I don't believe R can actually handle 64-bit integers, either, at least not in any easy way. Its 'numeric' type is internally a double, and even when you coerce it to an integer, it just becomes in effect the 53-bit integer part of a double.
Depends on the kind of antitrust suit. Certainly you can't be guilty of monopoly leveraging unless you have a monopoly, so you're right as far as that goes. =] And it's also true that that's generally the easiest kind of antitrust claim to prove (it's what Microsoft was accused of), because there's a very strong presumption that if: 1. you're a monopoly; and 2. you're leveraging it; then that's bad, regardless of what "legitimate reasons" you have for doing so. (It used to even be considered always illegal to leverage a monopoly, but recent Supreme Court decisions have chipped away at that.)
There are other kinds of anticompetitive trade practices, though, with lower threshholds. At the one end of the scale is being a monopoly, where your conduct is closely scrutinized; at the other end is being a tiny player, whose market power is so small that any claim you were engaged in anticompetitive behavior is implausible. But in between there's a whole area of restraint-of-trade and anticompetitive tying, where you can be guilty of an antitrust violation without being a monopoly, if you have sufficient market power to influence another market improperly, and you did in fact do so. For example, some car manufacturers lost antitrust cases relating to car parts, despite not having a monopoly on automobiles: courts found that e.g. Mercedes banning anyone from making Mercedes-compatible parts was anticompetitive tying between the car market and the replacement-parts market, even though Mercedes isn't a car monopoly.
Those kinds of cases often come down to what the intent of the tying was. Mercedes argued that they wanted to control the replacement-parts market not for anticompetitive reasons, but for quality-assurance reasons; the courts ended up not buying that. But often courts do buy arguments that there was a legitimate reason for tying (possibly even most of the time; non-monopoly tying prosecutions succeeding isn't that common). In Apple's case, they would presumably argue that the purpose of requiring XCode and certain languages isn't mainly to restrain trade in dev-tools markets or to make it harder to port apps, but because of reasons related to its app-review process ("it's easier to review apps if they're all in Obj-C on XCode" or something). I could see a court giving reasonable deference to that, though predictions are always dodgy. The main place I could see that failing is if there's some smoking-gun email that ends up as evidence, where a VP or someone says, "hey we should institute this new requirement because of [reasons that sound like restraint of trade]".
Most (not all) antitrust legislation is aimed at preventing monopoly exploitation of alternate markets.
This falls into the "not all" category though, I think, since it's an investigation of whether Apple's engaged in anticompetitive tying. The iPhone / Apple devtools tying is somewhat reminiscent of the car-makers / car-parts antitrust investigations, which happened even though no carmaker actually held a monopoly in the car market.
I think it'd be a problem for Apple if it were possible to show (e.g. by discovering an unwise internal memo) that the main purpose of the restrictions is to restrain trade, e.g. to harm the market for competitor development tools, or to make it harder to port apps to/from Android. Apple would want to be able to show that the tying was done for legitimate reasons not related to restraining trade.
I'd consider the legitimate scope of that law to be investigating whether they used the grant funds to by themselves a private yacht, and similar cases of fraudulent misuse of grant funds--- not the state investigating the scientific merits of the work.
That's basically what the United States Code is. Laws are compiled and sectioned into logical areas, so e.g. a bill that relates to both bankruptcy law and mining regulation (many laws have absurd things packaged together), will be refactored so the bankruptcy parts get codified into Title 11, while the mining regulations go into Title 30.
There are some decent points there, but a lot of them aren't really related to IEEE 754 compatibility. For example, bullet point #5 on their first-page list of five "gratuitous mistakes" is that Java doesn't support operator overloading. But by that standard, C sucks too, and yet is somehow used in lots of floating-point libraries.
By your logic, we wouldn't need any integer type longer than 2 bits. You could certainly design an integer arithmetic scheme on that basis, but I doubt you'd want to.
I think the argument is more that, in practice, 32 bits is a decent sweet spot for the changeover from native ints to arbitrary-precision bigints (whereas 2 bits is not a sweet spot). Are there that many cases where someone needs integers between 32 and 64 bits, but doesn't need to account for the possibility of >64-bit integers, and therefore doesn't need to use routines that can handle arbitrary-size integers? If there were a lot of such cases, changing from 32- to 64-bit as the switchover point from native to bigint arithmetic would be justified. But the guess is that those cases aren't common, at least among matlab users.
Probably because it is politicizing science regardless of the merits. The way science operates is not generally by having attorneys general investigating the merits of scientific papers. If something was wrong or fraudulent, that's a job for journal editorial staff and university misconduct boards to sort out.
Similarly, it'd be correctly considered "politicizing science" if democrats launched a fraud investigation of a libertarian economist, regardless of whether that economist did or didn't fabricate evidence. The attorney general is just not the right person to do it.
That seems vaguely reasonable to me, based on my experience getting people to do things. Some of the best stuff I've gotten from other people has been stuff that I've gotten on a totally "I'll do it when I get to it" basis. You get a lot of un-accounted-for work in those cases, because people aren't "really" working for you, but are thinking about your problem in the shower, or procrastinating from their "real" work by reading Google Scholar entries related to your problem, etc. Eventually, you might get back something pretty good. (Not always, of course; so you could also say it has a higher variance.)
Indeed, it seems the most likely effect is that, if the data-collection becomes easy, they'll outsource the data-analysis to someone else. It won't be empowering people to make decisions about their own lives with more information than they had before. Rather, it'll just strengthen the tendency many people already have to abdicate responsibility for their own lives, and expect someone else to tell them what they should do. In this glorious future, they can collect a bunch of data about all aspects of their life, and someone will tell them what they're doing right/wrong, and what they should change.
seems like it'll enable more cargo-cult stuff
on
The Data-Driven Life
·
· Score: 1
One of the natural brakes on ridiculous cargo-cult self-help, diet, motivation, and other such fads is that nobody bothers to follow them too religiously. Now it'll be easier than ever to actually know for sure if you're following the latest pseudoscientific fad, because you'll have the data right there to validate! Hey, my friend John told me you should make sure your Baz data reading always stays under 7.2, except in the evenings it's okay if it goes up!
To be somewhat more accurate, it's not "now" called hypertext: it was called hypertext before gopher even existed. Gopher was first released in 1991, while Ted Nelson coined "hypertext" in 1965, and there were dozens of implementations before the WWW (the most popular outside academia was probably Apple's HyperCard, released in 1987).
Indeed, and we have a lot of windy land. It's actually not entirely clear to me why, economically, anyone would build offshore wind farms when there's so much onshore capacity still untapped. Surely it's a lot more expensive to put these turbines in the water 15 miles offshore, and the maintenance will be worse? Is it due to the way subsidies cut different ways? Or due to transmission-line issues?
The law seems intended to keep officers from using that sort of discretion, though. It specifically makes it a misdemeanor, punishable by up to six months in jail and a fine of up to $2,500, for an immigrant to fail to carry ID, even if they are legally in the country. And being caught a 2nd time without ID on you, again even if you really do have it at home, would be a felony.
California used to have a stop-and-identify statute that required you to show a cop "credible and reliable" identification upon request, but it was struck down in 1983.
Of course, if they're wrong they're going to be very unhappy to find that a driver's license isn't even close to documentation of citizenship.
While I don't like the law at all, to be fair, it does appear to explicitly say that a currently valid AZ driver's license is to be taken as sufficient evidence of citizenship for the purposes of these kinds of stops.
That does seem to leave a problem with out-of-state driver's licenses. I've driven through Arizona many times, with my only identification being a TX driver's license (I don't generally carry my passport or birth certificate when driving around the U.S.). If I'm stopped on suspicion of being a non-citizen, will my TX license be taken as sufficient? The law doesn't appear to guarantee it would be, so there's the possibility that, if I'm detailed while driving through Arizona on I-10, I could be held by the police for a day or two while I get someone to overnight-mail my passport. Makes me somewhat wary of driving through Arizona.
If the problem is statist bastards asking for ID, why shouldn't I blame the statist bastards asking for ID? Big businesses employing illegals aren't the ones stopping me on the street and violating my rights as a U.S. citizen by demanding my papers. The people doing that are the people supporting these kinds of anti-American laws.
The case you link explicitly notes that the person in question was not required to carry or produce papers, only to identify himself verbally. The opinion makes pretty clear that they were upholding Nevada's "stop-and-identify" statute on the understanding that the "identify" part included no requirement to produce ID:
As we understand it, the statute does not require a suspect to give the officer a driver's license or any other document. Provided that the suspect either states his name or communicates it to the officer by other means--a choice, we assume, that the suspect may make--the statute is satisfied and no violation occurs.
I generally don't, if I'm not driving or going to a bar/restaurant. If you're out jogging in your neighborhood, you really carry your driver's license with you?
Heck, when I was in college, under 21 and with no car, I probably went months at a time without taking my driver's license out of my drawer, because there was no reason to keep it in my wallet.
Yes, but in Canada at least, you can't be arrested and charged with a criminal offense for forgetting your papers. If you say, "my passport is back in my hotel", they aren't going to charge you with a crime for leaving your passport in your hotel. They will, instead, as you say, just let you explain where your papers are located.
That depends on the state. Several states require that government-issued photo ID be on your person at all times.
I'm not aware of any such state. I'm not even aware of any state that requires a U.S. citizen to have a government-issued photo ID. If you don't drive, you aren't required to get a driver's license, and if you don't exit the country, you aren't required to get a passport.
Slashdot Mirror
I don't believe R can actually handle 64-bit integers, either, at least not in any easy way. Its 'numeric' type is internally a double, and even when you coerce it to an integer, it just becomes in effect the 53-bit integer part of a double.
Depends on the kind of antitrust suit. Certainly you can't be guilty of monopoly leveraging unless you have a monopoly, so you're right as far as that goes. =] And it's also true that that's generally the easiest kind of antitrust claim to prove (it's what Microsoft was accused of), because there's a very strong presumption that if: 1. you're a monopoly; and 2. you're leveraging it; then that's bad, regardless of what "legitimate reasons" you have for doing so. (It used to even be considered always illegal to leverage a monopoly, but recent Supreme Court decisions have chipped away at that.)
There are other kinds of anticompetitive trade practices, though, with lower threshholds. At the one end of the scale is being a monopoly, where your conduct is closely scrutinized; at the other end is being a tiny player, whose market power is so small that any claim you were engaged in anticompetitive behavior is implausible. But in between there's a whole area of restraint-of-trade and anticompetitive tying, where you can be guilty of an antitrust violation without being a monopoly, if you have sufficient market power to influence another market improperly, and you did in fact do so. For example, some car manufacturers lost antitrust cases relating to car parts, despite not having a monopoly on automobiles: courts found that e.g. Mercedes banning anyone from making Mercedes-compatible parts was anticompetitive tying between the car market and the replacement-parts market, even though Mercedes isn't a car monopoly.
Those kinds of cases often come down to what the intent of the tying was. Mercedes argued that they wanted to control the replacement-parts market not for anticompetitive reasons, but for quality-assurance reasons; the courts ended up not buying that. But often courts do buy arguments that there was a legitimate reason for tying (possibly even most of the time; non-monopoly tying prosecutions succeeding isn't that common). In Apple's case, they would presumably argue that the purpose of requiring XCode and certain languages isn't mainly to restrain trade in dev-tools markets or to make it harder to port apps, but because of reasons related to its app-review process ("it's easier to review apps if they're all in Obj-C on XCode" or something). I could see a court giving reasonable deference to that, though predictions are always dodgy. The main place I could see that failing is if there's some smoking-gun email that ends up as evidence, where a VP or someone says, "hey we should institute this new requirement because of [reasons that sound like restraint of trade]".
This falls into the "not all" category though, I think, since it's an investigation of whether Apple's engaged in anticompetitive tying. The iPhone / Apple devtools tying is somewhat reminiscent of the car-makers / car-parts antitrust investigations, which happened even though no carmaker actually held a monopoly in the car market.
I think it'd be a problem for Apple if it were possible to show (e.g. by discovering an unwise internal memo) that the main purpose of the restrictions is to restrain trade, e.g. to harm the market for competitor development tools, or to make it harder to port apps to/from Android. Apple would want to be able to show that the tying was done for legitimate reasons not related to restraining trade.
I'd consider the legitimate scope of that law to be investigating whether they used the grant funds to by themselves a private yacht, and similar cases of fraudulent misuse of grant funds--- not the state investigating the scientific merits of the work.
For the federal codes, you can get full-text-searchable ASCII directly from the source.
That's basically what the United States Code is. Laws are compiled and sectioned into logical areas, so e.g. a bill that relates to both bankruptcy law and mining regulation (many laws have absurd things packaged together), will be refactored so the bankruptcy parts get codified into Title 11, while the mining regulations go into Title 30.
There are some decent points there, but a lot of them aren't really related to IEEE 754 compatibility. For example, bullet point #5 on their first-page list of five "gratuitous mistakes" is that Java doesn't support operator overloading. But by that standard, C sucks too, and yet is somehow used in lots of floating-point libraries.
I think the argument is more that, in practice, 32 bits is a decent sweet spot for the changeover from native ints to arbitrary-precision bigints (whereas 2 bits is not a sweet spot). Are there that many cases where someone needs integers between 32 and 64 bits, but doesn't need to account for the possibility of >64-bit integers, and therefore doesn't need to use routines that can handle arbitrary-size integers? If there were a lot of such cases, changing from 32- to 64-bit as the switchover point from native to bigint arithmetic would be justified. But the guess is that those cases aren't common, at least among matlab users.
Probably because it is politicizing science regardless of the merits. The way science operates is not generally by having attorneys general investigating the merits of scientific papers. If something was wrong or fraudulent, that's a job for journal editorial staff and university misconduct boards to sort out.
Similarly, it'd be correctly considered "politicizing science" if democrats launched a fraud investigation of a libertarian economist, regardless of whether that economist did or didn't fabricate evidence. The attorney general is just not the right person to do it.
That seems vaguely reasonable to me, based on my experience getting people to do things. Some of the best stuff I've gotten from other people has been stuff that I've gotten on a totally "I'll do it when I get to it" basis. You get a lot of un-accounted-for work in those cases, because people aren't "really" working for you, but are thinking about your problem in the shower, or procrastinating from their "real" work by reading Google Scholar entries related to your problem, etc. Eventually, you might get back something pretty good. (Not always, of course; so you could also say it has a higher variance.)
Indeed, it seems the most likely effect is that, if the data-collection becomes easy, they'll outsource the data-analysis to someone else. It won't be empowering people to make decisions about their own lives with more information than they had before. Rather, it'll just strengthen the tendency many people already have to abdicate responsibility for their own lives, and expect someone else to tell them what they should do. In this glorious future, they can collect a bunch of data about all aspects of their life, and someone will tell them what they're doing right/wrong, and what they should change.
One of the natural brakes on ridiculous cargo-cult self-help, diet, motivation, and other such fads is that nobody bothers to follow them too religiously. Now it'll be easier than ever to actually know for sure if you're following the latest pseudoscientific fad, because you'll have the data right there to validate! Hey, my friend John told me you should make sure your Baz data reading always stays under 7.2, except in the evenings it's okay if it goes up!
To be somewhat more accurate, it's not "now" called hypertext: it was called hypertext before gopher even existed. Gopher was first released in 1991, while Ted Nelson coined "hypertext" in 1965, and there were dozens of implementations before the WWW (the most popular outside academia was probably Apple's HyperCard, released in 1987).
You mean gibibyte, right?
*struggles to avoid giggling at the term*
I suppose playing along with Halo music is one way of doing it. But how about an FPS where you shoot up music venues?
Indeed, and we have a lot of windy land. It's actually not entirely clear to me why, economically, anyone would build offshore wind farms when there's so much onshore capacity still untapped. Surely it's a lot more expensive to put these turbines in the water 15 miles offshore, and the maintenance will be worse? Is it due to the way subsidies cut different ways? Or due to transmission-line issues?
also, store.apple.com
The law seems intended to keep officers from using that sort of discretion, though. It specifically makes it a misdemeanor, punishable by up to six months in jail and a fine of up to $2,500, for an immigrant to fail to carry ID, even if they are legally in the country. And being caught a 2nd time without ID on you, again even if you really do have it at home, would be a felony.
California used to have a stop-and-identify statute that required you to show a cop "credible and reliable" identification upon request, but it was struck down in 1983.
While I don't like the law at all, to be fair, it does appear to explicitly say that a currently valid AZ driver's license is to be taken as sufficient evidence of citizenship for the purposes of these kinds of stops.
That does seem to leave a problem with out-of-state driver's licenses. I've driven through Arizona many times, with my only identification being a TX driver's license (I don't generally carry my passport or birth certificate when driving around the U.S.). If I'm stopped on suspicion of being a non-citizen, will my TX license be taken as sufficient? The law doesn't appear to guarantee it would be, so there's the possibility that, if I'm detailed while driving through Arizona on I-10, I could be held by the police for a day or two while I get someone to overnight-mail my passport. Makes me somewhat wary of driving through Arizona.
If the problem is statist bastards asking for ID, why shouldn't I blame the statist bastards asking for ID? Big businesses employing illegals aren't the ones stopping me on the street and violating my rights as a U.S. citizen by demanding my papers. The people doing that are the people supporting these kinds of anti-American laws.
The case you link explicitly notes that the person in question was not required to carry or produce papers, only to identify himself verbally. The opinion makes pretty clear that they were upholding Nevada's "stop-and-identify" statute on the understanding that the "identify" part included no requirement to produce ID:
I generally don't, if I'm not driving or going to a bar/restaurant. If you're out jogging in your neighborhood, you really carry your driver's license with you?
Heck, when I was in college, under 21 and with no car, I probably went months at a time without taking my driver's license out of my drawer, because there was no reason to keep it in my wallet.
Yes, but in Canada at least, you can't be arrested and charged with a criminal offense for forgetting your papers. If you say, "my passport is back in my hotel", they aren't going to charge you with a crime for leaving your passport in your hotel. They will, instead, as you say, just let you explain where your papers are located.
I'm not aware of any such state. I'm not even aware of any state that requires a U.S. citizen to have a government-issued photo ID. If you don't drive, you aren't required to get a driver's license, and if you don't exit the country, you aren't required to get a passport.