Rockets are definitely not the answer, unless Werner Von Braun is in the question. Part of the problem comes back to the fact that ALL space scientists today were growing up in the 50s and 60s, and ALL sci-fi in the 50s and 60s was rocket-based. It's much harder to think outside of the box when the box is the size of the technological world.
The sole question, and I'm pretty sure you agree with me on this, is what to replace the rocket system with. My personal preference is to push R&D in every possible direction. Ideas that don't work will fall off the edge on their own, leaving the best ones standing. That way, the pick is based on what works, not what someone thinks should work.
Let us say that you are assigned a public encryption key anonymously, and the server has a dictionary of all valid private encryption keys.
You enter your vote and encrypt it anonymously. (It would need salting, to prevent trivial breech of anonymity by theft of the encryption key.) The encryption key would then need to be shredded or otherwise destroyed, if total anonymity was to be preserved. For nearly-total anonymity, you don't have to do this and you get the benefit of validation.
The server, on counting, then runs through the list of decryption keys for each vote. When it finds a key that decrypts to a valid vote, it records the vote and disposes of the key.
The total number of votes that should be recorded will be equal to the total number of keys minus the number of keys still left after all votes are counted.
An encrypted vote for which no decryption key exists (which will occur if two or more votes exist for the same key, OR if a vote has been submitted for an unlisted key) is then counted as a "spoiled ballot". In the case of duplicate votes, it would then be possible to use the decryption keys to determine the prevelence of voter fraud in a given district.
Since you can hand copies of the encrypted votes and decryption keys to anyone, without fear of threatening anonymity or risking voter fraud, it would be possible for others to independently verify the tallies.
Assuming near-total anonymity is fine, then the voter can validate their vote by obtaining the complete set of encrypted votes and decryption keys. They then run through the decryption keys, looking for the decryption key that matches their encryption key, followed by looking for the vote that matches their decryption key. The result should match what they voted.
None of this is impossible to implement. A paper trail would be a benefit, as you could then compare the paper tallies with the electronic tallies, INCLUDING the tallies of spoiled votes (although you cannot identify votes rejected for any other reason by means of paper). Paper would not be essential, though.
I'd probably run both BeOS and Plan9 in parallel - say by using Xen or IBM's Hypervisor - and then add support for direct OS-to-OS calls through the Hypervisor. However, Plan9-over-BeOS would probably work very nicely as well. (You should even be able to run parts of the Plan9 kernel as a layer over BeOS, in a similar way to the way Wine works, or even the full kernel as with RT-Linux or UML Linux.)
I'd probably rip out BeOS' resource manager and replace it with 9fs, then rip out 9fs' local file support and tie it directly into BeFS. (The Linux clustering FS 'Lustre' is a good example of how to do this kind of engineering.)
A really, really trivial linkage would be to simply rip out Be's resource manager, link in 9fs for that purpose, and split the desktop system entirely from the server side (similar to how mainframes operate).
There's no shortage of ways to couple two OS' - the interesting thing will be to see whether anyone tries any of them out!
First, most distributions provide mechanisms for pulling packages - be it Yum, up2date, apt or whatever.
Second, large applications (the worst offenders) are often so large that supplemental libraries would add very little extra in the way of bulk. The difference between an hour download and one hour, four minutes is insignificant, even if you'd normally curse the four minute download if it was on its own.
As a result, it would be trivial to have an installer that checked dependencies and ensured that all relevent software was installed by some means or other. Even if there's no "package manager" per-se, there's still slocate! You're not powerless to find things.
The flow-chart would look something like this:
Does the user want a statically-linked executable installed (where the choice exists)?
If yes, then we don't care about library dependencies and we just install the package.
If no, we need to find out how to reach the library.
Is library X installed?
If yes, then is it on the default library path?
If yes, then nothing needs to be done.
If no, then add it to the LD_LIBRARY_PATH for that application.
If no, then we need to obtain the library somehow.
Is the application installation off a CD/DVD image or some other local storage?
If yes, then install the library from that - the vendor DID supply everything needed, right?
If no, then we'll need to fetch it from somewhere else.
Is there a network connection? (We want the most recent library of the right version, so we prefer remote versions over distro versions.)
If yes, then pull the required version (or latest) version of the library from one of the user's preferred repositories or the application provider's repository, whichever is the better fit for that system. (One has to match, provided the application provider provides all dependencies as well.)
If no, then we'll need an alternative source.
Does the user have a CD/DVD with the required package (or perhaps source tarball, if a compile environment exists) on it?
If yes, then install the library from that.
If yes, then ask the user to insert the CD/DVD and install the llibrary from it.
If no, then (and only then) are you stuck and need the user to install the library manually.
Repeat for applications, replacing LD_LIBRARY_PATH with PATH, for the application's environment. (Applications can include scripting engines.) Finally, repeat for scripts, data files and anything else you might need to actually run the system correctly.
Where the application can link against different library APIs and/or different groups of libraries, then link it to a standard wrapper and provide one wrapper per API or API group you support -or- use dlopen() and support the lot within a single wrapper. Makes it much easier to support new libraries, or a lack of a specific library, since you're not hard-coding a specific link to something outside of your control.
None of this is impossible or even difficult. (How hard is it, really, to do a 'yum install' or a 'set LD_LIBRARY_PATH' within an application?) Since the developers have copies of the required packages (they DID test the software, right?...right? Oh.) they would need to have the packages they compiled against, so must be capable of copying those dependencies onto a CD or a networked repository.
The LSB is a solution created because many of the commercial software vendors out there regard users as necessary evils rather than the primary purpose for having the application in the first place. If vendors could sell packages without ever needing to have users, they'd love it.
Think - what part of the Earth travels the greatest distance in a 24-hour period? The equator. This absolutely guarantees high windspeeds. Because of the atmospheric conditions, it also pretty much guarantees lots of thunderstorms and other nasties.
The poles would seem a better choice, as although the elevator would need to climb further to leave the atmosphere, you would have a much easier time of it because it would be a rotating point, not a huge swinging arm.
You would need to contend with the aurora, though, as you'd have lots of charged particles flying around, which would potentially give you a huge potential difference between the ends of the elevator. (One reason the tether system experimented on by the Space Shuttle failed - even a "short" tether can generate massive potential differences and be fried.)
This is why I don't think the Space Elevator is a good approach to the problem. The requirements conflict, such that I'm not sure there IS a good place to have one.
(This would not be true on, say, Mars. I think there is a lot of potential for a space elevator on a planet like that, as the charge at the poles is much lower and the atmosphere is much less stormy.)
...many designers are looking to using ramjets or oxygen-breathing rockets, as that reduces the fuel you need. They're also looking at propellent-free systems (such as blasting an object with a high-power laser).
Totally self-contained atmospheric rockets are so... last century.:)
...is WHY they would choose the version of MPI that is regarded as a great reference edition but not that good for serious work. MPICH2 does not support shared libraries under Microsoft's C compiler or under Windows (MPICH2's docs) or external data representations. Their version of ROMIO also hangs when using MPI threads.
(Also, MPICH sucks when used with multiple devices - you have to compile it with the device(s) you're using, and can only configure it for one device type at a time. So if you're planning on using a mix of Infiniband, Globus and Ethernet, forget it. It won't work.)
Probably the best MPI library out there is Open MPI, which supports the MPI2 standard, supports MPI threads and progress threads, is much more optimizable for different platforms and was developed by groups ranging from Los Alamos Laboratories (yes, the nuke place) and the LAM/MPI development team.
Ok, you have a choice between two implementations. One is slow, has a poor release cycle and has been forked numerous times (MPICH, MPICH2, MP-MPICH, Globus MPICH, GAMMA MPICH and MVICH are all forks off the same code-base). The other is partially written in assembler, is developed by a broad consortium of MPI experts and is unlikely to fork as the maintainers are really good about integrating new code. Which would you pick?
I am also concerned about Microsoft's history of "Embrace and Extend' - are they planning on breaking the MPI-2 specifications for their own purposes? I can't see any value in them doing so, but I don't see any value in 'Embrace and Extend' anyway.
Such as using their desktop monopoly to dominate the server market, by making the de-facto client/server technology a wholly Microsoft product which Microsoft could then ensure was only (fully) available on their servers and their clients?
Oh, well, in that case I take it back. It would only be a crime if Microsoft released a new technology (let's call it.Net) and made sure that everyone used it by making it a major part of both the technological and marketting side of their business, then strangled all competitors who might use clones (Mono) or rival technologies (Java/Jini, Java/JRI, Corba 3, etc)
Obviously, they wouldn't do something like that. They only threatened to kill one competitor and "knife the baby" of another. Peaceful folk, through and through.
BeOS also chopped-and-changed who it supported, which didn't help. It was originally for the PPC, then switched to Intel, dropping all the PPC users. Arguably, it would have done better to support the entire userbase and kept the chip-specifics confined enough that you only needed a few programmers to deal with that.
What I would like to see is BeOS and Plan9 (now Inferno) hook up. Inferno makes a great low-level environment, as it makes the entire network seem like a single system. However, the front-end isn't that great and there are very few user applications for it.
BeOS has plenty of user apps, but sucks at clustering and handling distributed resources. It also doesn't have much in the way of "serious" server applications. As such, using BeOS as a user-side OS and Inferno as the server-side, would seem the obvious match.
...the statements I've been making for a while now that I simply don't see Microsoft playing fair over Mono, the same way they've attacked all of their competitors using ethically dubious means or outright illegal methods.
It is correct to say that Microsoft can choose who they want to attend the conference. There's no disputing that. Technically, they don't even have to give a reason. HOWEVER, when a reason is given and it is blatantly and willfully deceptive or untrue, then it is not so much the barring as the use of FUD to damage competition unfairly.
Forget the barring. Ignore it. It isn't the important part of the situation. What is important is whether it is correct to say that other conference-goers are being given a line intended to intimidate or coerce. THAT is the important part, the conference itself is irrelevant.
You should also forget the rights a normal competitor has in the US. As a legally-declared monopolist, supposedly monitored for potential malpractice as ordered by the courts, and as an organization fighting the necessity for increased openness as decided by EU courts, Microsoft is (in theory) limited in what it can do to use negative advertising for causing willful harm to competitors.
If this was a "normal" situation, with a "normal" company, very little of this would matter one way or the other. This is NOT a normal situation, and Microsoft was ruled a monopolist by both the US and EU, making it definitely NOT a typical player in a free market.
Actually, the EU situation is probably the most relevant here, as it is entirely possible that the example of Mono may well be usable by the EU as proof that Microsoft's counter-case over the penalties and openness of its standards are without merit. If Microsoft is willing to obstruct a free market, even when in court for doing so, then it cannot be trusted to not do so by choice at any other time.
The world has plenty of futures to choose from, but it only has one past and if you lose something from it, it can't be replaced. As such, I'd argue that serious Government funding of archaeological projects (as a science, as a method of recording the past and as a method of conserving heritage) makes considerable sense.
In England, you can do almost nothing in the way of construction without an archaeological survey of a site. Which is a sound and rational policy. Or would be, if the Government contributed towards the cost, because then the survey might have some quality to it.
In consequence, you might well expect construction firms to be interested in finding where sites were, so they could be somewhere else. Or, finding them now BEFORE they buy the land, to give the archaeologists time to dig everything of importance up by the time the land deal goes through.
So there are plenty of people who might very well have a use for such data. Certainly, the current hit-or-miss method doesn't help in conservation (obscurity simply offers more opportunity to damage or destroy in ignorance). I guess the way I look at it is that a gravity wave detector was built in Scotland for about $30bln. and it was really just a repeat of the M-M ether experiment, so guaranteed to fail.
If you can afford to throw away thirty billion dollars on something you know won't work when you've finished it, you've enough spare cash to completely excavate and document virtually every potential archaeological site in England, Scotland, Wales and Northern Ireland. And STILL have change for a quiet evening down the pub.
Rebuilding New Orleans is going to cost something like $200 billion, but without new taxes, the costs are going to end up on the US national debt to be paid later. If you can do that, you can surely borrow enough to get every unemployed person in America in one gigantic human chain to go through the Amazon, locating ruins. It would probably cost less and if the Governments agreed on profit-sharing, the added tourism would be far more likely to pay the costs back than President Bush's New Orleans vision ever will.
In other words, the money isn't the problem. There will always be someone with money you could convince. The problem is getting enough data together to be able to convince people that it is worth investing the money.
Oh, and you're absolutely right about the importance of the data and context. Which is why I was bloody furious with the fiasco over Seahenge, where the archaeologists kept only one copy of all the notes in a single building which was poorly maintained. They'd excavated the entire site, so the ONLY existant data was in the notes, which were destroyed when an electrical fire burned the place to the ground.
Serious archaeologists will never convince the average person not to be destructive, with incidents like that. You only need one or two - reputations can be destroyed far more easily than they can be built up.
It didn't help matters when the US forces in Iraq were shown to have damaged or destroyed ancient sites (including Babylon) and to have stolen artifacts from archaeological sites. Again, what does this teach others? That such stuff is yours for the taking, if you're "bad" enough.
Nor did it help when an ancient North American site, held secret for many years, was handed over to the US Government and promptly pilfered. Quite probably by people within the Government, as they're the ones who knew about it.
You teach by example, and the examples that have been shown aren't good. As I see it, the only way to build a good image is to make communities more actively involved in preservation - not as a burden but as an opportunity. But you'll only succeed if those with the money (and authority) back that up by making it more profitable to be honest than corrupt.
You can rapidly build up charges of a few tens of thousands of volts at very close to zero current. It's not that hard to build a few million volts, provided the current is low enough and the surroundings are insulating enough. The key, as you've pointed out, is power - and you don't have a whole lot without current.
A Van De Graaf generator is basically a band of insulating material being rotated in a tower with some means of transferring a charge to it. There are relatively cheap desktop and home models that'll produce nearly half a million volts. Schools use such devices all the time, so if the fireman hasn't seem a voltage that high, he skipped classes.
Having said that, early atom-smashers used Van De Graaf generators only capable of producing five million or so volts. It seems reasonable to suspect something will burn before it is blasted out of existence. So, somewhere between 400,000 volts and 5,000,000 volts, you might be able to ignite something.
However, here we get a problem. You can't just carry around half a million volts and not notice it. Your hair tends to stand on end, for a start. ANYTHING metal - even a doorknob - will cause a discharge to occur. Getting into his car certainly would have - even if the car were carbin-fiber, the key would be metal and the distance short enough for an arc to occur.
There's also the problem of where you lodge a charge that great. A capacitor is basically two electrostatic devices with an insulator between them. In this case, the insulator would be the shoes, and the electrostatic device the person. I'll assume there are enough nails holding the carpet down to act as the other electrostatic device.
But what is the capacitance of a person? The figure I've been able to get with a Google search is an average of 204 pF with a typical range of 95 to 398 pF. (It varies according to height and weight, so a seven-foot sumo wrestler might have a higher capacitance than this range shows.)
In other words, not really what you'd need to carry half a million volts around. The jacket would have carried more, but unless it was made of Tantallum or some other material with very high capacitance, I doubt you'd be able to store enough charge to start setting things on fire.
In other words, there is nothing credible about the story. The voltages are abnormally low for a static device and way too low to actually do any fire damage, there's nowhere a higher charge could have been stored and there would have been too many points at which positively violent arcing would have occurred if it had been stored.
The 12,000 year old site currently being excavated in the town I grew up was discovered by chance during a severe drought - discoloration clearly marked outlines of ancient structures. The site has been worked by archaeologists for about 7 years now and they're uncovering a vast amount each year.
(Having said that, the entire settlement is believed to be hundreds - if not thousands - of times larger than the area actually examined by archaeologists. Add in nearby standing stones and round barrows, and the area in need of study is maybe hundreds of thousands of times larger than what they've studied. Makes you wonder what they haven't found!)
You can't expect a good pair of eyes (and a brain) to exist in every town or village that has ancient remains. On the other hand, with something like Google Maps, all it really requires is someone anywhere taking the time to look through the images.
Well, if they're sophisticated enough, all they really need to do is write a good image processing algorithm that detects definite artifacts in the image (straight lines, circles, etc) that do NOT correspond to anything that is a definite surface structure. All the person need do then is search through the candidate images, not the entire database, which would be a much more practical task to do.
Ideally, you'd use several layers of image processing, to whittle down the pool of images to highly probable cases, then subtract out known archaeological sites from a database.
Really, really ideally, you'd program the individual layers as BOING components and run the computation part of it as a gigantic @Home venture, as this would be massively parallelizable and sufficiently CPU intensive for most academics who would be interested in such work to not be able to afford a computer (or cluster) that could actually carry out the work in a reasonable timeframe.
Hmmmm. It's a pity Google don't cover enough of the UK in enough depth to be able to do good work there.
The only way to effectively deal with that sort of situation is to export aggressively (which helps with the trade balance anyway and therefore keeps the economists happy).
IMHO, the "ideal" in Europe would be for countries to specialize to some degree, so as to be competing more with other countries than between themselves. Competing within Europe is like a soccer team where the players on that team obstruct each other - they can't win that way.
When you specialize in this way, your "domestic" market essentially expands to the whole of Europe, dealing with a sizable part of the overproduction problem. You also have much more power behind you for export, because you don't have different parts of Europe fighting between themselves over foreign markets - it becomes in EVERYONE's interest for country X to export Y to Z, across Europe.
But it's worth remembering that one of the "heroes" of the Industrial Revolution, Robert Owen, worked on the basis that an educated workforce produces more, faster and cheaper.
Sponsoring individual students, however, is hit-or-miss and expensive. On the other hand, sponsoring a teacher (who may well have a class of anywhere between several dozen to several hundred) is cheap, efficient and much more likely to produce "good results" (in terms of useful employees).
My guess is that IBM is planning on pushing their names to the classes and on aggressively recruiting the best. That way, they get to "upgrade" their workforce to a younger generation (who will be cheaper, as a result) who know everything IBM needs them to know (because it's IBM staffers doing the teaching) AND who also know all the latest-generation stuff on top of that (because they're learning today, as opposed to a decade ago).
That would cut costs, multiply workforce quality AND massively boost their image, all in one go.
Oh, and because the IBM staffers will do the steering, the students may well end up worse employees for competitors...
Any CO2 releases that occur from carbon actively in the carbon cycle does NOT add to that cycle, by definition. Burning hydrocarbons, or other locked carbon sources, releases carbon that was NOT free (duh!).
The next thing to consider is that volcanos are relatively instantaneous events, whereas industry is a continuous process. Because of this, we're talking two very different mechanisms. The two are not directly comparable.
Finally, volcanos blast CO2 to quite high altitudes, whereas (say) cars produce a very low altitude smog. (Especially when you get temperature inversions.) You don't get temperature inversions occuring during a volcanic eruption. Or, at least, not for very long. In consequence, we're not even talking about the same environment. Again, the two scenarios cannot be directly compared.
Anyone who tries to draw a direct analogy between two completely different systems as though they operated identically is guaranteed to produce the results they want, NOT the results that would actually occur. Those who model what is actually going on, on the other hand, are bound to see what IS.
You are correct, MD4 and MD5 are hashing functions. and hashing functions are one-way. Hashing is typically used within encryption - say to generate reasonable keys from pass-phrases. You could use them as an "encryption mode" (a method of using different keys per block) by hashing the prior unencrypted block and using the result as the new key.
Well, there is ONE possible exception to this. You can use hashes for error-correction. If you have enough hashes over enough slices of the data, you could actually regenerate the original dataset from just the hashes. Whether there would be any advantage in using such a method is debatable, though I suspect there may be a variant you could use where conventional error-correction codes would be unable to handle the noise levels.
And in a moral sense, you're perfectly right. However, the fact is, a great many scientific discoveries are made by theft (physical or intellectual) and it is a bogus morality to insist on only certain discoveries being correctly attributed.
Personally, I would wholeheartedly agree with anyone who insisted on all textbooks, journals, etc, correctly crediting those who did the work. I doubt a single textbook (and very few journals) would meet that standard, which would make for a very limited educational system.
From a more "flexible" perspective, there are other points to consider. Scientists often place their work in the public domain because concepts of ownership really don't apply. There isn't a warehouse, somewhere, where they stock up on laws of physics to be sold on eBay. A planet or a comet exists, whether you are there to observe it or not. 1+1=2, and will continue to do so long after Sol has become a red giant, destroying the planet.
From this perspective, science is communal. It exists, it always will exist, it always has existed. All you do is enter into that realm. You can't change it, you can't add to it, you can't take any of it away, all you can do is observe.
Of course, to be self-consistant and honest with oneself, if you DO hold such a view, you cannot claim credit or ownership when disputes arise. If something is communal enough for you to "borrow" from someone else, then it's communal enough for you to have no particular rights yourself. You can't have it both ways.
My personal view is that scientists (and communities in general) need a mix of these two perspectives. There are some things I can fully understand being private or privately owned, and the true ownership should be acknowledged and accepted by all. If it is an identifiable entity that can be labelled and distinguished, and which is also mutable, then private ownership is unquestionably the right thing.
If it cannot be identified, if it is some universal property, or if it is something you can't actually do anything with, then you cannot truly be said to own it. If you think of ownership in a hierarchical sense, then something cannot be both subclass and superclass of the same thing at the same time. If you ban paradoxes then anything that is paradoxical to own should not qualify as ownable.
In this particular case, I don't see that it neatly falls into either of the above two scenarios. There was actual theft of something that (by my standards above) was privately owned. On the other hand, the data on which the theft was based was universal and therefore could not be truly owned.
(The Earth simply isn't big enough for observers within the same hemisphere to notice a significant difference in position of an object that distant, when using a standard coordinate system, so the values are not the product of the observer, they are the product of the observed.)
Finally, both teams did work which only when unified produced a result that was meaningful. In consequence, neither team could be said to have actually made the discovery, per se, because neither had the data to do so.
Yes and no. There were two important factors - the PR involved (the Russians had beaten the Americans to every major goal up to then, despite the fact that the Americans had the pick of the German rocket scientists and the Russians ended up with the V2 grunts), and the guidance system boast.
The latter was militarily the more important, as it demonstrated the ability to hit a specific target over a long distance. Essential, if your defence relies almost entirely on extreme-range ballistic missiles and cruise missiles.
Even so, it was widely known that the Russian guidance systems were, by and large, superior for the ballistic systems. If the Americans could hit an American football field from a thousand miles away, the Russians could place a missile through the posts.
This is one reason the Americans largely moved to cruise missiles - there, their more sophisticated computing and imaging technologies gave them an edge. The moon landings weren't much use in this arena, though, so developing ballistic guidance systems was no longer of much importance. This is a much more plausible explanation for the abandonment of the Apollo program and the closing of all the Saturn V construction facilities than the Apollo 13 + PR disaster hypothesis. Governments ignore public opinion when it suits them, the military is another matter.
The task of reaching the moon today does indeed involve a different range of problems. The military is still not interested, NASA is now staffed more by managers than engineers (they outsource everything) and they do have different objectives - they're just not sure what they are.
Remember, although President Bush has told NASA it should return to the moon, Congress hasn't given it any money to do so with and with all the emergency spending it has had to do with Iraq, Afghanistan and Katrina, it might well hold back for the remainder of his time in office.
It should also be borne in mind that engineers involved in such a program will be considering the complete failures of computer modelling with the space shuttle. (Computer models predicted no foam loss greater than 1mm cube from the external tank on the last launch, and were hopelessly wide of the mark on airflow round heaters and other protrusions.)
If they cannot accurately model a system they have had something like 20 years worth of actual hard data on, they are NOT going to be confident they can accurately model a totally new system they have zero flight data for. However, America being America, they will nonetheless make the system as complex as humanly possible. This means they will be forced into doing far more extensive testing, as that's the only way they'll have any confidence in the results.
Further, America being America, the engineering tasks will be farmed out by political considerations and not expertise or capability. This is the only reason the Shuttle's booster rockets even have the O-rings which caused the Challenger disaster - the way the jobs were handed out made it impossible to build a single-piece rocket.
Given that Halliburton has been awarded a no-compete contract to rebuild New Orleans, it doesn't take a genius to figure out who would get the contract to build key components of a new moon rocket, if the budget for it goes through before 2008.
One last complicating factor - space technology is widely distributed across the globe. However, after the ISS fiasco, in which America has either scaled back or simply not delivered key parts, a joint venture with other nations seems unlikely. Europe might also be wary of doing anything much with America, after the politics of the fusion reactor placement, numerous trade wars and other international scandals.
A new administration might help, but then a new administration might simply cancel the program altogether. NASA knows this, other nations know this, and nobody is going to invest money on a program they don't know will even exist long enough to deliver.
"Up to" means that there was a second in which they may have reached 240 megabits per second. I'm not sure if this is the data before or after compression, so 240 bits of 1's run-length encoded might easily be transmitted in one second.
It is also unclear as to whether the data was actually intact or not, how much error-correction the network card needed to perform, how many resends were required, etc.
In other words, even a transmitted rate of 240 megabits per second need not equal 240 megabit transfer rates. There are plenty of ways to fudge the numbers.
A trivial example: A network card operates at 240 megabits per second, but needs 240 retries to get enough data across for a genetic algorithm to build the most probable originating packet that could produce the data received, where the genetic algorithm adds several minutes to the transmission time of a single packet. At what speed does the card operate?
Microsoft has a habit of buying companies it starts joint ventures with. Assuming it continues in this trend, the combined MSN/AOL system will have a very large userbase and the means to tune the OS such that no other Internet Provider will work as well - whether it uses that means or not.
This need not require deliberately crippling other providers - though, as the DR-DOS case showed, they're not above that. It could be as simple as merging AOL-specific code into the Windows kernel, which would make anything using AOL faster than anything that ran purely in userspace.
The second aspect of this is that AOL and Time-Warner are still very connected. If Microsoft controlled AOL - now or in the future - it could also gain control of Time-Warner. Logically, they would then combine Time-Warner with all their other media interests (ZDNet, MSNBC, etc) which would substantially reduce the diversity of outlets out there.
It wouldn't be the end of the world, but I would not be keen to see Microsoft in the driving-seat of both conventional AND Internet media.
Slashdot Mirror
The sole question, and I'm pretty sure you agree with me on this, is what to replace the rocket system with. My personal preference is to push R&D in every possible direction. Ideas that don't work will fall off the edge on their own, leaving the best ones standing. That way, the pick is based on what works, not what someone thinks should work.
You enter your vote and encrypt it anonymously. (It would need salting, to prevent trivial breech of anonymity by theft of the encryption key.) The encryption key would then need to be shredded or otherwise destroyed, if total anonymity was to be preserved. For nearly-total anonymity, you don't have to do this and you get the benefit of validation.
The server, on counting, then runs through the list of decryption keys for each vote. When it finds a key that decrypts to a valid vote, it records the vote and disposes of the key.
The total number of votes that should be recorded will be equal to the total number of keys minus the number of keys still left after all votes are counted.
An encrypted vote for which no decryption key exists (which will occur if two or more votes exist for the same key, OR if a vote has been submitted for an unlisted key) is then counted as a "spoiled ballot". In the case of duplicate votes, it would then be possible to use the decryption keys to determine the prevelence of voter fraud in a given district.
Since you can hand copies of the encrypted votes and decryption keys to anyone, without fear of threatening anonymity or risking voter fraud, it would be possible for others to independently verify the tallies.
Assuming near-total anonymity is fine, then the voter can validate their vote by obtaining the complete set of encrypted votes and decryption keys. They then run through the decryption keys, looking for the decryption key that matches their encryption key, followed by looking for the vote that matches their decryption key. The result should match what they voted.
None of this is impossible to implement. A paper trail would be a benefit, as you could then compare the paper tallies with the electronic tallies, INCLUDING the tallies of spoiled votes (although you cannot identify votes rejected for any other reason by means of paper). Paper would not be essential, though.
I'd probably rip out BeOS' resource manager and replace it with 9fs, then rip out 9fs' local file support and tie it directly into BeFS. (The Linux clustering FS 'Lustre' is a good example of how to do this kind of engineering.)
A really, really trivial linkage would be to simply rip out Be's resource manager, link in 9fs for that purpose, and split the desktop system entirely from the server side (similar to how mainframes operate).
There's no shortage of ways to couple two OS' - the interesting thing will be to see whether anyone tries any of them out!
Second, large applications (the worst offenders) are often so large that supplemental libraries would add very little extra in the way of bulk. The difference between an hour download and one hour, four minutes is insignificant, even if you'd normally curse the four minute download if it was on its own.
As a result, it would be trivial to have an installer that checked dependencies and ensured that all relevent software was installed by some means or other. Even if there's no "package manager" per-se, there's still slocate! You're not powerless to find things.
The flow-chart would look something like this:
Does the user want a statically-linked executable installed (where the choice exists)?
Is library X installed?
Is the application installation off a CD/DVD image or some other local storage?
Is there a network connection? (We want the most recent library of the right version, so we prefer remote versions over distro versions.)
Does the user have a CD/DVD with the required package (or perhaps source tarball, if a compile environment exists) on it?
Repeat for applications, replacing LD_LIBRARY_PATH with PATH, for the application's environment. (Applications can include scripting engines.) Finally, repeat for scripts, data files and anything else you might need to actually run the system correctly.
Where the application can link against different library APIs and/or different groups of libraries, then link it to a standard wrapper and provide one wrapper per API or API group you support -or- use dlopen() and support the lot within a single wrapper. Makes it much easier to support new libraries, or a lack of a specific library, since you're not hard-coding a specific link to something outside of your control.
None of this is impossible or even difficult. (How hard is it, really, to do a 'yum install' or a 'set LD_LIBRARY_PATH' within an application?) Since the developers have copies of the required packages (they DID test the software, right?
The LSB is a solution created because many of the commercial software vendors out there regard users as necessary evils rather than the primary purpose for having the application in the first place. If vendors could sell packages without ever needing to have users, they'd love it.
The poles would seem a better choice, as although the elevator would need to climb further to leave the atmosphere, you would have a much easier time of it because it would be a rotating point, not a huge swinging arm.
You would need to contend with the aurora, though, as you'd have lots of charged particles flying around, which would potentially give you a huge potential difference between the ends of the elevator. (One reason the tether system experimented on by the Space Shuttle failed - even a "short" tether can generate massive potential differences and be fried.)
This is why I don't think the Space Elevator is a good approach to the problem. The requirements conflict, such that I'm not sure there IS a good place to have one.
(This would not be true on, say, Mars. I think there is a lot of potential for a space elevator on a planet like that, as the charge at the poles is much lower and the atmosphere is much less stormy.)
Charlie and the Great Glass Elevator. It is full of the perils of space elevators, particularly when encountering knids.
Totally self-contained atmospheric rockets are so... last century.
(Also, MPICH sucks when used with multiple devices - you have to compile it with the device(s) you're using, and can only configure it for one device type at a time. So if you're planning on using a mix of Infiniband, Globus and Ethernet, forget it. It won't work.)
Probably the best MPI library out there is Open MPI, which supports the MPI2 standard, supports MPI threads and progress threads, is much more optimizable for different platforms and was developed by groups ranging from Los Alamos Laboratories (yes, the nuke place) and the LAM/MPI development team.
Ok, you have a choice between two implementations. One is slow, has a poor release cycle and has been forked numerous times (MPICH, MPICH2, MP-MPICH, Globus MPICH, GAMMA MPICH and MVICH are all forks off the same code-base). The other is partially written in assembler, is developed by a broad consortium of MPI experts and is unlikely to fork as the maintainers are really good about integrating new code. Which would you pick?
I am also concerned about Microsoft's history of "Embrace and Extend' - are they planning on breaking the MPI-2 specifications for their own purposes? I can't see any value in them doing so, but I don't see any value in 'Embrace and Extend' anyway.
Oh, well, in that case I take it back. It would only be a crime if Microsoft released a new technology (let's call it
Obviously, they wouldn't do something like that. They only threatened to kill one competitor and "knife the baby" of another. Peaceful folk, through and through.
What I would like to see is BeOS and Plan9 (now Inferno) hook up. Inferno makes a great low-level environment, as it makes the entire network seem like a single system. However, the front-end isn't that great and there are very few user applications for it.
BeOS has plenty of user apps, but sucks at clustering and handling distributed resources. It also doesn't have much in the way of "serious" server applications. As such, using BeOS as a user-side OS and Inferno as the server-side, would seem the obvious match.
It is correct to say that Microsoft can choose who they want to attend the conference. There's no disputing that. Technically, they don't even have to give a reason. HOWEVER, when a reason is given and it is blatantly and willfully deceptive or untrue, then it is not so much the barring as the use of FUD to damage competition unfairly.
Forget the barring. Ignore it. It isn't the important part of the situation. What is important is whether it is correct to say that other conference-goers are being given a line intended to intimidate or coerce. THAT is the important part, the conference itself is irrelevant.
You should also forget the rights a normal competitor has in the US. As a legally-declared monopolist, supposedly monitored for potential malpractice as ordered by the courts, and as an organization fighting the necessity for increased openness as decided by EU courts, Microsoft is (in theory) limited in what it can do to use negative advertising for causing willful harm to competitors.
If this was a "normal" situation, with a "normal" company, very little of this would matter one way or the other. This is NOT a normal situation, and Microsoft was ruled a monopolist by both the US and EU, making it definitely NOT a typical player in a free market.
Actually, the EU situation is probably the most relevant here, as it is entirely possible that the example of Mono may well be usable by the EU as proof that Microsoft's counter-case over the penalties and openness of its standards are without merit. If Microsoft is willing to obstruct a free market, even when in court for doing so, then it cannot be trusted to not do so by choice at any other time.
If it's magic, it must be all that Harry Potter's fault.
In England, you can do almost nothing in the way of construction without an archaeological survey of a site. Which is a sound and rational policy. Or would be, if the Government contributed towards the cost, because then the survey might have some quality to it.
In consequence, you might well expect construction firms to be interested in finding where sites were, so they could be somewhere else. Or, finding them now BEFORE they buy the land, to give the archaeologists time to dig everything of importance up by the time the land deal goes through.
So there are plenty of people who might very well have a use for such data. Certainly, the current hit-or-miss method doesn't help in conservation (obscurity simply offers more opportunity to damage or destroy in ignorance). I guess the way I look at it is that a gravity wave detector was built in Scotland for about $30bln. and it was really just a repeat of the M-M ether experiment, so guaranteed to fail.
If you can afford to throw away thirty billion dollars on something you know won't work when you've finished it, you've enough spare cash to completely excavate and document virtually every potential archaeological site in England, Scotland, Wales and Northern Ireland. And STILL have change for a quiet evening down the pub.
Rebuilding New Orleans is going to cost something like $200 billion, but without new taxes, the costs are going to end up on the US national debt to be paid later. If you can do that, you can surely borrow enough to get every unemployed person in America in one gigantic human chain to go through the Amazon, locating ruins. It would probably cost less and if the Governments agreed on profit-sharing, the added tourism would be far more likely to pay the costs back than President Bush's New Orleans vision ever will.
In other words, the money isn't the problem. There will always be someone with money you could convince. The problem is getting enough data together to be able to convince people that it is worth investing the money.
Oh, and you're absolutely right about the importance of the data and context. Which is why I was bloody furious with the fiasco over Seahenge, where the archaeologists kept only one copy of all the notes in a single building which was poorly maintained. They'd excavated the entire site, so the ONLY existant data was in the notes, which were destroyed when an electrical fire burned the place to the ground.
Serious archaeologists will never convince the average person not to be destructive, with incidents like that. You only need one or two - reputations can be destroyed far more easily than they can be built up.
It didn't help matters when the US forces in Iraq were shown to have damaged or destroyed ancient sites (including Babylon) and to have stolen artifacts from archaeological sites. Again, what does this teach others? That such stuff is yours for the taking, if you're "bad" enough.
Nor did it help when an ancient North American site, held secret for many years, was handed over to the US Government and promptly pilfered. Quite probably by people within the Government, as they're the ones who knew about it.
You teach by example, and the examples that have been shown aren't good. As I see it, the only way to build a good image is to make communities more actively involved in preservation - not as a burden but as an opportunity. But you'll only succeed if those with the money (and authority) back that up by making it more profitable to be honest than corrupt.
A Van De Graaf generator is basically a band of insulating material being rotated in a tower with some means of transferring a charge to it. There are relatively cheap desktop and home models that'll produce nearly half a million volts. Schools use such devices all the time, so if the fireman hasn't seem a voltage that high, he skipped classes.
Having said that, early atom-smashers used Van De Graaf generators only capable of producing five million or so volts. It seems reasonable to suspect something will burn before it is blasted out of existence. So, somewhere between 400,000 volts and 5,000,000 volts, you might be able to ignite something.
However, here we get a problem. You can't just carry around half a million volts and not notice it. Your hair tends to stand on end, for a start. ANYTHING metal - even a doorknob - will cause a discharge to occur. Getting into his car certainly would have - even if the car were carbin-fiber, the key would be metal and the distance short enough for an arc to occur.
There's also the problem of where you lodge a charge that great. A capacitor is basically two electrostatic devices with an insulator between them. In this case, the insulator would be the shoes, and the electrostatic device the person. I'll assume there are enough nails holding the carpet down to act as the other electrostatic device.
But what is the capacitance of a person? The figure I've been able to get with a Google search is an average of 204 pF with a typical range of 95 to 398 pF. (It varies according to height and weight, so a seven-foot sumo wrestler might have a higher capacitance than this range shows.)
In other words, not really what you'd need to carry half a million volts around. The jacket would have carried more, but unless it was made of Tantallum or some other material with very high capacitance, I doubt you'd be able to store enough charge to start setting things on fire.
In other words, there is nothing credible about the story. The voltages are abnormally low for a static device and way too low to actually do any fire damage, there's nowhere a higher charge could have been stored and there would have been too many points at which positively violent arcing would have occurred if it had been stored.
(Having said that, the entire settlement is believed to be hundreds - if not thousands - of times larger than the area actually examined by archaeologists. Add in nearby standing stones and round barrows, and the area in need of study is maybe hundreds of thousands of times larger than what they've studied. Makes you wonder what they haven't found!)
You can't expect a good pair of eyes (and a brain) to exist in every town or village that has ancient remains. On the other hand, with something like Google Maps, all it really requires is someone anywhere taking the time to look through the images.
Well, if they're sophisticated enough, all they really need to do is write a good image processing algorithm that detects definite artifacts in the image (straight lines, circles, etc) that do NOT correspond to anything that is a definite surface structure. All the person need do then is search through the candidate images, not the entire database, which would be a much more practical task to do.
Ideally, you'd use several layers of image processing, to whittle down the pool of images to highly probable cases, then subtract out known archaeological sites from a database.
Really, really ideally, you'd program the individual layers as BOING components and run the computation part of it as a gigantic @Home venture, as this would be massively parallelizable and sufficiently CPU intensive for most academics who would be interested in such work to not be able to afford a computer (or cluster) that could actually carry out the work in a reasonable timeframe.
Hmmmm. It's a pity Google don't cover enough of the UK in enough depth to be able to do good work there.
IMHO, the "ideal" in Europe would be for countries to specialize to some degree, so as to be competing more with other countries than between themselves. Competing within Europe is like a soccer team where the players on that team obstruct each other - they can't win that way.
When you specialize in this way, your "domestic" market essentially expands to the whole of Europe, dealing with a sizable part of the overproduction problem. You also have much more power behind you for export, because you don't have different parts of Europe fighting between themselves over foreign markets - it becomes in EVERYONE's interest for country X to export Y to Z, across Europe.
Sponsoring individual students, however, is hit-or-miss and expensive. On the other hand, sponsoring a teacher (who may well have a class of anywhere between several dozen to several hundred) is cheap, efficient and much more likely to produce "good results" (in terms of useful employees).
My guess is that IBM is planning on pushing their names to the classes and on aggressively recruiting the best. That way, they get to "upgrade" their workforce to a younger generation (who will be cheaper, as a result) who know everything IBM needs them to know (because it's IBM staffers doing the teaching) AND who also know all the latest-generation stuff on top of that (because they're learning today, as opposed to a decade ago).
That would cut costs, multiply workforce quality AND massively boost their image, all in one go.
Oh, and because the IBM staffers will do the steering, the students may well end up worse employees for competitors...
The next thing to consider is that volcanos are relatively instantaneous events, whereas industry is a continuous process. Because of this, we're talking two very different mechanisms. The two are not directly comparable.
Finally, volcanos blast CO2 to quite high altitudes, whereas (say) cars produce a very low altitude smog. (Especially when you get temperature inversions.) You don't get temperature inversions occuring during a volcanic eruption. Or, at least, not for very long. In consequence, we're not even talking about the same environment. Again, the two scenarios cannot be directly compared.
Anyone who tries to draw a direct analogy between two completely different systems as though they operated identically is guaranteed to produce the results they want, NOT the results that would actually occur. Those who model what is actually going on, on the other hand, are bound to see what IS.
Well, there is ONE possible exception to this. You can use hashes for error-correction. If you have enough hashes over enough slices of the data, you could actually regenerate the original dataset from just the hashes. Whether there would be any advantage in using such a method is debatable, though I suspect there may be a variant you could use where conventional error-correction codes would be unable to handle the noise levels.
Personally, I would wholeheartedly agree with anyone who insisted on all textbooks, journals, etc, correctly crediting those who did the work. I doubt a single textbook (and very few journals) would meet that standard, which would make for a very limited educational system.
From a more "flexible" perspective, there are other points to consider. Scientists often place their work in the public domain because concepts of ownership really don't apply. There isn't a warehouse, somewhere, where they stock up on laws of physics to be sold on eBay. A planet or a comet exists, whether you are there to observe it or not. 1+1=2, and will continue to do so long after Sol has become a red giant, destroying the planet.
From this perspective, science is communal. It exists, it always will exist, it always has existed. All you do is enter into that realm. You can't change it, you can't add to it, you can't take any of it away, all you can do is observe.
Of course, to be self-consistant and honest with oneself, if you DO hold such a view, you cannot claim credit or ownership when disputes arise. If something is communal enough for you to "borrow" from someone else, then it's communal enough for you to have no particular rights yourself. You can't have it both ways.
My personal view is that scientists (and communities in general) need a mix of these two perspectives. There are some things I can fully understand being private or privately owned, and the true ownership should be acknowledged and accepted by all. If it is an identifiable entity that can be labelled and distinguished, and which is also mutable, then private ownership is unquestionably the right thing.
If it cannot be identified, if it is some universal property, or if it is something you can't actually do anything with, then you cannot truly be said to own it. If you think of ownership in a hierarchical sense, then something cannot be both subclass and superclass of the same thing at the same time. If you ban paradoxes then anything that is paradoxical to own should not qualify as ownable.
In this particular case, I don't see that it neatly falls into either of the above two scenarios. There was actual theft of something that (by my standards above) was privately owned. On the other hand, the data on which the theft was based was universal and therefore could not be truly owned.
(The Earth simply isn't big enough for observers within the same hemisphere to notice a significant difference in position of an object that distant, when using a standard coordinate system, so the values are not the product of the observer, they are the product of the observed.)
Finally, both teams did work which only when unified produced a result that was meaningful. In consequence, neither team could be said to have actually made the discovery, per se, because neither had the data to do so.
The latter was militarily the more important, as it demonstrated the ability to hit a specific target over a long distance. Essential, if your defence relies almost entirely on extreme-range ballistic missiles and cruise missiles.
Even so, it was widely known that the Russian guidance systems were, by and large, superior for the ballistic systems. If the Americans could hit an American football field from a thousand miles away, the Russians could place a missile through the posts.
This is one reason the Americans largely moved to cruise missiles - there, their more sophisticated computing and imaging technologies gave them an edge. The moon landings weren't much use in this arena, though, so developing ballistic guidance systems was no longer of much importance. This is a much more plausible explanation for the abandonment of the Apollo program and the closing of all the Saturn V construction facilities than the Apollo 13 + PR disaster hypothesis. Governments ignore public opinion when it suits them, the military is another matter.
The task of reaching the moon today does indeed involve a different range of problems. The military is still not interested, NASA is now staffed more by managers than engineers (they outsource everything) and they do have different objectives - they're just not sure what they are.
Remember, although President Bush has told NASA it should return to the moon, Congress hasn't given it any money to do so with and with all the emergency spending it has had to do with Iraq, Afghanistan and Katrina, it might well hold back for the remainder of his time in office.
It should also be borne in mind that engineers involved in such a program will be considering the complete failures of computer modelling with the space shuttle. (Computer models predicted no foam loss greater than 1mm cube from the external tank on the last launch, and were hopelessly wide of the mark on airflow round heaters and other protrusions.)
If they cannot accurately model a system they have had something like 20 years worth of actual hard data on, they are NOT going to be confident they can accurately model a totally new system they have zero flight data for. However, America being America, they will nonetheless make the system as complex as humanly possible. This means they will be forced into doing far more extensive testing, as that's the only way they'll have any confidence in the results.
Further, America being America, the engineering tasks will be farmed out by political considerations and not expertise or capability. This is the only reason the Shuttle's booster rockets even have the O-rings which caused the Challenger disaster - the way the jobs were handed out made it impossible to build a single-piece rocket.
Given that Halliburton has been awarded a no-compete contract to rebuild New Orleans, it doesn't take a genius to figure out who would get the contract to build key components of a new moon rocket, if the budget for it goes through before 2008.
One last complicating factor - space technology is widely distributed across the globe. However, after the ISS fiasco, in which America has either scaled back or simply not delivered key parts, a joint venture with other nations seems unlikely. Europe might also be wary of doing anything much with America, after the politics of the fusion reactor placement, numerous trade wars and other international scandals.
A new administration might help, but then a new administration might simply cancel the program altogether. NASA knows this, other nations know this, and nobody is going to invest money on a program they don't know will even exist long enough to deliver.
It is also unclear as to whether the data was actually intact or not, how much error-correction the network card needed to perform, how many resends were required, etc.
In other words, even a transmitted rate of 240 megabits per second need not equal 240 megabit transfer rates. There are plenty of ways to fudge the numbers.
A trivial example: A network card operates at 240 megabits per second, but needs 240 retries to get enough data across for a genetic algorithm to build the most probable originating packet that could produce the data received, where the genetic algorithm adds several minutes to the transmission time of a single packet. At what speed does the card operate?
Being nocturnal will work well with not going outdoors, and the hooked beak will be great for typing. The grubs'll make a mess, though.
No, the headers get stuck to the maple syrup when travelling down the ethernet cables.
This need not require deliberately crippling other providers - though, as the DR-DOS case showed, they're not above that. It could be as simple as merging AOL-specific code into the Windows kernel, which would make anything using AOL faster than anything that ran purely in userspace.
The second aspect of this is that AOL and Time-Warner are still very connected. If Microsoft controlled AOL - now or in the future - it could also gain control of Time-Warner. Logically, they would then combine Time-Warner with all their other media interests (ZDNet, MSNBC, etc) which would substantially reduce the diversity of outlets out there.
It wouldn't be the end of the world, but I would not be keen to see Microsoft in the driving-seat of both conventional AND Internet media.