Time in evolution is not measured in years, but in generations. An evolutionary software process can go through millions of generations in a day, compared to humans, which take roughly 20 years for each generation.
Mechanical computers are one of the worst nanotech ideas. They are a "proof of concept" thing to demonstrate that nano-scale computers are possible, because in absence of experiment it is easier to convince people that bumpy nanosticks can compute than that the quantum effects can be dealt with. Drexler needed something possible to point at, I doubt even he seriously thinks mechanical computers are the way of the future.
They are definitely way slower than an equivalent-scaled electronic or light-based design. They are indeed theoretically a bit faster (about 1 GHz and 1000 MIPS) than your current average production chip (when "Engines of Creation" came out, the gap was much wider), but it is a theoretical high estimate which conventional electronic designs have already overtaken.
Realistically, do you think it's more efficient to move around hundreds of whole atoms or a few electrons or photons? (the concept of single-particle signalling is a very exciting one to me)
Incidentally, I have a copy of Nanosystems on my lap as I type this. I've read it, but I don't have the expertise to check his work. It is very thorough and quite interesting if you like theoretical engineering.
In the comics on the website, the human "batteries" are used for other things that make a whole lot more sense. I won't say what, but I will say that they are worth reading, and they only contradict the movie to improve on it, tearing up that "little safety net" you mention.
I agree with the sequel potential, but it would have to be in the real world, and thus it would be a very different movie.
Back in the Doom days, when they talked about Quake, they talked about a real-physics universe where the player is basically Thor and the main weapon is the hammer.
I remember in particular one line about how you were going to be able to knock someone down and then go over and kick the s**t out of him while he was down.
It was like Daikatana, really, in that it revolved around this one apparently primitive weapon that gained power. I think you were supposed to be able to cause local earthquakes with the hammer and open up cracks in the ground or something like that.
Let's face it, nothing in the premise was logical. It was internally consistent, but had no connection to reality. I could pick it apart, but everyone who's seen it and has basic knowledge of computers, biology, and thermodynamics (i.e./. readers) knows what I'm talking about.
OTOH, I staggered out of the theatre. It released all my adrenaline reserves and left me a hollow shell. I was two blocks away from the theatre before I regained any sense of reality.
The only thing I could compare it to in terms of adrenaline rush is Ninja Scroll (no, I don't suggest they're similar in any other way). And Ninja Scroll didn't pluck at my geeky heartstrings the same way.
However, definitely a one-shot movie. I don't ever want to watch it again. I was blown away in a similar manner by Starship Troopers (which must have been produced by someone who hated the book, BTW), but when I rented it trying to recapture the feeling, I was disappointed by a campy sci-fi flick. I'm sure it would be fun to watch clips of those great fight scenes again (and watch for digital artifacts ^_^), but I'm not sure I could sit through the rest.
Neither China nor Cuba are true democratic communist states. They just have better PR in the US because they weren't the US's main military rival.
It's important to note that the true splits are democracy/dictatorship and capitalism/communism; US propaganda from the cold war days commonly redefined communism as the opposite of democracy (innacurate but fair, because none of the countries which flew the flag of communism were democratic). Nobody knows how a large democratic communism would work because we've never seen one. Communism relies on central control to allocate resources, and naturally evolves into a fascist state if it reaches a certain size.
Communism isn't evil, just unworkable at large scales. I've heard of moderate successes at small scales.
Scientific research in academia works that way too. It works pretty well.
OpenSource is not crumbling, it's just getting so big that consensus is taking more work (or just not happening). More useful work is being done, it's just being done a little less efficiently with more duplication of effort. Arguments are no sign of decay; the FreeBSD/Linux split, for example, hasn't destroyed the free software movement. A certain amount of duplication of effort makes the system more robust: if Linus had just decided to wait for the HURD kernel, we wouldn't have Linux today, and probably wouldn't have such a well-developed platform for the GNU (and other copylefted) tools.
I don't believe OpenSource is the answer for all software development, but I believe it will coexist well in the future with commercial development, in the same manner that scientific research coexists well with commercial engineering projects. We're still just learning how to divide up the work.
What free software is really about is the realization that the most generalized programs (OS's and compilers, for example) are like learned information, not mechanisms. Hackers are equivalent to researchers, and the time and effort of brilliant researchers is the scarcest resource of them all.
We only got one Einstein, one Newton, one Leibniz (list ended arbitrarily here). We can't make them any faster than a few per century. If somebody locks that information up in a vault and uses it for private purposes only, it is a huge loss for humanity.
If UNIX is a proprietary operating system with a huge license fee, none of the hackers can use it at home, and none of them can distribute the UNIX specific programs they make. This is a brutal blow to the technological development of humanity, perhaps as much as physics textbooks being similarly restricted to a few thousand copies hidden away on the shelves of engineers.
Writing and supporting one Linux is much cheaper in terms of sparce developer time than supporting two dozen proprietary OS's, and giving it to everyone who wants it provides far more value to humanity and ends the whole mess of writing new ones.
Imagine a tractor developed around the turn of the century, like a modern tractor but with a button you can push that instantly creates an identical tractor out of thin air at zero cost. Naturally, you'd expect every farmer to have a tractor. But what if the creator of the first tractor (or his employer) built a factory in which the tractors are created in the same costless manner, but then workers pry out the replicator button before shipping them off and charged the price of 10% of the land it could work? Or worse yet, the tractors are sold under a no-replication licence, and anyone caught replicating a tractor is dragged off to jail, regardless of how many mouths it would feed or hands it would free. Well, you might expect war, but the tractor company would become very wealthy and powerful very quickly, and all the tractor owners would leap to protect their competitive advantage. You would certainly expect other people to expend the huge effort to make their own tractors from scratch, but most of them would also want lots of money and would follow the example of the original. Inevitably, though, those farmers who could not afford to buy tractors would pool their resources and remake the tractor themselves, and would not, of course, place restrictions on themselves or the friends they replicated it for, because they want maximum value and they want to give maximum value. Humanity would then forever have tractors, and would enter the post-agrarian age where only 5% of the population can provide all the food, the survival need for the scarce resource of human labor being greatly reduced.
This analogy applies to arithmetic, trigonometry, Newtonian physics, and UNIX equally well. If it's valuable and cheap to reproduce, eventually it will be free, even if empires are built between the time when one group of people gets it and the time it becomes common property, and even if one group of people has to seemingly act against their own interest in releasing it. It is the lower energy state into which all things are doomed to fall and from which they cannot return.
...neither does communism (to each his needs...). No system has ever worked that way, except within an organization.
But capitalism (and this application, sale/licensing of proprietary software)is the best model we have for fairly compensating developers while placing the costs on the shoulders of those who gain the benefits.
Capitalism, of course, would work a lot better with an informed public who doesn't, say, continue buying minor upgrades to a crappy OS for a ridiculously high profit margin (relative to dev cost/sales) without even considering the many alternatives. Branding and ignorance have a way of destroying the logic of capitalism (similar to buying a Club Monaco shirt for 3X the price of an identical unbranded shirt).
I like FSF-style free software, but it only makes sense for 1) mature products which are really commonly used, like UNIX and spreadsheets, and 2) products which are not all that hard to develop (meaning 1 person could complete the inital product on his own, not that it would be easy for him) and people probably wouldn't bother with unless it was free, like Perl (regardless of how indispensible it has become, do you really think anyone would have bothered with the original Perl with a sticker price of $250 and no source?).
An ideal model would be to compensate all software developers in direct relation to the value of their code. Of course, determining the value is impossible. Market value is a fiction, a number which varies from person to person and from second to second within each person. Real value is even less tangible, if it is, in fact, even defineable. Capitalism is the best we can do, until the cost of development drops below the "Oh, what the hell, (I/we)'ll do it for free" threshold.
I have never been impressed with the honesty of the FSF.
It takes either an exceptionally stupid (which RMS ain't) or blatantly dishonest person to say "free software has nothing to do with money" (quoted directly from the article).
Realistically, FSF/GPL free software is unavoidably free as in "free beer."
For example, I use RedHat Linux, mostly because of rpm binaries all over the place (I know, I could use rpm without RedHat, but it was simpler this way), but I don't pay for it. When you pay for a linux distribution, you are paying for 1) the media, and 2) support. If you are paying more than, say, 150% the cost of the media, you are definitely in support land, because competition will obviously spring up and kill any higher profit margins otherwise (this is what happened with me & RedHat, I didn't want the support so I found the media for about $3).
Stepping back from the business view, you could also call paying for a commercial distribution a form of donation to encourage development. But if I decide to donate money, I'll decide exactly how much I want to give and when I want to give it. I will not have donations required of me.
The only real money to be made in free software is in support or slim profit margin distribution. These are activities of large companies, and do not directly reward the originators or developers of software projects.
A small company or individual which creates a software product can not profit from releasing GPLed software, unless it is tied to sellable proprietary stuff (ex. hardware, other software, or documentation; profit through documentation is risky, risky business, esp. for a small company, O'Reilly could eat you for breakfast).
If it wasn't about money, the licence would not allow free redistribution to all and sundry, but only to others who had purchased the product. It would require the distribution of source at cost only to customers who paid for the binaries. This would be "free" software as in free to maintain, and free to modify, but not as in "free beer." It would deal with every gripe about proprietary software except the price (of course, if the user base is limited because of the price, other gripes would follow, but those apply to unpopular free software too). But you don't exactly see the FSF supplying and promoting a legally valid licence that works like this, do you?
He's been saying all along that a commercial release would happen if manual sales were insufficient, but just as they can't keep up with manual orders they decide to go shareware.
What gives? This is not consistent with previous statements. Apparently he was just kidding when he said there would be no commercial release if manual sales were adequate.
I hate to be the ingrate that points this out, but I think it's much more likely to go GPL if they don't sell many keys or manuals. Food for thought.
Slashdot Mirror
Time in evolution is not measured in years, but in generations. An evolutionary software process can go through millions of generations in a day, compared to humans, which take roughly 20 years for each generation.
Mechanical computers are one of the worst nanotech ideas. They are a "proof of concept" thing to demonstrate that nano-scale computers are possible, because in absence of experiment it is easier to convince people that bumpy nanosticks can compute than that the quantum effects can be dealt with. Drexler needed something possible to point at, I doubt even he seriously thinks mechanical computers are the way of the future.
They are definitely way slower than an equivalent-scaled electronic or light-based design. They are indeed theoretically a bit faster (about 1 GHz and 1000 MIPS) than your current average production chip (when "Engines of Creation" came out, the gap was much wider), but it is a theoretical high estimate which conventional electronic designs have already overtaken.
Realistically, do you think it's more efficient to move around hundreds of whole atoms or a few electrons or photons? (the concept of single-particle signalling is a very exciting one to me)
Incidentally, I have a copy of Nanosystems on my lap as I type this. I've read it, but I don't have the expertise to check his work. It is very thorough and quite interesting if you like theoretical engineering.
In the comics on the website, the human "batteries" are used for other things that make a whole lot more sense. I won't say what, but I will say that they are worth reading, and they only contradict the movie to improve on it, tearing up that "little safety net" you mention.
I agree with the sequel potential, but it would have to be in the real world, and thus it would be a very different movie.
Back in the Doom days, when they talked about Quake, they talked about a real-physics universe where the player is basically Thor and the main weapon is the hammer.
I remember in particular one line about how you were going to be able to knock someone down and then go over and kick the s**t out of him while he was down.
It was like Daikatana, really, in that it revolved around this one apparently primitive weapon that gained power. I think you were supposed to be able to cause local earthquakes with the hammer and open up cracks in the ground or something like that.
Let's face it, nothing in the premise was logical. It was internally consistent, but had no connection to reality. I could pick it apart, but everyone who's seen it and has basic knowledge of computers, biology, and thermodynamics (i.e. /. readers) knows what I'm talking about.
OTOH, I staggered out of the theatre. It released all my adrenaline reserves and left me a hollow shell. I was two blocks away from the theatre before I regained any sense of reality.
The only thing I could compare it to in terms of adrenaline rush is Ninja Scroll (no, I don't suggest they're similar in any other way). And Ninja Scroll didn't pluck at my geeky heartstrings the same way.
However, definitely a one-shot movie. I don't ever want to watch it again. I was blown away in a similar manner by Starship Troopers (which must have been produced by someone who hated the book, BTW), but when I rented it trying to recapture the feeling, I was disappointed by a campy sci-fi flick. I'm sure it would be fun to watch clips of those great fight scenes again (and watch for digital artifacts ^_^), but I'm not sure I could sit through the rest.
Neither China nor Cuba are true democratic communist states. They just have better PR in the US because they weren't the US's main military rival.
It's important to note that the true splits are democracy/dictatorship and capitalism/communism; US propaganda from the cold war days commonly redefined communism as the opposite of democracy (innacurate but fair, because none of the countries which flew the flag of communism were democratic). Nobody knows how a large democratic communism would work because we've never seen one. Communism relies on central control to allocate resources, and naturally evolves into a fascist state if it reaches a certain size.
Communism isn't evil, just unworkable at large scales. I've heard of moderate successes at small scales.
Scientific research in academia works that way too. It works pretty well.
OpenSource is not crumbling, it's just getting so big that consensus is taking more work (or just not happening). More useful work is being done, it's just being done a little less efficiently with more duplication of effort. Arguments are no sign of decay; the FreeBSD/Linux split, for example, hasn't destroyed the free software movement. A certain amount of duplication of effort makes the system more robust: if Linus had just decided to wait for the HURD kernel, we wouldn't have Linux today, and probably wouldn't have such a well-developed platform for the GNU (and other copylefted) tools.
I don't believe OpenSource is the answer for all software development, but I believe it will coexist well in the future with commercial development, in the same manner that scientific research coexists well with commercial engineering projects. We're still just learning how to divide up the work.
What free software is really about is the realization that the most generalized programs (OS's and compilers, for example) are like learned information, not mechanisms. Hackers are equivalent to researchers, and the time and effort of brilliant researchers is the scarcest resource of them all.
We only got one Einstein, one Newton, one Leibniz (list ended arbitrarily here). We can't make them any faster than a few per century. If somebody locks that information up in a vault and uses it for private purposes only, it is a huge loss for humanity.
If UNIX is a proprietary operating system with a huge license fee, none of the hackers can use it at home, and none of them can distribute the UNIX specific programs they make. This is a brutal blow to the technological development of humanity, perhaps as much as physics textbooks being similarly restricted to a few thousand copies hidden away on the shelves of engineers.
Writing and supporting one Linux is much cheaper in terms of sparce developer time than supporting two dozen proprietary OS's, and giving it to everyone who wants it provides far more value to humanity and ends the whole mess of writing new ones.
Imagine a tractor developed around the turn of the century, like a modern tractor but with a button you can push that instantly creates an identical tractor out of thin air at zero cost. Naturally, you'd expect every farmer to have a tractor. But what if the creator of the first tractor (or his employer) built a factory in which the tractors are created in the same costless manner, but then workers pry out the replicator button before shipping them off and charged the price of 10% of the land it could work? Or worse yet, the tractors are sold under a no-replication licence, and anyone caught replicating a tractor is dragged off to jail, regardless of how many mouths it would feed or hands it would free. Well, you might expect war, but the tractor company would become very wealthy and powerful very quickly, and all the tractor owners would leap to protect their competitive advantage. You would certainly expect other people to expend the huge effort to make their own tractors from scratch, but most of them would also want lots of money and would follow the example of the original. Inevitably, though, those farmers who could not afford to buy tractors would pool their resources and remake the tractor themselves, and would not, of course, place restrictions on themselves or the friends they replicated it for, because they want maximum value and they want to give maximum value. Humanity would then forever have tractors, and would enter the post-agrarian age where only 5% of the population can provide all the food, the survival need for the scarce resource of human labor being greatly reduced.
This analogy applies to arithmetic, trigonometry, Newtonian physics, and UNIX equally well. If it's valuable and cheap to reproduce, eventually it will be free, even if empires are built between the time when one group of people gets it and the time it becomes common property, and even if one group of people has to seemingly act against their own interest in releasing it. It is the lower energy state into which all things are doomed to fall and from which they cannot return.
...neither does communism (to each his needs...). No system has ever worked that way, except within an organization.
But capitalism (and this application, sale/licensing of proprietary software)is the best model we have for fairly compensating developers while placing the costs on the shoulders of those who gain the benefits.
Capitalism, of course, would work a lot better with an informed public who doesn't, say, continue buying minor upgrades to a crappy OS for a ridiculously high profit margin (relative to dev cost/sales) without even considering the many alternatives. Branding and ignorance have a way of destroying the logic of capitalism (similar to buying a Club Monaco shirt for 3X the price of an identical unbranded shirt).
I like FSF-style free software, but it only makes sense for 1) mature products which are really commonly used, like UNIX and spreadsheets, and 2) products which are not all that hard to develop (meaning 1 person could complete the inital product on his own, not that it would be easy for him) and people probably wouldn't bother with unless it was free, like Perl (regardless of how indispensible it has become, do you really think anyone would have bothered with the original Perl with a sticker price of $250 and no source?).
An ideal model would be to compensate all software developers in direct relation to the value of their code. Of course, determining the value is impossible. Market value is a fiction, a number which varies from person to person and from second to second within each person. Real value is even less tangible, if it is, in fact, even defineable. Capitalism is the best we can do, until the cost of development drops below the "Oh, what the hell, (I/we)'ll do it for free" threshold.
I have never been impressed with the honesty of the FSF.
It takes either an exceptionally stupid (which RMS ain't) or blatantly dishonest person to say "free software has nothing to do with money" (quoted directly from the article).
Realistically, FSF/GPL free software is unavoidably free as in "free beer."
For example, I use RedHat Linux, mostly because of rpm binaries all over the place (I know, I could use rpm without RedHat, but it was simpler this way), but I don't pay for it. When you pay for a linux distribution, you are paying for 1) the media, and 2) support. If you are paying more than, say, 150% the cost of the media, you are definitely in support land, because competition will obviously spring up and kill any higher profit margins otherwise (this is what happened with me & RedHat, I didn't want the support so I found the media for about $3).
Stepping back from the business view, you could also call paying for a commercial distribution a form of donation to encourage development. But if I decide to donate money, I'll decide exactly how much I want to give and when I want to give it. I will not have donations required of me.
The only real money to be made in free software is in support or slim profit margin distribution. These are activities of large companies, and do not directly reward the originators or developers of software projects.
A small company or individual which creates a software product can not profit from releasing GPLed software, unless it is tied to sellable proprietary stuff (ex. hardware, other software, or documentation; profit through documentation is risky, risky business, esp. for a small company, O'Reilly could eat you for breakfast).
If it wasn't about money, the licence would not allow free redistribution to all and sundry, but only to others who had purchased the product. It would require the distribution of source at cost only to customers who paid for the binaries. This would be "free" software as in free to maintain, and free to modify, but not as in "free beer." It would deal with every gripe about proprietary software except the price (of course, if the user base is limited because of the price, other gripes would follow, but those apply to unpopular free software too). But you don't exactly see the FSF supplying and promoting a legally valid licence that works like this, do you?
He's been saying all along that a commercial release would happen if manual sales were insufficient, but just as they can't keep up with manual orders they decide to go shareware.
What gives? This is not consistent with previous statements. Apparently he was just kidding when he said there would be no commercial release if manual sales were adequate.
I hate to be the ingrate that points this out, but I think it's much more likely to go GPL if they don't sell many keys or manuals. Food for thought.