I have frequently seen cars dirving at the average speed of traffic reach the end of a long passageway at the same time as another car that, whenever a space opens in an adjacent lane, speeds up to get ahead.
I'll grant you that if the road is empty of traffic, then a car that goes twice as fast as another gets to the end in half the time. That's not a common occurance. And even then, the advantage of 75 over 55 is nowhere near that large. And over 65 it's even less. (15/13's as fast. I.e., rarely significant.)
7 Develop the EU as a reference player for a democratic and neutral governance of the internet
That sounds pretty simple. Just start up some new root servers with new high-level domains, say: biz. fct, grp instead of com, edu, org REPLACE ICANN! Adopt the current standards without change for those domains, and then start tinkering. (Tinker while you only have a very few web pages.) Decide whether you want to be exclusively IPv6. Swtich to utf-8 identifiers, with some way to indicate whether or not the character is a part of the BLP. (Yeah, that still gives preference to ASCII-7 chars, but that works pretty well for most EU languages. Not so well for Greek, but they are a minor power.)
A rule of thumb is the smaller the country, the less interested it is on spying on it's citizens. This isn't guaranteed, of course. It's also true that the more honest the police are, the less the country is interested in spying on its citizens...but this can be hard to determine.
OTOH, ALL governments are interested in spying on their citizens. No exceptions. Not even at the city level. How much they do about it varies. The larger and more powerful they are, the more they do about it.
All skyhooks depend on a large mass in orbit to hang from. The orbital mass takes the hit as things are lifted, and rises as things are lowered. You've got to conserve angular momentum.
On Earth, I agree. On the Moon? That means you need to fly up to the contact point without air to support you. This makes connecting quite tricky. And it means you need rockets (or some other, unidentified approach) to get the altitude. This doesn't strike me as a good way to send volumes of freight. Every miss causes a new impact crater. (On earth you can either have a parachute or wings to get back with.)
P.S.: As with all skyhooks, you need to ship equal masses up and down over time, or the orbit of the skyhook decays. Which, if you're planning on shipping lots of freight, a catapult might be a better answer.
I think you're wrong. Import/export between Earth and Space is going to be extremely expensive until we build some form of skyhook (personally I favor starting with a PinWheel). We can't do that until we start moving asteroids. So that's not an early stage of the process.
And that means that stuff made in/obtained from space is going to be sufficiently expensive that it won't have noticable effect on local pollution.
Yeah, but to profitably get materials back from the moon you need a catapult. Or, possibly, a beanstalk. Neither of those are cheap to build. And the moon seem not to have many readily accessible minerals. (It's got other advantages, and I can see building an industrial plant there to support a colony established for other reasons, but not to ship things back to Earth. At least not this decade, and probably not the next. But it would be a great place for a radio telescope, and also for a regular telescope, though space is better for that. Still, it would eliminate the station-keeping problem. [Note that these should both be built on the far face of the moon, to escape noise from Earth.])
Solar panels are foolish for that application. Use a large curved mirror. If you spin it right, it could be quite thin, so even aluminized mylar would work, until the radiation destroyed it. Aluminum foil would work indefinitely. (You need weights around the rim, and some weak springs, to ensure that it folds out into the right shape. Not difficult. IIRC, it's already been done to test out a solar sail.)
Note that the mirror would be a bit better if it were thicker, and you could use a half-cylinder and not spin it, but that one is probably better built on site. And you will need solar cells to provide electricity, so not just the mirror. But the mirror is the best way to get heat for smelting.
If you know you can handle it in ASCII, then you can handle utf8 just like the ASCII. If you don't know you can handle it in ASCII, then you can handle it in utf8 more easily.
I will agree than handling utf8 in C++ is difficult, which is a part of why I said that it was poorly implemented. And nobody EVER displays enough text for the difference between speed of utf8 and ascii display to make any difference. I don't think it's possible, due to lags in non-computational parts of the system. I will agree that for some embedded systems the difference in the amount of memory required by the code can be significant. The number of cases where that is significant, however, is so small that it's irrelevant, and should be handled in some other way. (There are already dialects that strip out some parts of the language, so it's not an insuperable problem.)
If you use utf8, then unicode is as efficient as ascii...for everything you can do with ascii. Nobody should be required to use utf32 except when it's the most convenient choice. (And there's never a good argument for using utf16, unless you're on a computer with 16 bit words.)
Storing it in ROM wouldn't suffice, though it would help a lot. I think your first statement was better: "it's not possible to design a perfect computer system".
Umh... US parties *claim* to differ on a large number of areas. When the party out of power gets into power, however, the same policies are continued, with one minor change. The Republicans are generally less sensitive to how much the majority of the citizens like them. There are exceptions, usually on the Republican side. Regan, e.g., was very interested in people liking him. So was Eisenhower. (His very campaign slogan was "I like Ike".) I *think* Obama is concerned with, just not very capable of, getting people to like him. This leads to a derived difference where the Republicans tend to institute repressive policies, and the Democrats make use of them.
We *probably* could. It's not certain. We have not yet solved the problems of building a closed eco-system. (I.e., one that only needs resources that can be obtained locally. At totally closed one is an obvious impossibility, or thermodynamics is wrong.)
Please note that the MIR depends on a continual stream of supplies from Earth. So that's not an example. BioSphere 2 was a disaster, because they hadn't predicted that concrete would be absorbing CO2. Etc.
Now I don't *think* that this is an intractable problem, but it's one that hasn't been seriously tackled, and until it is cracked long-term presence off the earth is not feasible. And until you've actually solved it, you don't know how hard the problem is going to be.
IIRC though (this is several decades ago) you only get the simplified form by dropping higher order terms. There was a good reason to do so (IIRC they were *extremely* small in value), but it does mean that it's an approximation. However, since even the deviation of Einsteinian theory from Newtonian is difficult to detect, I expect that the corrected version would differ from the Einsteinian version to such a slight degree that it would still be impossible to detect. And (again, IIRC) there was an infinite series of such terms, each (nearly) infinitesimally small when compared to the prior term. So you couldn't find the exact value no matter what you did. And the length of the calcularions would more than double for each term you added.
Sorry, but though the claim was made General Relativity didn't predict the correct deviation of Mercury's orbit. The key turned out to be the flattening of the sun at the poles. (I think once that was taken into account, General Relativity gave a better answer than Newton, but the Newtonian answer was good enough that it wouldn't have been noticed originally.)
He's probably insisting on a more direct form of observation, even though he actually knows this is impossible. (Well, perhaps not, if the new lens design that allows one to see things fractions of a wave length long can be built at the right wave length. I've got my doubts...and even then the wavelength would be too short to see directly.)
OTOH, I'm also not very pleases with the quality of the evidence that we need to deal with, and would like something a lot less indirect. But right now that's the best we can do. So we are only "statistically certain" that certain observations mean what we think they mean. The statistics are conseervative enough that we can place a great deal of reliance on them, but...
Yet again, even direct physical observation of things of the appropriate size and opacity can't really be trusted, as was proved by experiments where a guy was stabbed with a banana, or a guy wandered through a basket ball game wearing a gorilla suit. (The cameras saw him, but very few of the observers did. They were too busy counting baskets.)
So I understand people being dissatisfied by "statistical certainty", but really, it can be better than eye-witness observation...at least unless that "eye-witness observation" is backed up by a camera.
That will give the wrong answer. The reason is step size. Remember the relative positions are constantly changing, except for certain special cases.
FWIW, professional astronomers can't do an accurate model of the solar sysstem. The can do one that's "almost correct", but that's not the same. And the solar system can be handled with Newtonian mechanics. Once you bring in pulsars emitting gravity waves...Yi!
P.S.: IIRC, the best models of the solar system can't predict on which side of the sun the earth will be located in exactly 100 million years from now. When you project further, the results get fuzzier.
OTOH, in this case we're only looking at three stars. We're ignoring minor dust like planets. So it's guaranteed that we won't be able to find really minor effects. But, IIRC, the expectation is that emission of gravity waves will cause the orbits to decay more rapidly. (And I seem to also recall that in a previous case that effect wasn't found, but they weren't sure their observations were sensitive enough to have found it.) So this could be really interesting.
Please note: I am not an astronomer of any variety. I just read populaarizations. So I can't point you to any original sources any better than Google could.
There are a large number of misfeatures. Compare the C11 std implementation with Vala, Python, D, hell, even Java. (I don't like Java's implementation because it uses the utf16 unicode, but the design is good.)
There's no reason that a string should be modifiable, but if you don't allow that, you need to allow arrays of char that can be easily converted into a string. Some languages pick one choice, some another, both are reasonable. (Unmodifiable strings are desireable because it makes compiling/designing for parallel execution easier and safer. How you achieve it is less important.)
Perhaps some of my distaste with the C++11 std unicode handling relates to my dislike of C++ templates, but when I tried to use the C++11 unicode, I really hated it, unlike that of any other language that implemented it. I'd actually rather use byte strings and the Unicode std. library, which means it's (from my point of view) a really terrible design. I literally moved over to doing the project in C before deciding that I didn't need efficiency enough to fight that battle. So that part is not being done in Python. (I could have done it in D, but I'd need too many unsupported external libraries.)
Yeah, but C++ has made a series of bad choices. Now it's to the point where I would prefer to use Ada. Actually, neither is my preference, but C++ is a nightmare. It's ok as long as it's just my own code, but there are so many dangerouly "neat" ways to do things, that it's an abomination.
Hell, even C, with free use of the macro processor, can easily turn into something incomprehensible. And they guy who did that can justify each choice. C++ takes that up an octave. Even the guy who did it is often unable to explain what it was intended to do, even if the code is successfully doing it.
I will grant that custom subsets of C/C++ can be good, and even optimal, for a given problem or SMALL group of users. But that's not what you commonly see. So, depending on the problem, I tend to prefer Python, Ada, or D...and very occasionally Java. C I mainly use for linking libraries and routines in different languages.
P.S.: I wonder if Vala will ever become usable. It's a language design with great promise. But the last time I checked they hadn't even documented many of the most basic features. Perhaps they presume that "well, since it's almost like C, we don't really need to document it", but to me it just looks incomplete.
More to the point, why don't they implement decent Unicode handling. That's much more suitable to a compiler standard than graphics. (Actually, the reason is probably because MS and Java settled on a totally rediculous implementation of Unicode [it was more reasonable at the time], and you'd need to get one of them to change to a sane standard. Either utf8, utf32, or optionally either would work, though. There are arguments in favor of each, but utf16 is indefensible.
You could give D a look. ( http://dlang.org/index.html ). To me it seems like one of the best languages that compile "to the metal". It's problem is the perennial one of new languages: Not enough libraries. Of course, you can link against any library with a C calling convention (and I think those with a C++ calling convention), but there are problems with data structure translation. Even strings require special handling. And it doesn't recognize C/C++ macros, so all the header files need to be translated...which is a bit of a difficult problem, and currently each case needs to be handled manually. Etc.
It's enough of a problem that I'm not doing my current project in D. But I still wish I could.
Slashdot Mirror
I have frequently seen cars dirving at the average speed of traffic reach the end of a long passageway at the same time as another car that, whenever a space opens in an adjacent lane, speeds up to get ahead.
I'll grant you that if the road is empty of traffic, then a car that goes twice as fast as another gets to the end in half the time. That's not a common occurance. And even then, the advantage of 75 over 55 is nowhere near that large. And over 65 it's even less. (15/13's as fast. I.e., rarely significant.)
7 Develop the EU as a reference player for a democratic and neutral governance of the internet
That sounds pretty simple. Just start up some new root servers with new high-level domains, say:
biz. fct, grp instead of com, edu, org
REPLACE ICANN!
Adopt the current standards without change for those domains, and then start tinkering. (Tinker while you only have a very few web pages.) Decide whether you want to be exclusively IPv6. Swtich to utf-8 identifiers, with some way to indicate whether or not the character is a part of the BLP. (Yeah, that still gives preference to ASCII-7 chars, but that works pretty well for most EU languages. Not so well for Greek, but they are a minor power.)
You're a bit more confident than I am. They won't grab him, but they might assassinate him. If they're feeling snarky they could use Polonium.
A rule of thumb is the smaller the country, the less interested it is on spying on it's citizens. This isn't guaranteed, of course. It's also true that the more honest the police are, the less the country is interested in spying on its citizens...but this can be hard to determine.
OTOH, ALL governments are interested in spying on their citizens. No exceptions. Not even at the city level. How much they do about it varies. The larger and more powerful they are, the more they do about it.
All skyhooks depend on a large mass in orbit to hang from. The orbital mass takes the hit as things are lifted, and rises as things are lowered. You've got to conserve angular momentum.
On Earth, I agree. On the Moon? That means you need to fly up to the contact point without air to support you. This makes connecting quite tricky. And it means you need rockets (or some other, unidentified approach) to get the altitude. This doesn't strike me as a good way to send volumes of freight. Every miss causes a new impact crater. (On earth you can either have a parachute or wings to get back with.)
P.S.: As with all skyhooks, you need to ship equal masses up and down over time, or the orbit of the skyhook decays. Which, if you're planning on shipping lots of freight, a catapult might be a better answer.
I think you're wrong. Import/export between Earth and Space is going to be extremely expensive until we build some form of skyhook (personally I favor starting with a PinWheel). We can't do that until we start moving asteroids. So that's not an early stage of the process.
And that means that stuff made in/obtained from space is going to be sufficiently expensive that it won't have noticable effect on local pollution.
Yeah, but to profitably get materials back from the moon you need a catapult. Or, possibly, a beanstalk. Neither of those are cheap to build. And the moon seem not to have many readily accessible minerals. (It's got other advantages, and I can see building an industrial plant there to support a colony established for other reasons, but not to ship things back to Earth. At least not this decade, and probably not the next. But it would be a great place for a radio telescope, and also for a regular telescope, though space is better for that. Still, it would eliminate the station-keeping problem. [Note that these should both be built on the far face of the moon, to escape noise from Earth.])
Solar panels are foolish for that application. Use a large curved mirror. If you spin it right, it could be quite thin, so even aluminized mylar would work, until the radiation destroyed it. Aluminum foil would work indefinitely. (You need weights around the rim, and some weak springs, to ensure that it folds out into the right shape. Not difficult. IIRC, it's already been done to test out a solar sail.)
Note that the mirror would be a bit better if it were thicker, and you could use a half-cylinder and not spin it, but that one is probably better built on site. And you will need solar cells to provide electricity, so not just the mirror. But the mirror is the best way to get heat for smelting.
If you know you can handle it in ASCII, then you can handle utf8 just like the ASCII. If you don't know you can handle it in ASCII, then you can handle it in utf8 more easily.
I will agree than handling utf8 in C++ is difficult, which is a part of why I said that it was poorly implemented. And nobody EVER displays enough text for the difference between speed of utf8 and ascii display to make any difference. I don't think it's possible, due to lags in non-computational parts of the system. I will agree that for some embedded systems the difference in the amount of memory required by the code can be significant. The number of cases where that is significant, however, is so small that it's irrelevant, and should be handled in some other way. (There are already dialects that strip out some parts of the language, so it's not an insuperable problem.)
If you use utf8, then unicode is as efficient as ascii...for everything you can do with ascii. Nobody should be required to use utf32 except when it's the most convenient choice. (And there's never a good argument for using utf16, unless you're on a computer with 16 bit words.)
Storing it in ROM wouldn't suffice, though it would help a lot. I think your first statement was better: "it's not possible to design a perfect computer system".
Umh... US parties *claim* to differ on a large number of areas. When the party out of power gets into power, however, the same policies are continued, with one minor change. The Republicans are generally less sensitive to how much the majority of the citizens like them. There are exceptions, usually on the Republican side. Regan, e.g., was very interested in people liking him. So was Eisenhower. (His very campaign slogan was "I like Ike".) I *think* Obama is concerned with, just not very capable of, getting people to like him. This leads to a derived difference where the Republicans tend to institute repressive policies, and the Democrats make use of them.
Thanks. As I said, it's been a LONG time since I did this.
Maybe you should dig the relief tube with nukes. That way the tube would be large enough that there'd be no need for a volcano explosion. Plowshare!
We *probably* could. It's not certain. We have not yet solved the problems of building a closed eco-system. (I.e., one that only needs resources that can be obtained locally. At totally closed one is an obvious impossibility, or thermodynamics is wrong.)
Please note that the MIR depends on a continual stream of supplies from Earth. So that's not an example. BioSphere 2 was a disaster, because they hadn't predicted that concrete would be absorbing CO2. Etc.
Now I don't *think* that this is an intractable problem, but it's one that hasn't been seriously tackled, and until it is cracked long-term presence off the earth is not feasible. And until you've actually solved it, you don't know how hard the problem is going to be.
IIRC though (this is several decades ago) you only get the simplified form by dropping higher order terms. There was a good reason to do so (IIRC they were *extremely* small in value), but it does mean that it's an approximation. However, since even the deviation of Einsteinian theory from Newtonian is difficult to detect, I expect that the corrected version would differ from the Einsteinian version to such a slight degree that it would still be impossible to detect. And (again, IIRC) there was an infinite series of such terms, each (nearly) infinitesimally small when compared to the prior term. So you couldn't find the exact value no matter what you did. And the length of the calcularions would more than double for each term you added.
Sorry, but though the claim was made General Relativity didn't predict the correct deviation of Mercury's orbit. The key turned out to be the flattening of the sun at the poles. (I think once that was taken into account, General Relativity gave a better answer than Newton, but the Newtonian answer was good enough that it wouldn't have been noticed originally.)
He's probably insisting on a more direct form of observation, even though he actually knows this is impossible. (Well, perhaps not, if the new lens design that allows one to see things fractions of a wave length long can be built at the right wave length. I've got my doubts...and even then the wavelength would be too short to see directly.)
OTOH, I'm also not very pleases with the quality of the evidence that we need to deal with, and would like something a lot less indirect. But right now that's the best we can do. So we are only "statistically certain" that certain observations mean what we think they mean. The statistics are conseervative enough that we can place a great deal of reliance on them, but...
Yet again, even direct physical observation of things of the appropriate size and opacity can't really be trusted, as was proved by experiments where a guy was stabbed with a banana, or a guy wandered through a basket ball game wearing a gorilla suit. (The cameras saw him, but very few of the observers did. They were too busy counting baskets.)
So I understand people being dissatisfied by "statistical certainty", but really, it can be better than eye-witness observation...at least unless that "eye-witness observation" is backed up by a camera.
That will give the wrong answer. The reason is step size. Remember the relative positions are constantly changing, except for certain special cases.
FWIW, professional astronomers can't do an accurate model of the solar sysstem. The can do one that's "almost correct", but that's not the same. And the solar system can be handled with Newtonian mechanics. Once you bring in pulsars emitting gravity waves...Yi!
P.S.: IIRC, the best models of the solar system can't predict on which side of the sun the earth will be located in exactly 100 million years from now. When you project further, the results get fuzzier.
OTOH, in this case we're only looking at three stars. We're ignoring minor dust like planets. So it's guaranteed that we won't be able to find really minor effects. But, IIRC, the expectation is that emission of gravity waves will cause the orbits to decay more rapidly. (And I seem to also recall that in a previous case that effect wasn't found, but they weren't sure their observations were sensitive enough to have found it.) So this could be really interesting.
Please note: I am not an astronomer of any variety. I just read populaarizations. So I can't point you to any original sources any better than Google could.
There are a large number of misfeatures. Compare the C11 std implementation with Vala, Python, D, hell, even Java. (I don't like Java's implementation because it uses the utf16 unicode, but the design is good.)
There's no reason that a string should be modifiable, but if you don't allow that, you need to allow arrays of char that can be easily converted into a string. Some languages pick one choice, some another, both are reasonable. (Unmodifiable strings are desireable because it makes compiling/designing for parallel execution easier and safer. How you achieve it is less important.)
Perhaps some of my distaste with the C++11 std unicode handling relates to my dislike of C++ templates, but when I tried to use the C++11 unicode, I really hated it, unlike that of any other language that implemented it. I'd actually rather use byte strings and the Unicode std. library, which means it's (from my point of view) a really terrible design. I literally moved over to doing the project in C before deciding that I didn't need efficiency enough to fight that battle. So that part is not being done in Python. (I could have done it in D, but I'd need too many unsupported external libraries.)
Could it be that the "flamebait" moderation was itself a joke? Think about it.
Yeah, but C++ has made a series of bad choices. Now it's to the point where I would prefer to use Ada. Actually, neither is my preference, but C++ is a nightmare. It's ok as long as it's just my own code, but there are so many dangerouly "neat" ways to do things, that it's an abomination.
Hell, even C, with free use of the macro processor, can easily turn into something incomprehensible. And they guy who did that can justify each choice. C++ takes that up an octave. Even the guy who did it is often unable to explain what it was intended to do, even if the code is successfully doing it.
I will grant that custom subsets of C/C++ can be good, and even optimal, for a given problem or SMALL group of users. But that's not what you commonly see. So, depending on the problem, I tend to prefer Python, Ada, or D...and very occasionally Java. C I mainly use for linking libraries and routines in different languages.
P.S.: I wonder if Vala will ever become usable. It's a language design with great promise. But the last time I checked they hadn't even documented many of the most basic features. Perhaps they presume that "well, since it's almost like C, we don't really need to document it", but to me it just looks incomplete.
More to the point, why don't they implement decent Unicode handling. That's much more suitable to a compiler standard than graphics. (Actually, the reason is probably because MS and Java settled on a totally rediculous implementation of Unicode [it was more reasonable at the time], and you'd need to get one of them to change to a sane standard. Either utf8, utf32, or optionally either would work, though. There are arguments in favor of each, but utf16 is indefensible.
You could give D a look. ( http://dlang.org/index.html ). To me it seems like one of the best languages that compile "to the metal". It's problem is the perennial one of new languages: Not enough libraries. Of course, you can link against any library with a C calling convention (and I think those with a C++ calling convention), but there are problems with data structure translation. Even strings require special handling. And it doesn't recognize C/C++ macros, so all the header files need to be translated...which is a bit of a difficult problem, and currently each case needs to be handled manually. Etc.
It's enough of a problem that I'm not doing my current project in D. But I still wish I could.