number of generations was small enough that extremely distinct genetic markers can be tracked.
If you can do that - ie: show that there exists a set of genetic markers that come from a common genetic family, not merely an entire species - you've established a much more definite relationship between species A and species B. The genetic code is not merely similar, it shows strong tell-tale signs of a direct family relationship.
I don't know if these options are really practical. I don't know if we have enough genetic data to know. It seems to me that we have far far too little genetic data about the world to know much of anything. But I do believe that - eventually - we will indeed have a clear-cut proof of evolution that goes so far beyond reasonable doubt that even the unreasonable doubters would have a hard time of it.
They run FreeBSD, which rarely needs any kind of patches. Also, patching should cause a fairly even distribution of downtimes, but the description suggests it was much more random.
But I've known 2-3 DAY delays in receiving mail. That, to me, is amazingly bad. Yeah, yeah, they're free. So what? If they're badly configured, then they won't run, no matter how much Yahoo was paid, but if they're well-configured, then they'll run, no matter how little Yahoo gets from them.
Load in Linux + LinuxBIOS + a watchdog, with a copy of Postfix, for the mail handling, and you should not be getting a 50% downtime. It should be closer to 0.05% downtime, even under the kinds of loads described. (I'm including LinuxBIOS here because you can reboot with it rapidly. The watchdog is for auto-reboot on crash. The longest any machine would remain down with such a configuration is 35 seconds.)
To me, the "obvious" candidates for solutions would include investing in R&D for more efficient internal combustion engines and more efficient industrial methods. (Efficient is important. Reducing emissions is irrelevent if you also reduce output, especially if the reduction is disproportionate the wrong way. You still need the same total output, so if one company reduces emissions by 10% by reducing output by 50%, you actually worsen the problem by 180%.)
If it turns out that there's a problem, you've got to rduce the NET emissions, while still maintaining output. Ideally, you also want to equal or increase profits in the proces. This should not be a problem, because you have to pay for the waste you generate (even if it's nothing more advanced than maintaining a venting system). It is not profitable to create pollution. Nobody pays you to do that. The more useful work you can do, then, with the same resources, the more money you will make.
And this is where I differ from the political hacks. A CORRECT solution won't ruin the economy, it'll bolster it. The R&D alone would inject vast amounts of money into the system, and the improved productivity that could potentially result would sustain those profits for a long time.
Now, if it turns out that there's no problem at all, then the above solution is even better, as it improves the scalability of everything out there. Better scalability means greater resiliance against economic downturns, faster reaction to changing requirements, greater profitability, cheaper products and greater availability.
The R&D would need to be Government-funded, but as I'm expecting it to result in a massive upswing in the economy, they'd get paid back very rapidly through the taxes it would bring in. Were the Government to go to each of the key industries and offer substantial contracts to invent whole new replacement processes and mechanisms for key polluters, we could be seeing returns within a few years.
This approach is similar to President Bush's advocacy of letting industry deal with the problem, but it is much more aggressive and it provides industry with both means and motivation. You can't get solutions by wishing for them, but you CAN get solutions by both paying for them and pushing for them.
As for the research into global warming, you are correct that we have insufficient data. We do need much more data, we also need much better computer simulations, and we therefore need much better computers to run those simulations on. Because of the chaotic nature of the system, no amount of data will be "sufficient", but we do nonetheless need far more than we have. (We have next to no information for the south pole, for example, where air flows are demonstrably 100% different from the assumptions made in most simulations.)
This is an extremely complex subject. I would add one further point. The climate is a chaotic system. This is not a trivial thing, because it means that you have massive discontinuities and instabilities at all points.
Put into simple terms, a rise of N degrees between two points in time tells you next to nothing about what happens over any other interval. A small rise in temperature one day may trigger off a series of events which lead to a massive plunge in temperature the next. A tiny drop in temperature may, likewise, lead to temperatures soaring a short time later.
(Chaos mathematicians refer to this as sensitivity to initial conditions and often talk in terms of Lorenz Waterwheels and African butterflies. What they really mean is that predictions are extremely hard and usually wrong, and that the consequences can be sudden, violent and unexpected in nature, but there again, they might not, and you probably won't know in advance which it's going to be.)
...for MIT to employ professors who put petty party politics over and above their own discipline. A scientific field is called a discipline for a reason, and those who claim eminent status within it are expected to maintain a level of professionalism that is clearly far beyond the mental capacity of this professor.
To those who think this is over-the-top, please remember that I'm from England - the country that brought you Oxford University in 800 AD, and Cambridge in 1100 AD, where there is a natural cultural antipathy towards those sordid enough to bring politics into academia.
You also have to bear in mind that I hold the VERY passionate belief that academics can only be free to study when their study has no hidden agenda, no political motivation and no political reward. I believe utterly that any educational facility where these are not adhered to utterly are doomed to drag their subjects into disrepute and their students into quagmires.
A title - such as that of "professor" - is NOT an excuse, it is a responsibility. To deliberately flout that responsibility in a vainglorious effort to pursue money or influence in political circles is sordid and contemptible.
However, I find it odd that the claims of suppression are coming from an institute in a country noted solely for its dissent over Global Warming, where NASA even had to write policy documents to prohibit religious propoganda being included in statements. I also find it curious that these complaints are surfacing AFTER the scandal at NASA, rather than before, when people might have been more sympathetic.
No, I question the credibility and the timing of these claims, and I find it disgraceful of MIT to be associated with what appears to be little more than a political stunt. Note I said "appears". There may well be some basis to the claims, but the timing and nature of their presentation destroys all of that. If you want to be taken seriously, you can't come across as a spoiled brat whose toy - possibly NASA - has been taken away from you.
The claims against Global Warming may or may not have any validity. That isn't being discussed, so I'm not going to address that here. What I am going to address is efforts to turn the debate into propoganda - by whatever side. Whether you're talking about the Swift Veterans For Toothfairies, or some other totally political movement, is also unimportant. Reality cares not one whit for opinion polls or campaign financing. If the climate is shot, it's shot, and all the PR in the world won't change that. If it isn't, it isn't, and again PR isn't worth a damn.
Politics has no business meddling in the affairs of science. The reverse is not true, as science has the knowledge necessary to define politics. But politicians should learn their place - they are the servants of the public and the slaves of the forces of nature. Allowing politics to control anything is the ultimate recipe for disaster.
Slashdot Mirror
Another option is to package the values up into blocks and use a cryptographically authenticating block chaining method. That way, if someone attempts to build their own chain, the chaining method will fail to authenticate correctly. That way, even if a MitM found a way to block reception of the original OTP, the recipient would still know that the key had been intercepted.
Could this be reduced to the point where the general user would be satisfied? Well, I believe it to be already superior to many other popular general-purpose systems. There's some lack in third-party applications, X is still horrible for abstract graphics and distros aren't merging in drivers that are necessary (eg: madwifi), but those are deficiencies in imagination, not bugs in the code.
Nonetheless, there are bugs in the software, and yes these can be fixed. If there's a code audit, a random number of bugs will be found on each sweep, so the time taken to fix all of them will be indeterminate. The correctness of any given fix will also be indeterminate. If we propose an average rate of 1 bugfix every 3 days (this includes time for tracking down even the most obscure of bugs, testing the fix, and having it filter through the usual chain of developers), it would take about twenty thousand programmers a decade to reduce the number of bugs to the absolute minimum - I'm guessing that this would be in the low hundreds. (As the code is being developed, new bugs will be being added, so there will come a time when the bugs being added by the core developers will exceed the bugs being developed by this hypothetical army.)
A more rigorous method would be to use formal software methods to validate each function is doing what is intended. You could then be sure no new bugs were being added by the bug-stompers. However, because ANY change would totally trash the calculations being done, you would need to fix entire functions at a time, regardless of how many bugs there were. I'd say that it would take the same sized mob about the same length of time, but with greater certainty as to what was fixed and what was left to be fixed.
By this logic, Microsoft or IBM could have fixed their flagship Operating Systems to the point of being pretty nearly invulnerable to defects of any kind. They've had the manpower, the time and the money to perform the kind of total code review suggested. They have clearly not done so, suggesting that the marketplace is largely unaffected by the number of defects in software, but is far more influenced by application availability. Where applications are equal, then the fight is decided by brand.
Is there anything that can independently confirm this suspicion? Yes. The BBC Microcomputer had branding and software. The Acorn Business Computer had neither, even though it was technically made by the same people. OS/2 and NT4 were built from the same underlying codebase, but NT had the applications. Guess which won. The history of computing is littered with such examples.
This is why I am always skeptical of the argument that large numbers of bugs are inevitable. I believe it is closer to the truth that users don't give a damn (except when they lose data, and then they blame the computer, the dog, the telephone - anything but the software). I also believe that computer companies know that the users don't care, and therefore lower their standards to meet those expectations.
At present, Coverity's software seems to indicate that Open Source is down to the low hundreds of bugs per million lines of code, for those classes of bug their software can detect - which is some unknown fraction of the number of possible bug types that could exist. (I'd guess 10%, which would mean there's about a thousand bugs per million lines of code.) I see no reason why, within a decade, we could not be seeing reports of Open Source having fewer than a hundred bugs per trillion lines of code - other than nobody wanting to invest in that much clean-up when almost none of the mainstream users
The one big vulnerability with OTPs is that you've now got to send the key securely. Since it is equal in size to the message and is only valid for one message, it is equally hard to send the key securely as it is to send the message securely. Because the pad is pure randomness, it is possible (using existing methods) to send the pad by public key encryption, as it is non-trivial for someone intercepting the message to know how to decrypt it, as it's hard to know when you've broken the encryption. One piece of randomness looks much like another.
Generally, though, people take shortcuts. Instead of using a full-sized one-time pad, a much smaller, repeatedly-used pad is used instead, with some form of pseudo-random mangling to churn things up so that it acts in a very similar manner to a one-time pad. This is generally how stream ciphers work.
Quantum Cryptography - if used sensibly - would involve transmitting a gigantic OTP. Far bigger than the one you need. You then drop all of the bytes that are intercepted. The only bytes used in the pad are the ones the intercepting person does NOT have, so you know the pad is free of holes.
A "better" solution would be to not transmit the key at all, but somehow exploit photon teleportation to deliver the key in a secure manner. However, if you could do that, you wouldn't need encryption in the first place.
My point with Linux was definitely not to claim Linux was better for that platform - it isn't. It was half intended to be vaguely humerous and half intended to provoke any Linux user reading it into wondering just what else is out there in the way of extensions and capabilities that they don't know about. It's too easy to assume that the mainstream kernel (or the one that comes with the distro) is all there is, when really it's only the leading edge of the very first part of the beginning of what's available.
(I wouldn't presume to do that for NetBSD, but only because I don't know of any extensions for it. If I did, and I had enough familiarity with that kernel, I'd probably be looking to stretch a few mental muscles there as well.)
I can run Linux on a VAX, too!
In practice, nobody writes complex programs purely in hardware any more. It's much more power-efficient, it is usually faster, but there's a price to pay. It takes more space and it's often impossible to update. In consequence, the optimal solution is a compromise - not realy "pure" hardware solutions, but still "mostly" hardware with a bunch of programmable registers and gates.
Even then, this won't do everything you'd want. You still need some software. But it should be kept to an absolute minimum, for a low-power device. Software is cheap on space, but costly on power. It is also dog-slow, which means that you need faster (and therefore more power-hungry) CPUs just to get the same performance.
Ok, given that, what needs to be done to wring as much performance as possible, for as little power? Offload engines for CPU-intensive work (such as cryptography) is an obvious must. You also really want offload engines for anything that is used frequently enough that even a small saving would make a huge impact. A good memory manager would be useful, by that reasoning.
You also want to use every trick in the book to squeeze resources as far as possible. If the CPU is only doing something 50% of the time, then halve the clock rate. Don't waste power on NO-OPs. You don't need to have a fully-powered network card to detect the presence of a signal - use the signal's own power for detection purposes. I'm not sure if arial types are mandated, but if not, three loop arials at 90' to each other (with only the arial with the strongest signal being in use) would reduce your power requirements for transmit significantly.
Ok, so you've used every trick in the book (and written three further volumes besides) to save power through hardware where at all humanly possible. What else can you do?
Reduce the workload the software needs to do. By freezing images and restoring them, you eliminate startup requirements on time and power. By compiling with profiling, then compiling using the profiling to improve optimization, you can reduce space requirements by a fair amount. Eliminate costly programs (such as X) and use KGI or Framebuffer instead to eliminate a lot of overhead on space, power and internal bandwidth.
So there's a mix of software and hardware solutions, but it is NEVER more software, and it is ESSENTIAL that the software and hardware is squeezed as much as possible.
If your GUI can be crammed into the kernel, say by seriously mutilating KGI or framebuffers to provide everything you need directly, then have your desktop as an init replacement, you could even get away without a pre-booted image.
The next step after that would be to have the kernel run in non-volatile RAM, so that if the machine is turned off and back on, the system doesn't need to reboot at all. NVR is slower than volatile, so it would depend on whether the last few seconds of boot time were more obnoxious than the degradation in performance.
The power problem should be easy for them to fix - you simply enable CPU frequency scaling, then scale the speed of the CPU so that it never runs faster than the slowest speed that isn't painful for the applications you're running. So the more you run, the faster the device runs (up to maximum speed) so that performance remains fairly constant but doesn't waste power idling for the next task. My guess is that this would improve the lifespan of the battery by a fair amount.
If wireless is a power killer, then it might be wise to look at the wireless device for opportunities. For example, there's no point blasting out full power when the remote device is right next to you, so don't transmit more power than needed to get a working signal to the WAP. Although the received power wouldn't be enough to drive the network chips, it would be enough to drive a transistor to let the device know that a nearby signal is present (although it wouldn't be good enough to detect any signal within the normal range of the wireless device). There may also be more energy-efficient wireless NIC chips out there, which could save on power.
I imagine power is also why cryptography isn't really supported. There are plenty of crypto chips (Motorola's S1 looks good) and those will take less power than doing crypto in software. All it would take is adding the necessary hooks to OpenSSL's crypto engine support. Problem solved.
So if I'm so smart, why aren't I building these and making myself rich? Well, I don't have a few million dollars to burn. Nor do I have a red paperclip. (That guy HAS to be Ferrengi!)
...an infinite number of dumbasses replying to each other will eventually produce the complete works of USENET.
What it means is that if a rootkit alters the internal IP tables for a Microsoft address, most virus checkers won't pick up on it (the Hosts file will be untouched) and it will be impossible for the user to override the problem in order to get to Microsoft's website to download the necessary patches.
The BBC has a recording of an Anglo-Saxon speech that would have gone down well in many churches this morning... if they'd understood a word of it.
...the fact that they decided it would offend non-Christians. This makes as much sense as banning an Odinist Longship-burning ceremony on the grounds that it might offend non-Baptists. Hmmm. There again, that might explain some things.
Technocrats and intellectual elites are a Bad Idea, but utilizing the intellectual resources as far as possible for the benefit of the country is a Good Idea. My suggestion - almost a jury pool of advisors - is probably not the best option anyone can come up with, but it would seem better than the fiasco we have at present.
For the purpose of conveying a single person, with near-zero baggage, to and from work, a car is stupid and an SUV is moronic. Mass transit - if correctly designed - would be the "ideal" solution, but I'd certainly regard motorbikes that are efficient as vastly preferable to SUVs and hatchbacks.
(In fact, it would be great if, instead of having one HOV lane and the rest of the road to whatever's out there, they had two HOV, one bike lane, and if there's anything left, cars/trucks get squished into that. If people were adopting sound transport strategies, they wouldn't need a single-occupancy lane except for very rare situations and emergencies.)
Individual rights exist only at the expense of communal rights, so the only sound way of living is to balance these two as best you can. If individual rights trump everything, then no communal rights can exist at all and there will be no community and no individual can exist for long in utter, irrevocable isolation. If communal rights trump everything, then the individual ceases to exist and the community will stagnate, rot and die.
Common sense dictates that the middle way is the only long-term way, that extremes (of any kind) are degenerate and will fail. This fact was not discovered by a "tree hugger" (not totally sure what that means, as REAL environmentalists are way too busy to hug trees, and would consider it damaging to the micro ecosphere of the tree anyway) but was known for sitting around in the shade. It's an interesting worldview that politicos would likely profit from.
Common sense also dictates that venom should be reserved for the true offenders (and I'm more than happy to consider the current EPA as one of the worst offenders of all time). There seems little point and less sense in condemning those who may even agree with you, if there was any kind of meaningful communication. I tend to agree with a lot of different perspectives, because I work hard at finding out what is meant and why. (I also tend to be in a bunch of flamewars, because I also speak my mind whether it is popular or not, and my views tend not to be popular with large segments of society.)
I have no problems with motorbikes. Well, I'm not completely convinced 2-stroke engines are capable of delivering the same power for the same fuel as a 4-stroke engine, if optimally designed, and I'm definitely not convinced either would be superior to a well-designed rotary engine, but that's for a highly optimized design. For right now, I'd be willing to believe that motorbikes have superior performance to a car.
I do not drive an SUV, I do not regard SUVs as suitable for any purpose whatsoever, and I'd regard any environmentalist who owned an SUV as brain-dead.
Yes, cars are too complex these days. This has nothing to do with efficiency, however. There is nothing efficient about a car that cannot be maintained. That should be obvious, since if it cannot be maintained, it cannot be kept at an optimum for the conditions the vehicle is subject to. A hyper-complex car can, at best, only be any good under absolutely average test conditions and nothing else.
The add-ons do nothing for efficiency within themselves, either. They aren't designed to. They're designed to add "value" to a car (read: money in the hands of the manufacturer) but really don't benefit anyone. Cruise control? Well, anything that can control Cruise can't be bad, but I digress. There are numerous reports of cruise controls in cars malfunctioning and activating in a manner that turned the car into an uncontrollable missile - even though there are supposed to be failsafes to prevent just that. The failsafes don't exist. There are also reports of car electronics sparking even when the car is switched off, resulting in houses being burned to the ground and a number of fatalities.
But what have these things added? Most fuel is used in stop-go traffic, not highway cruising. Such additions look "cool", have added nothing and have caused (fortunately only a few, but still far too many) fatalities.
Am I in favour of complexity? No. I am in favour of efficiency. Often, efficiency can be gained through simplicity. Not always - sometimes you need to add enough complexity to be able to do something well. To misquote one famous scientist, things should be as simple as possible, but no simpler.
Am I an environmentalist? I don't like the term, because I don't see things as being necessarily in conflict, and many of my beliefs that are friendly towards the environment are so because they are as simple as possible but no simpler. Those things that are good at achieving results with minimal effort must necessarily also achieve results with the minimum of impact. I like the idea of preserving as many species as possible, but that's just as much because diverse environments are stable AND offer the greatest possiblity of providing me with whatever I might need.
I despise those who believe things because it is their religion. That earns no respect from me. I DO respect those who believe things because they have established something to be true, and I definitely respect those who work hard on pushing their understanding to the limits. That is without regard to the beliefs, including whether I hold them myself. I reserve my greatest disgust and contempt, though, for those who turn meaningless labels (such as "environmentalism") into religious objects of worship or fanatical objects of hate. Such people have no feelings towards that which they profess to worship or hate - how could they, when they don't even know what they are? So not only are they claiming prophetic powers of received wisdom, they are also claiming the divine right of prejudice. Contempt is far better than such people deserve, but in the spirit of not wasting needlessly, I will waste nothing more on them.
Second, you can use less energy by being more efficient. This is my preferred option for many areas. Many devices have efficiencies in the 5-10% range, with the other 90-95% of the energy being completely wasted. To me, this leaves considerable scope for improvement. You get the same amount done, but with far less waste. If it turns out the waste is unimportant, you can then ramp up and do far more with the same input as you currently have. This doesn't involve broken windows, this involves preferring insulation to burning money.
Third, you could always adjust the needs. Hauling pig iron around the country at high speed, then letting it rust in some storage area, is wasteful of energy in two ways. You waste energy accelerating a large mass by more than is useful to you, then you waste energy reducing the iron oxide back to iron. If, on the other hand, you moved "non-perishable" raw materials much more slowly, with better containment, it would require much less energy to do so at all points along the way.
(If you've a full pipeline of canal barges, it makes no difference that they travel at a few miles a day, as you can add/remove exactly the same amount to the pipeline over the same time as you could by truck and your stockpile would be no different, the only difference is that you need no storage yards, you use almost no energy and it's vastly easier to protect a barge against the elements.)
Your "broken window" theory also assumes that you can directly compare R&D of efficient techniques (which add value to society and to business) with the activities of any non-contributing industry. Circulating money in a manner that does not add value will actually decrease value. (Money suffers from a form of entropy called inflation.) But the entire game changes when you invest in something that provides direct returns through improved scalability and reduced requirements.
Better scalability, better pipelining, lower energy requirements - sounds like a shopping list of things computer geeks are always asking Intel for, but really these are the things that will make or break ANY industry. It is in failing in these areas that doomed many British industries - the products themselves were generally superior to their American counterparts, but the processes used were too primitive, too neolithic, to survive any serious pressure.
(The example I gave - 10% reduction in pollution, but with a 50% reduction in work done - would be a good example of an illusary improvement. We may have achieved the magic number, but only by almost doubling the overall pollution.)
Some things could be done. Car millage is currently based on a tiny random sample, and I believe it's only an arithmetic average (mean value), not a minimum or even the average by other measures (median and mode) and is a fixed value according to the car's class. If this is the path to be followed, a minimum (or variance) and a median should be specified as well. Because cars are NOT uniformly travelling at 55mph along perfectly smooth roads, you want to know the efficiency under a wide range of conditions, and you want to define what is within acceptable limits under all of those conditions.
This is NOT the work of politicians - this is an area for experts in mechanical engineering, chemistry, physics and perhaps even civil engineering (since we'd want to know how the variables we're operating in change over time). However committees rarely design anything useful, so they should be required to actually produce proofs of concept that can achieve what they claim is achievable, within the constraints of what they claim is required to achieve it.
Politicians should then be handed a document outlining what is typically achieved (which should be the minimum standards) and what could be realistically achieved (which should be the standards at which manufacturers would be rewarded somehow). The politicians should be limited to signing off on the work of those who know what they're doing.
Practical examples and demonstration implementations should make for laws that corporations would WANT to follow, because if their competitors do so before them, their competitors will take away their business.
You are correct that this is not political. Either you're investing in better management of resources and waste, or you are not. There is nothing inherently left-wing or right-wing about superior management technques. Sure, certain groups emphasise certain aspects of the issue, but the groups cover the whole political spectrum. All that happens is that each group emphasises the issues that are important to them. This is a Good Thing, overall, but can create the illusion of politics where none exist.
That's also assuming that there's no duress involved (Linda Lovelace claimed to have worked at gunpoint, and the UN reports that slavery is on the increase in the US) and that the alternatives are viable (minimum wage is less than half of what's needed for a single person to actually live). One of the problems is that the education needed to get a job worthy of the name is expensive, even though well-paid jobs pay the Government through extra income tax.
It's impossible for anyone to be 100% sure that consent has been given, but if you are confident that it has (within whatever threshold you set for yourself) and that it is likewise genuine, then I can see nothing wrong.
I would add one other proviso - if fMRI scans subsequently prove that there is permanent harm to those sections of the brain involving judgement, then you would also need to give yourself consent to cause that harm, in full knowledge. At this time, no such harm has been proven - all studies have been statistical in nature and no study of actual mechanisms have been bothered with. (Both sides seem to prefer clearly defined regions of doubt and uncertainty, suggesting that most studies are actually being carried out by philosophical white mice.)
Note that this last bit requires a demonstration of actual harm, BUT ALSO allows all individuals concerned the right to override this through personal choice.
Freedom of choice is only meaningful if it allows you to choose something someone else might not like. But it is also only meaningful if you are free to choose, and have a working knowledge of what the choices are.
However, there ARE other kinds of proof, that is also mathematical in nature, but aren't generally covered in these discussions - one being the existance/non-existance proof. This kind of proof establishes if an answer exists (or does not exist) to a specific question, without attempting to actually produce any examples. It is a useful technique when you only want to know if a counter-example must exist, even if there is no actual way of discovering it.
Let us take the theory of evolution. Evolution requires a series of random mutations to accumulate, eventually diversifying species. This has a number of obvious requirements, the first of which is that for all species that exist or have ever existed, there must exist (or have existed) a one and only one other species, where the genetic variation within either does not overlap with the genetic variation of the other, and where a sequence of mutations can be defined such that no step in the sequence involves changes in excess of those that naturally occur between generations and all steps produce viable genetic code for an organism with no significant disadvantage over others within the species, and where independent evidence exists that the two genetic sequences are directly related to each other.
Those are some very tough requirements and it's not at all clear you could solve these requirements for ANY species that currently exist. However, the problem is not one of solving the requirements. This is an existance proof, not a solution. So long as we can show that a solution must exist, we don't actually need to know what the solution is.
A second kind of proof is proof by induction. A proof by induction requires that we prove that a specific step occurs and that a given initial condition occurs. Everything else then follows. Again, we use the above argument. We start by saying that a species forks from another species, according to evolution. Every step along that fork must be valid, or it would have died off. We can use the initial species as our first step, then prove that a step will exist such that it mutates the organism without impairing it. From this, it would be possible to formulate (for any given branch) a proof by induction of the entire branch. Once you have done that for an arbritary branch, you can use that as a proof of the general mechanism and prove by induction that the entire tree must also be valid (from a given starting point).
Relationships between branches is also important, as this eliminates the possibility of parallel evolution, isolates or independent creation by any other means. It's also one of the hardest problems. There is no evidence from mDNA earlier that "mitochondrial Eve", even though there were plenty of humans existing before then. What you'd want is a species where individual lifetimes were vastly longer AND where the number of generations was small enough that extremely distinct genetic markers can be tracked.
If you can do that - ie: show that there exists a set of genetic markers that come from a common genetic family, not merely an entire species - you've established a much more definite relationship between species A and species B. The genetic code is not merely similar, it shows strong tell-tale signs of a direct family relationship.
I don't know if these options are really practical. I don't know if we have enough genetic data to know. It seems to me that we have far far too little genetic data about the world to know much of anything. But I do believe that - eventually - we will indeed have a clear-cut proof of evolution that goes so far beyond reasonable doubt that even the unreasonable doubters would have a hard time of it.
They run FreeBSD, which rarely needs any kind of patches. Also, patching should cause a fairly even distribution of downtimes, but the description suggests it was much more random.
Load in Linux + LinuxBIOS + a watchdog, with a copy of Postfix, for the mail handling, and you should not be getting a 50% downtime. It should be closer to 0.05% downtime, even under the kinds of loads described. (I'm including LinuxBIOS here because you can reboot with it rapidly. The watchdog is for auto-reboot on crash. The longest any machine would remain down with such a configuration is 35 seconds.)
If it turns out that there's a problem, you've got to rduce the NET emissions, while still maintaining output. Ideally, you also want to equal or increase profits in the proces. This should not be a problem, because you have to pay for the waste you generate (even if it's nothing more advanced than maintaining a venting system). It is not profitable to create pollution. Nobody pays you to do that. The more useful work you can do, then, with the same resources, the more money you will make.
And this is where I differ from the political hacks. A CORRECT solution won't ruin the economy, it'll bolster it. The R&D alone would inject vast amounts of money into the system, and the improved productivity that could potentially result would sustain those profits for a long time.
Now, if it turns out that there's no problem at all, then the above solution is even better, as it improves the scalability of everything out there. Better scalability means greater resiliance against economic downturns, faster reaction to changing requirements, greater profitability, cheaper products and greater availability.
The R&D would need to be Government-funded, but as I'm expecting it to result in a massive upswing in the economy, they'd get paid back very rapidly through the taxes it would bring in. Were the Government to go to each of the key industries and offer substantial contracts to invent whole new replacement processes and mechanisms for key polluters, we could be seeing returns within a few years.
This approach is similar to President Bush's advocacy of letting industry deal with the problem, but it is much more aggressive and it provides industry with both means and motivation. You can't get solutions by wishing for them, but you CAN get solutions by both paying for them and pushing for them.
As for the research into global warming, you are correct that we have insufficient data. We do need much more data, we also need much better computer simulations, and we therefore need much better computers to run those simulations on. Because of the chaotic nature of the system, no amount of data will be "sufficient", but we do nonetheless need far more than we have. (We have next to no information for the south pole, for example, where air flows are demonstrably 100% different from the assumptions made in most simulations.)
Put into simple terms, a rise of N degrees between two points in time tells you next to nothing about what happens over any other interval. A small rise in temperature one day may trigger off a series of events which lead to a massive plunge in temperature the next. A tiny drop in temperature may, likewise, lead to temperatures soaring a short time later.
(Chaos mathematicians refer to this as sensitivity to initial conditions and often talk in terms of Lorenz Waterwheels and African butterflies. What they really mean is that predictions are extremely hard and usually wrong, and that the consequences can be sudden, violent and unexpected in nature, but there again, they might not, and you probably won't know in advance which it's going to be.)
To those who think this is over-the-top, please remember that I'm from England - the country that brought you Oxford University in 800 AD, and Cambridge in 1100 AD, where there is a natural cultural antipathy towards those sordid enough to bring politics into academia.
You also have to bear in mind that I hold the VERY passionate belief that academics can only be free to study when their study has no hidden agenda, no political motivation and no political reward. I believe utterly that any educational facility where these are not adhered to utterly are doomed to drag their subjects into disrepute and their students into quagmires.
A title - such as that of "professor" - is NOT an excuse, it is a responsibility. To deliberately flout that responsibility in a vainglorious effort to pursue money or influence in political circles is sordid and contemptible.
No, I question the credibility and the timing of these claims, and I find it disgraceful of MIT to be associated with what appears to be little more than a political stunt. Note I said "appears". There may well be some basis to the claims, but the timing and nature of their presentation destroys all of that. If you want to be taken seriously, you can't come across as a spoiled brat whose toy - possibly NASA - has been taken away from you.
The claims against Global Warming may or may not have any validity. That isn't being discussed, so I'm not going to address that here. What I am going to address is efforts to turn the debate into propoganda - by whatever side. Whether you're talking about the Swift Veterans For Toothfairies, or some other totally political movement, is also unimportant. Reality cares not one whit for opinion polls or campaign financing. If the climate is shot, it's shot, and all the PR in the world won't change that. If it isn't, it isn't, and again PR isn't worth a damn.
Politics has no business meddling in the affairs of science. The reverse is not true, as science has the knowledge necessary to define politics. But politicians should learn their place - they are the servants of the public and the slaves of the forces of nature. Allowing politics to control anything is the ultimate recipe for disaster.