This is one possible definition, but it's hardly the only interesting one. Lots of people have worked on Turing Machines using minimal numbers of states and symbols which are universal under very convoluted encodings. Requiring the "same" encoding may not be possible, for example in the Turing Machine case, there is no definitive simplest encoding from a set of symbols on a tape into a state table. Therefore by your definition, it is impossible to make a Universal Turing Machine.
Export regs specifically don't apply to authentication systems. If your reg code system can only be used for authentication, it is OK to export. I imagine this would almost universally be the case.
Bush has stated that he wants to resume underground testing. I'm not fully informed on all of the politics that went into creating the ASCI program, but I don't think that the Republicans would want to halt any program that has to do with nuclear weapons. The fundamental fact remains that contrary to the original poster's claims, Bush had nothing to do with this. The CTBT was the primary motivation for buying these huge computers, and Clinton was the one backing the CTBT.
The reason for the big push in computational testing is the comprehensive test ban treaty, which was supported by the Democrats, and opposed by the Republicans. The main reason for the ASCI program is the CTBT, which is now dead in the water, due to Republican opposition.
Well, the Swiss (CERN) used to control web standards way back in 1994 or so. Then they announced that they didn't want to employ Tim Berners-Lee anymore or continue development of standards, the cern http server, or the line-mode browser. TBL then, with CERN's blessing went to MIT to start the W3C, which currently controls web standards. So blame the Swiss. Oh, and by the way, the U.S. is the first modern Republic, since way back when you guys across the pond still had kings.
I think most U.S. users are more concerned with being able to roam in most of the U.S., which the Kyocera offers, but the Treo does not. GSM coverage is pretty poor in the U.S. Most users can live with getting another phone when they cross the pond, but not when they drive fifty miles out of town.
I don't understand how Handspring expects to sell too many of these with single-mode GSM. GSM only may be fine in Europe where everything is GSM, but in the USA, if you travel, you need to be able to fall back to Analog. A great part of the utility of having a mobile phone is being able to use it while travelling, particularly in a semi-remote area. You would have to buy another phone for backup with the Treo. It's built in on the Kyocera or Samsung phones.
Yes, Trumpet winsock was the best I found, but it constantly dropped connections and crashed. The Win95 TCP/IP was much better -- more reliable and easier to configure as it knew how to log into most terminal servers without a script.
I thought a command economy was one of the reasons that the Soviets failed. Humans can't manage such huge and complex pieces of social machinery. Nature has to be left to take its course. However, now we're doing the same damn thing, except it's the corporation instead of the government. I fail to see why the same rules don't apply.
And you can't statically link against it because of lgpl, but there are different bugs in every version. Then you tell yourself, self, why don't you just ship a version of glibc with the package and dynamically link? Well, it turns out that ld-linux.so is also a buggy piece of crap, and different versions of ld-linux will only work with certain versions of glibc. Then people whine that commercial apps say "requires redhat 6.2." THIS IS THE REASON WHY!
You're making it excessively complicated. First, laws don't have to make logical sense, they just have to be passed by congress. They can prohibit you from doing almost anything. All that needs to be done is to extend copyright law, or have a judicial opinion which says that you can't display a page except as authorized by the copyright holder. Throw in a new anti-circumvention law ala DMCA, and it is illegal to own ad-blocking software. Presto. You are arguing that each of your five points is impossible. The government simply has to rule that you must do each of your five things without any compensation from the government.
Now ads are starting to show up which don't have window frames around them. This may be an IE specific "feature", however, they don't need to have a close button on them. All I've seen so far have close links on them, but this is clearly not a requirement. Once these come into widespread use, nearly everyone will get ad-blockers. Unfortunately, this will probably mean that a law will be passed outlawing ad-blocking proxies.
This of course assumes the Church-Turing thesis. I think that Kolmogorov complexity is still a valuable concept, however, given stuff like Chaitin's Omega, which would appear to have an undefined K. entropy, it is difficult to give priority to computation. I think that in the next hundred years we may see Church-Turing disproved.
Having worked on Mosaic for a time, I can say that Microsoft didn't get any advantage in their later releases from it. IE 1.0 was largely Mosaic. I can't believe that there's any substantial Mosaic code left in 5.0. Remember that this was a browser that until near the end of its life couldn't even display a page at all until all of the images were completely loaded.
You use a satellite with a mirror on it. The only problem is the atmosphere degrades the signal which means that you have to sacrifice some security for error correction. This isn't a problem unless your opponent owns the satellite and has extremely sophisticated technology.
I don't know what kind of codes you are compiling, however in my experience VC++ produces much faster code than GCC. You can clearly see this if you look at the assembly that GCC produces from a tight loop. It is garbage. GCC doesn't know how to deal with a machine that is highly register-constrained such as x86. Now if you want the best performance, you use the Intel compiler, which is far better than either GCC or VC++. The Intel compiler is also now available on Linux.
This is one possible definition, but it's hardly the only interesting one. Lots of people have worked on Turing Machines using minimal numbers of states and symbols which are universal under very convoluted encodings. Requiring the "same" encoding may not be possible, for example in the Turing Machine case, there is no definitive simplest encoding from a set of symbols on a tape into a state table. Therefore by your definition, it is impossible to make a Universal Turing Machine.
Faster seek times
Export regs specifically don't apply to authentication systems. If your reg code system can only be used for authentication, it is OK to export. I imagine this would almost universally be the case.
When google doesn't work for me, I go to Teoma.
Teoma often provides better results than google.
I believe that should be:
$ = Mhz * 2($/Mhz)
$ = Mhz/Mhz * 2 $
$ = 2 $
Which, if we stop there, is the mathematical basis for Enron.
However this can be simplified to:
$ = 0
Which is the mathematical basis for WTO protesting hippies.
Therefore:
Ken Lay = Dirty Tree Hugger.
Bush has stated that he wants to resume underground testing. I'm not fully informed on all of the politics that went into creating the ASCI program, but I don't think that the Republicans would want to halt any program that has to do with nuclear weapons. The fundamental fact remains that contrary to the original poster's claims, Bush had nothing to do with this. The CTBT was the primary motivation for buying these huge computers, and Clinton was the one backing the CTBT.
The reason for the big push in computational testing is the comprehensive test ban treaty, which was supported by the Democrats, and opposed by the Republicans. The main reason for the ASCI program is the CTBT, which is now dead in the water, due to Republican opposition.
It's hard to base an entire economy on the Military/Industrial complex when you only spend 3.2% of GDP on it.
Well, the Swiss (CERN) used to control web standards way back in 1994 or so. Then they announced that they didn't want to employ Tim Berners-Lee anymore or continue development of standards, the cern http server, or the line-mode browser. TBL then, with CERN's blessing went to MIT to start the W3C, which currently controls web standards. So blame the Swiss. Oh, and by the way, the U.S. is the first modern Republic, since way back when you guys across the pond still had kings.
You're grasping at ideas which you don't understand. You have made several factually incorrect statements. Do some more reading.
I thought you were joking about relativistic effects from going so fast.
I think most U.S. users are more concerned with being able to roam in most of the U.S., which the Kyocera offers, but the Treo does not. GSM coverage is pretty poor in the U.S. Most users can live with getting another phone when they cross the pond, but not when they drive fifty miles out of town.
I don't understand how Handspring expects to sell too many of these with single-mode GSM. GSM only may be fine in Europe where everything is GSM, but in the USA, if you travel, you need to be able to fall back to Analog. A great part of the utility of having a mobile phone is being able to use it while travelling, particularly in a semi-remote area. You would have to buy another phone for backup with the Treo. It's built in on the Kyocera or Samsung phones.
I guess you aren't on the right lists then Jon. Too bad they're invite-only.
Yes, Trumpet winsock was the best I found, but it constantly dropped connections and crashed. The Win95 TCP/IP was much better -- more reliable and easier to configure as it knew how to log into most terminal servers without a script.
I thought a command economy was one of the reasons that the Soviets failed. Humans can't manage such huge and complex pieces of social machinery. Nature has to be left to take its course. However, now we're doing the same damn thing, except it's the corporation instead of the government. I fail to see why the same rules don't apply.
And you can't statically link against it because of lgpl, but there are different bugs in every version. Then you tell yourself, self, why don't you just ship a version of glibc with the package and dynamically link? Well, it turns out that ld-linux.so is also a buggy piece of crap, and different versions of ld-linux will only work with certain versions of glibc. Then people whine that commercial apps say "requires redhat 6.2." THIS IS THE REASON WHY!
You're making it excessively complicated. First, laws don't have to make logical sense, they just have to be passed by congress. They can prohibit you from doing almost anything. All that needs to be done is to extend copyright law, or have a judicial opinion which says that you can't display a page except as authorized by the copyright holder. Throw in a new anti-circumvention law ala DMCA, and it is illegal to own ad-blocking software. Presto. You are arguing that each of your five points is impossible. The government simply has to rule that you must do each of your five things without any compensation from the government.
Now ads are starting to show up which don't have window frames around them. This may be an IE specific "feature", however, they don't need to have a close button on them. All I've seen so far have close links on them, but this is clearly not a requirement. Once these come into widespread use, nearly everyone will get ad-blockers. Unfortunately, this will probably mean that a law will be passed outlawing ad-blocking proxies.
This of course assumes the Church-Turing thesis. I think that Kolmogorov complexity is still a valuable concept, however, given stuff like Chaitin's Omega, which would appear to have an undefined K. entropy, it is difficult to give priority to computation. I think that in the next hundred years we may see Church-Turing disproved.
Having worked on Mosaic for a time, I can say that Microsoft didn't get any advantage in their later releases from it. IE 1.0 was largely Mosaic. I can't believe that there's any substantial Mosaic code left in 5.0. Remember that this was a browser that until near the end of its life couldn't even display a page at all until all of the images were completely loaded.
Grape6 is 32Tflops, not 32Gflops. You are off by three orders of magnitude. COTS cannot achieve anywhere near this level of speed currently.
You use a satellite with a mirror on it. The only problem is the atmosphere degrades the signal which means that you have to sacrifice some security for error correction. This isn't a problem unless your opponent owns the satellite and has extremely sophisticated technology.
I don't know what kind of codes you are compiling, however in my experience VC++ produces much faster code than GCC. You can clearly see this if you look at the assembly that GCC produces from a tight loop. It is garbage. GCC doesn't know how to deal with a machine that is highly register-constrained such as x86. Now if you want the best performance, you use the Intel compiler, which is far better than either GCC or VC++. The Intel compiler is also now available on Linux.