Which is why outsourcing generally sucks. Its not a question of "outsourcing is cheaper" in general, its a question of "we want cheaper labor" and outsourcing to a cheap firm is a way to do that.
But you still get what you pay for. If you outsource to a firm that has competent employees you'll generally get a reasonable product (or service) from them -- but you'll be paying similar rates to the staff you replaced. Maybe more since the outsourcing firm will want its cut on top of their worker's salaries.
"One" is not the same as "the government." Governments, not just in Germany but everywhere and pretty much by definition, have many powers that are not entrusted to individuals and even a few that we haven't yet entrusted to corporations.
If you personally went around fingerprinting 6 year olds, you would probably find yourself in a fair bit of trouble regardless of the powers the government gives to itself.
should a court order requested by the authorities give them the permission to do so in order to fight crime.
That part might is definitely important from a legal perspective, but technologically its still a clusterfuck. If the encryption can be broken at all, for any reason, then it can potentially be broken by hackers. And so far, history indicates that "potentially broken by hackers" is only a handful of days away from "has been broken by hackers."
You fail to acknowledge that fact that the majority of the terrorist attacks post-911 have been by home-grown Muslims. There's no "importing" required.
And worse, it becomes a bit of a death spiral since the more we persecute the Muslims in our community, the higher chance that at least a few of them are going to be thinking about getting involved with more accepting communities. Occasionally that new community will be a terrorist organization, and only a small number of those otherwise-displaced people need to be indoctrinated in order to start the cycle over. Just one person can be sufficient to instigate a terror attack that gets reported around the world -- driving a truck down a busy pedestrian area doesn't really need a lot of co-conspirators or detailed organizational skills. It just needs a single person with access to a truck and a willingness to die for their cause.
If you just stuck with #2 and #3, you'd have pretty much hit the nail on the head though.
Its pretty politically incorrect to abolish basic freedoms as well. But I mean, as long as you can wrap it in a coating of "AMG terrorists!" we seem pretty willing to get those in power do whatever they want. This is not a new phenomena, and its hardly limited to Germany. The US and UK have been (publicly and proudly) doing shit like this for years now, and many other countries that we hear less about aren't exactly far behind.
Not to mention there's no shortage of right wing fervor in basically all Western countries right now -- Allies and Axis alike, so blaming the trend in Germany specifically on the Nazis kind of ignores the rest of the world in a way that makes little sense.
If anything, fear of their Nazi history being brought up has slowed down Germany's adoption of these shiny new oppression techniques that other Western governments have been salivating over for the past 10-15 years. But slowed is not stopped and until the world as a whole decides that acting out of terror (ie: exactly what the terrorists want) isn't really solving the problem, we're just going to see more and more countries slowly joining the parade into nationalistic fear of "the other."
Again, I'm not disagreeing that Google isn't really a "full" data set, but its a very large data set and any biases (aside from Google's possible hidden data massaging) should be fairly easy to identify and adjusted for -- or at least no harder than doing so for any other data set.
So what I'm asking, as a not-statistician, is what I'm missing that would allow you to generate a better set of data either with no biases (which I would find hard to believe,) or with some guarantee that all biases can be known and accounted for in a way that's impossible to do on the Google data set.
I'm probably not going to go out and buy a book to justify a Slashdot post, but giving me a TL;DR of why I'm wrong (and more importantly, why the paper's author -- who presumably is also an actual statistician -- is wrong) rather than just arbitrarily claiming so based on "a few minutes of thought" would be helpful.
Sampling theory is also the foundation of inferential statistics
Sure. But my question is still why you would want to use statistics and extrapolate to a full data set when you have a full data set to start with.
If you want to know how many people ate eggs for breakfast today, you have two choices: Ask every single person, or ask a sample of people and come up with a statistical distribution. The former will give you an exact number, but is impractical for a large population (say, everyone in the US.) The latter is more practical sure, but it only gives you an approximation with an error margin. Yes statistics do a very good job of filling in when we can't directly sample every individual, but by definition they can never be as accurate as actually being able to sample every individual.
That said, I wasn't intending to imply Google's data set is perfect, or even complete (and I specifically stated a couple of ways where its obviously not going to be.)
My question was more along the lines of asking what you would consider to be a better data set. You made a whole bunch of statements claiming that using Google is bad almost to the point of useless, but without any indication as to why you believe that to be the case, or what data set you think they should be using instead.
As for some of your specifics,
why terms are entered
Presumably, because people want an answer to their query. I suppose you might want to validate that assumption just in case there's armies of people out there randomly entering queries regarding things they don't care about to skew the data significantly, but I'm feeling fairly confident in my assumption that the vast majority of searches are done by people looking for actual search results.
the terms thought but not entered
True, but I'd put the onus on you to try and prove that this wouldn't apply to any other method of data collection as well. Until we invent mind reading, there will always be things people think but don't say, in every context.
he terms modified by google's previews of popular search terms
Now we're getting into things that might legitimately skew the results in at least moderate ways. I would hesitate again to suggest this is unique to Google (how many surveys exist where the wording of the questions ends up being deemed as self-biasing?) That said though, the fact that someone selected the completion from the drop down still shows interest in the term, and (again presumably) the drop down was at least fairly closely related to their original intended search or it wouldn't be a useful feature of the search box in the first place.
the indication that a user entered them rather than a machine
That one's entirely on Google to try and detect and merge or otherwise remove automated searches from their statistics. I have no idea if they even attempt to do so, though I imagine the probably do since advertisers don't want to be paying for robot eyes.
This kind of gets into the issue of Google massaging the data in unknown ways as I mentioned, and that's definitely a valid concern when using Google's data as the basis for your claims.
And of course you didn't mention a few other potential issues. Language for one. I bet you get a lot more Spanish searches in SoCal than you do in Colorado for example. TFA doesn't say whether that kind of thing was accounted for (though hopefully the actual paper would have at least mentioned it.) Of course again, to suggest that this makes Google worse than other methods you would have to show how those other methods avoid the language issue (and there has been no shortage of complaints about much psychology research -- especially older stuff -- being biased toward white, college-age Americans due to them being a ready sample for many researchers to use.)
this method of data collection ignores the need for sampling
Sampling of what? We take samples in the real world when we don't have the ability to collect full data sets, and then we use statistics to extrapolate those samples out. If we have the full data set to start with, what's the point in sampling and having to revert to statistical methods to back-fill the stuff that we just finished intentionally removing?
a biased sample
Biased in what way? Google's database gives not one whit about your race, sex, gender identity, political or religious beliefs, or anything else. Its about the most unbiased sample you can get (but not quite -- there's a bias against the underprivileged since many of them don't have computer access and thus won't be using Google or any other search engine.) There's also an obvious bias against Bing users, but nobody cares about them anyway:P.
data that is never entered into a google search
Examples? I can't think of anything a person would be interested in that they don't have or want to search for more information about.. and Google is, by a very very large margin, the primary place people go to do that. They're not considered a monopoly due to their fancy top hats.
show the behavior of whatever subset of the area targeted participates
Which is fine when that subset is 80% of the population, with no reason to believe that the other 20% of the population that uses Bing or other search engines would have radically different queries (again, with the one bias being against the underprivileged who don't use any search engine.)
limited use in larger inference.
Larger than what? That's literally the largest data set (about people's general interests) that we've ever collected. I can't see how using a smaller data set would net you better results.
The only real argument you could make is maybe questioning whether Google is massaging the data in a way that would make it unsuitable to this sort of analysis (and then the next question is how would we know that unless they tell us themselves?) And of course if you try to apply the data to an inappropriate subject, your results will likely be garbage, but that goes for any data set. Trying to claim that its not a sufficiently large data set seems a little bit silly though.
Yep. The whole "checks and balances" plan kind of falls flat when Congress is more interested in covering the president's ass than being an independent branch as the constitution intended.
I've seen the danger of an overly powerful central government... The cities are handing over their authority to a central government...
I've seen the danger of an overly powerful central government... The neighborhoods are handing over their authority to a central government...
I'm not necessarily trying to disagree with this logic, but I've always been curious why Americans think that the state level should be the ones with the power? Why is the fed worse? And if we agree that the fed is worse, then why is the city level not better? What's special about the state level, beyond purely "because America did it that way so 'Murca!"?
No, I'm saying it should be smart enough to make that determination on its own.. Sure it won't be perfect, but it will likely be a better judge of whether or not 1GiB is sufficient than pretty much 100% of computer-illiterate people. And someone computer literate enough to make the choice themselves is also hopefully smart enough to a) understand the implications of their choice and b) be able to figure out (or at least look up) the extra 1 or 2 buttons they need to click to find the "hidden" option.
If you're a developer of any sort then you should damned well know the difference between 32 and 64 bit systems, and you should be smart enough to find the damned option even if its hidden. You know, by maybe using the web to search if nothing else?
I'm more aiming this at people who don't know the difference, or don't think it matters, or they read something in 2002 saying how 64 bit systems aren't well supported and haven't updated their knowledge since then and if they see an option they'll pick the wrong one based on old or incorrect knowledge without any research.
Basically, if you know what you're doing, then my argument doesn't apply to you. And if you don't know what you're doing, then giving you an option is just going to make you do the wrong thing 50% of the time.
Of course if you're one of the people who just believes every person on the planet should have intimate and complete knowledge of computers in order to use them well.. I guess enjoy your fantasy world. Here in reality that's just not the way things work whether you like it or not, so we may as well at least stop people from shooting themselves in the foot unnecessarily when we can.
You would think they could somehow detect whether your CPU supports 64bit instructions and if so, just force the 64bit install. That is, don't give users the choice (or at least make it well-hidden.)
They don't give users a choice for so many other things these days.. why do they still let you fuck yourself on that particular option? There is literally zero reason to allow it -- the extremely rare case where someone really would want to put a 32bit OS on a 64bit PC, the person in question almost certainly knows the difference and why they have to do that. And if they don't, they're probably just wrong and making the choice based on incorrect or outdated information.
Similarly, why don't they include things like the C++ redistributables and.NET libs? They usually include the newest one but pretty hit and miss on the older versions. Its no end of hassle to developers (and to end users when the devs don't bother or get it wrong) and again, the only people who would really care about the extra couple hundred MB of disk space would probably know how to remove the ones they don't want anyway so what's the downside?
This is the aspect I'm mostly referring to when I say "backed" by a government -- its controlled.
Its backing is production as a function of productive labor
"Mining" bitcoins is labor (whether you call it productive or not is a subjective measure.) In particular, it requires an investment in time and equipment. So from that perspective, its very much parallel to fiat currencies even if there isn't very much physically heavy lifting involved.
It has this representation because it is sufficiently-universal
Which was my second point, and the reason why I stated that cryptocurrencies need some form of control (whether governmental or other.) If everyone just makes up their own cryptocurrency, then none of them will become sufficiently universal to be useful.
Think of bitcoin like ICQ. Started off small with a few users but as more and more people joined the system, it became popular and widely used. But of course once that happened, other parties decided that they wanted a piece of that puzzle and soon you had a dozen instant messengers, at least 4 of which were popular in different regions of the world (ICQ, MSN, Yahoo, AIM are the ones I know of.. there was probably more.)
We're already starting to see the tip of this with cryptocurrencies -- people who want to try and duplicate bitcoin's success so that they can be the person who owns the first 100 Altracoins and can just sit back and watch the money roll in.
Now if we fast forward the IM market to today: All four of the original "big name" ones are basically defunct. We've replaced them with voice apps that happen to have an IM system tacked onto them (often poorly,) or they're built into websites like Facebook and we've effectively got dozens of completely incompatible IMs and everyone has 2 or 3 that they're stuck using in order to stay in touch with different people. Its a bloody disaster of a mess.
And assuming nobody steps up to somehow block such a future, I fully see cryptocurrencies going the same route. With potentially even more disastrous results since its not just a question of having to have 4 different (but fairly equivalent) programs sucking up your computer resources, but it could potentially affect your very livelihood and savings if you put your money into MyCoin or you leave it in FaceCoin for too long and now SnapCoin has all the market share and your FaceCoins are continually devaluing, or whatever parallel you want to draw.
they don't have the purchasing flexibility of currency
But they can be traded for currency, which solves that problem. In fact, bitcoin is far more flexible than gold by your definition: I can go to Newegg's website and buy a video card with bitcoin. If I take a chunk of gold into Best Buy though, I'm pretty sure they won't accept my payment.
bitcoins is scarce; cryptocurrencies aren't scarce
This here is the problem. Real currencies work because they're limited. I can't just go write "100 Slashdots" on a piece of paper and expect you to accept it as currency, even though its physically not much different from the piece of paper you're holding that says "100 US Dollars." The difference is in perception -- real currency is backed by a government somewhere. If I hand you a piece of paper without some form of national backing, you will reject it as being anything more than a piece of used paper.
At some point, cryptocurrencies are going to need that sort of authority-based backing (it doesn't necessarily have to be a nation or state, but some entity with the power and will to take responsibility for backing the currency) or everyone and their monkey is going to create their own cryptocurrency and nobody will know (or care) what any of them are worth, which will make all of them less than worthless since people will have no basis to trust any particular cryptocurrency will still be viable tomorrow or if it will be replaced by the next fad.
Humans have been happy to pay for perceived rather than intrinsic value since at least the first bank note was introduced a few hundred years ago.. and probably before that as well.
"Worth" is exactly what someone would pay for it, regardless of whether they base their decision on perceived or intrinsic value. In the case of bitcoin, that's obviously greater than zero.
If the common rhetoric is true -- that terrorists just hate our freedom -- then the most obvious way to stop them is to just have no freedom! Win-win, right!
The downside for the EU is a hell of a lot of pissed off EU residents when they're suddenly no longer able to use Google's search because Google improperly promoted their shopping service (they have one of those? Huh.. so they do) half a decade ago. Though I guess it might still be ongoing.. TFA wasn't clear on that point.
Unfortunately, these days the request would probably be worded more along the lines of "We got a bomb threat, we want the records of all phone calls made in the past 24hr."
Luckily, "anonymous" tipping has ramifications outside of the 911 service office. For example, if you were an anonymous caller and the police divulged your name in court, they'd be in a lot of hot water themselves.
I certainly wouldn't use my own home or cell phone to implicate a politician or other person who might be able to gain access to the 911 call records on their own authority (or via coercion,) but if you're making an anonymous tip about a more common crime like domestic abuse or whatever, you're generally still pretty safe even if the 911 operator could theoretically identify you since they wouldn't be legally allowed to release that information.
I'm not sure how this follows. If all of their lines are registered as "Comcast," then what do I care which particular one they use to call me with? Sure if I'm told the actual number then I'd in principle be able to tell when two Comcast techs call me using different lines but um.. so what? I still don't see how that would change anything significant.
What unmasking all callers would do is prevent people from hiding their numbers, which has certain privacy implications for individuals in addition to making it harder for autodialers and spam callers and the similar to get away with their annoyances.
Slashdot Mirror
Or they become a community that supports your existing Muslim population and a few less start looking toward ISIS and friends for a place to belong.
Fear and hatred have never done much to stop violence. Occasionally though, an olive branch can.
Which is why outsourcing generally sucks. Its not a question of "outsourcing is cheaper" in general, its a question of "we want cheaper labor" and outsourcing to a cheap firm is a way to do that.
But you still get what you pay for. If you outsource to a firm that has competent employees you'll generally get a reasonable product (or service) from them -- but you'll be paying similar rates to the staff you replaced. Maybe more since the outsourcing firm will want its cut on top of their worker's salaries.
"One" is not the same as "the government." Governments, not just in Germany but everywhere and pretty much by definition, have many powers that are not entrusted to individuals and even a few that we haven't yet entrusted to corporations.
If you personally went around fingerprinting 6 year olds, you would probably find yourself in a fair bit of trouble regardless of the powers the government gives to itself.
should a court order requested by the authorities give them the permission to do so in order to fight crime.
That part might is definitely important from a legal perspective, but technologically its still a clusterfuck. If the encryption can be broken at all, for any reason, then it can potentially be broken by hackers. And so far, history indicates that "potentially broken by hackers" is only a handful of days away from "has been broken by hackers."
You fail to acknowledge that fact that the majority of the terrorist attacks post-911 have been by home-grown Muslims. There's no "importing" required.
And worse, it becomes a bit of a death spiral since the more we persecute the Muslims in our community, the higher chance that at least a few of them are going to be thinking about getting involved with more accepting communities. Occasionally that new community will be a terrorist organization, and only a small number of those otherwise-displaced people need to be indoctrinated in order to start the cycle over. Just one person can be sufficient to instigate a terror attack that gets reported around the world -- driving a truck down a busy pedestrian area doesn't really need a lot of co-conspirators or detailed organizational skills. It just needs a single person with access to a truck and a willingness to die for their cause.
If you just stuck with #2 and #3, you'd have pretty much hit the nail on the head though.
Its pretty politically incorrect to abolish basic freedoms as well. But I mean, as long as you can wrap it in a coating of "AMG terrorists!" we seem pretty willing to get those in power do whatever they want. This is not a new phenomena, and its hardly limited to Germany. The US and UK have been (publicly and proudly) doing shit like this for years now, and many other countries that we hear less about aren't exactly far behind.
73 years is hardly "overnight."
Not to mention there's no shortage of right wing fervor in basically all Western countries right now -- Allies and Axis alike, so blaming the trend in Germany specifically on the Nazis kind of ignores the rest of the world in a way that makes little sense.
If anything, fear of their Nazi history being brought up has slowed down Germany's adoption of these shiny new oppression techniques that other Western governments have been salivating over for the past 10-15 years. But slowed is not stopped and until the world as a whole decides that acting out of terror (ie: exactly what the terrorists want) isn't really solving the problem, we're just going to see more and more countries slowly joining the parade into nationalistic fear of "the other."
Again, I'm not disagreeing that Google isn't really a "full" data set, but its a very large data set and any biases (aside from Google's possible hidden data massaging) should be fairly easy to identify and adjusted for -- or at least no harder than doing so for any other data set.
So what I'm asking, as a not-statistician, is what I'm missing that would allow you to generate a better set of data either with no biases (which I would find hard to believe,) or with some guarantee that all biases can be known and accounted for in a way that's impossible to do on the Google data set.
I'm probably not going to go out and buy a book to justify a Slashdot post, but giving me a TL;DR of why I'm wrong (and more importantly, why the paper's author -- who presumably is also an actual statistician -- is wrong) rather than just arbitrarily claiming so based on "a few minutes of thought" would be helpful.
Sampling theory is also the foundation of inferential statistics
Sure. But my question is still why you would want to use statistics and extrapolate to a full data set when you have a full data set to start with.
If you want to know how many people ate eggs for breakfast today, you have two choices: Ask every single person, or ask a sample of people and come up with a statistical distribution. The former will give you an exact number, but is impractical for a large population (say, everyone in the US.) The latter is more practical sure, but it only gives you an approximation with an error margin. Yes statistics do a very good job of filling in when we can't directly sample every individual, but by definition they can never be as accurate as actually being able to sample every individual.
That said, I wasn't intending to imply Google's data set is perfect, or even complete (and I specifically stated a couple of ways where its obviously not going to be.)
My question was more along the lines of asking what you would consider to be a better data set. You made a whole bunch of statements claiming that using Google is bad almost to the point of useless, but without any indication as to why you believe that to be the case, or what data set you think they should be using instead.
As for some of your specifics,
why terms are entered
Presumably, because people want an answer to their query. I suppose you might want to validate that assumption just in case there's armies of people out there randomly entering queries regarding things they don't care about to skew the data significantly, but I'm feeling fairly confident in my assumption that the vast majority of searches are done by people looking for actual search results.
the terms thought but not entered
True, but I'd put the onus on you to try and prove that this wouldn't apply to any other method of data collection as well. Until we invent mind reading, there will always be things people think but don't say, in every context.
he terms modified by google's previews of popular search terms
Now we're getting into things that might legitimately skew the results in at least moderate ways. I would hesitate again to suggest this is unique to Google (how many surveys exist where the wording of the questions ends up being deemed as self-biasing?) That said though, the fact that someone selected the completion from the drop down still shows interest in the term, and (again presumably) the drop down was at least fairly closely related to their original intended search or it wouldn't be a useful feature of the search box in the first place.
the indication that a user entered them rather than a machine
That one's entirely on Google to try and detect and merge or otherwise remove automated searches from their statistics. I have no idea if they even attempt to do so, though I imagine the probably do since advertisers don't want to be paying for robot eyes.
This kind of gets into the issue of Google massaging the data in unknown ways as I mentioned, and that's definitely a valid concern when using Google's data as the basis for your claims.
And of course you didn't mention a few other potential issues. Language for one. I bet you get a lot more Spanish searches in SoCal than you do in Colorado for example. TFA doesn't say whether that kind of thing was accounted for (though hopefully the actual paper would have at least mentioned it.) Of course again, to suggest that this makes Google worse than other methods you would have to show how those other methods avoid the language issue (and there has been no shortage of complaints about much psychology research -- especially older stuff -- being biased toward white, college-age Americans due to them being a ready sample for many researchers to use.)
Umm say what?
this method of data collection ignores the need for sampling
Sampling of what? We take samples in the real world when we don't have the ability to collect full data sets, and then we use statistics to extrapolate those samples out. If we have the full data set to start with, what's the point in sampling and having to revert to statistical methods to back-fill the stuff that we just finished intentionally removing?
a biased sample
Biased in what way? Google's database gives not one whit about your race, sex, gender identity, political or religious beliefs, or anything else. Its about the most unbiased sample you can get (but not quite -- there's a bias against the underprivileged since many of them don't have computer access and thus won't be using Google or any other search engine.) There's also an obvious bias against Bing users, but nobody cares about them anyway :P.
data that is never entered into a google search
Examples? I can't think of anything a person would be interested in that they don't have or want to search for more information about.. and Google is, by a very very large margin, the primary place people go to do that. They're not considered a monopoly due to their fancy top hats.
show the behavior of whatever subset of the area targeted participates
Which is fine when that subset is 80% of the population, with no reason to believe that the other 20% of the population that uses Bing or other search engines would have radically different queries (again, with the one bias being against the underprivileged who don't use any search engine.)
limited use in larger inference.
Larger than what? That's literally the largest data set (about people's general interests) that we've ever collected. I can't see how using a smaller data set would net you better results.
The only real argument you could make is maybe questioning whether Google is massaging the data in a way that would make it unsuitable to this sort of analysis (and then the next question is how would we know that unless they tell us themselves?) And of course if you try to apply the data to an inappropriate subject, your results will likely be garbage, but that goes for any data set. Trying to claim that its not a sufficiently large data set seems a little bit silly though.
guessing the one plan they won't come up with? Offer good content at an affordable price, in a timely and accessible manner.
Yep. The whole "checks and balances" plan kind of falls flat when Congress is more interested in covering the president's ass than being an independent branch as the constitution intended.
I've seen the danger of an overly powerful central government ... The cities are handing over their authority to a central government ...
I've seen the danger of an overly powerful central government ... The neighborhoods are handing over their authority to a central government ...
I'm not necessarily trying to disagree with this logic, but I've always been curious why Americans think that the state level should be the ones with the power? Why is the fed worse? And if we agree that the fed is worse, then why is the city level not better? What's special about the state level, beyond purely "because America did it that way so 'Murca!"?
The usual argument.. "Bad government! Stop trying to protect the people we want to abuse because capitalism!"
I mean if the prisoners don't like it, they can just go to a competing jail right? That's how it works isn't it?
No, I'm saying it should be smart enough to make that determination on its own.. Sure it won't be perfect, but it will likely be a better judge of whether or not 1GiB is sufficient than pretty much 100% of computer-illiterate people. And someone computer literate enough to make the choice themselves is also hopefully smart enough to a) understand the implications of their choice and b) be able to figure out (or at least look up) the extra 1 or 2 buttons they need to click to find the "hidden" option.
If you're a developer of any sort then you should damned well know the difference between 32 and 64 bit systems, and you should be smart enough to find the damned option even if its hidden. You know, by maybe using the web to search if nothing else?
I'm more aiming this at people who don't know the difference, or don't think it matters, or they read something in 2002 saying how 64 bit systems aren't well supported and haven't updated their knowledge since then and if they see an option they'll pick the wrong one based on old or incorrect knowledge without any research.
Basically, if you know what you're doing, then my argument doesn't apply to you. And if you don't know what you're doing, then giving you an option is just going to make you do the wrong thing 50% of the time.
Of course if you're one of the people who just believes every person on the planet should have intimate and complete knowledge of computers in order to use them well.. I guess enjoy your fantasy world. Here in reality that's just not the way things work whether you like it or not, so we may as well at least stop people from shooting themselves in the foot unnecessarily when we can.
You would think they could somehow detect whether your CPU supports 64bit instructions and if so, just force the 64bit install. That is, don't give users the choice (or at least make it well-hidden.)
They don't give users a choice for so many other things these days.. why do they still let you fuck yourself on that particular option? There is literally zero reason to allow it -- the extremely rare case where someone really would want to put a 32bit OS on a 64bit PC, the person in question almost certainly knows the difference and why they have to do that. And if they don't, they're probably just wrong and making the choice based on incorrect or outdated information.
Similarly, why don't they include things like the C++ redistributables and .NET libs? They usually include the newest one but pretty hit and miss on the older versions. Its no end of hassle to developers (and to end users when the devs don't bother or get it wrong) and again, the only people who would really care about the extra couple hundred MB of disk space would probably know how to remove the ones they don't want anyway so what's the downside?
Its supply is controlled by a government
This is the aspect I'm mostly referring to when I say "backed" by a government -- its controlled.
Its backing is production as a function of productive labor
"Mining" bitcoins is labor (whether you call it productive or not is a subjective measure.) In particular, it requires an investment in time and equipment. So from that perspective, its very much parallel to fiat currencies even if there isn't very much physically heavy lifting involved.
It has this representation because it is sufficiently-universal
Which was my second point, and the reason why I stated that cryptocurrencies need some form of control (whether governmental or other.) If everyone just makes up their own cryptocurrency, then none of them will become sufficiently universal to be useful.
Think of bitcoin like ICQ. Started off small with a few users but as more and more people joined the system, it became popular and widely used. But of course once that happened, other parties decided that they wanted a piece of that puzzle and soon you had a dozen instant messengers, at least 4 of which were popular in different regions of the world (ICQ, MSN, Yahoo, AIM are the ones I know of.. there was probably more.)
We're already starting to see the tip of this with cryptocurrencies -- people who want to try and duplicate bitcoin's success so that they can be the person who owns the first 100 Altracoins and can just sit back and watch the money roll in.
Now if we fast forward the IM market to today: All four of the original "big name" ones are basically defunct. We've replaced them with voice apps that happen to have an IM system tacked onto them (often poorly,) or they're built into websites like Facebook and we've effectively got dozens of completely incompatible IMs and everyone has 2 or 3 that they're stuck using in order to stay in touch with different people. Its a bloody disaster of a mess.
And assuming nobody steps up to somehow block such a future, I fully see cryptocurrencies going the same route. With potentially even more disastrous results since its not just a question of having to have 4 different (but fairly equivalent) programs sucking up your computer resources, but it could potentially affect your very livelihood and savings if you put your money into MyCoin or you leave it in FaceCoin for too long and now SnapCoin has all the market share and your FaceCoins are continually devaluing, or whatever parallel you want to draw.
they don't have the purchasing flexibility of currency
But they can be traded for currency, which solves that problem. In fact, bitcoin is far more flexible than gold by your definition: I can go to Newegg's website and buy a video card with bitcoin. If I take a chunk of gold into Best Buy though, I'm pretty sure they won't accept my payment.
bitcoins is scarce; cryptocurrencies aren't scarce
This here is the problem. Real currencies work because they're limited. I can't just go write "100 Slashdots" on a piece of paper and expect you to accept it as currency, even though its physically not much different from the piece of paper you're holding that says "100 US Dollars." The difference is in perception -- real currency is backed by a government somewhere. If I hand you a piece of paper without some form of national backing, you will reject it as being anything more than a piece of used paper.
At some point, cryptocurrencies are going to need that sort of authority-based backing (it doesn't necessarily have to be a nation or state, but some entity with the power and will to take responsibility for backing the currency) or everyone and their monkey is going to create their own cryptocurrency and nobody will know (or care) what any of them are worth, which will make all of them less than worthless since people will have no basis to trust any particular cryptocurrency will still be viable tomorrow or if it will be replaced by the next fad.
Humans have been happy to pay for perceived rather than intrinsic value since at least the first bank note was introduced a few hundred years ago.. and probably before that as well.
"Worth" is exactly what someone would pay for it, regardless of whether they base their decision on perceived or intrinsic value. In the case of bitcoin, that's obviously greater than zero.
If the common rhetoric is true -- that terrorists just hate our freedom -- then the most obvious way to stop them is to just have no freedom! Win-win, right!
So you're a happy Bing user then?
The downside for the EU is a hell of a lot of pissed off EU residents when they're suddenly no longer able to use Google's search because Google improperly promoted their shopping service (they have one of those? Huh.. so they do) half a decade ago. Though I guess it might still be ongoing.. TFA wasn't clear on that point.
Agreed.
Unfortunately, these days the request would probably be worded more along the lines of "We got a bomb threat, we want the records of all phone calls made in the past 24hr."
Luckily, "anonymous" tipping has ramifications outside of the 911 service office. For example, if you were an anonymous caller and the police divulged your name in court, they'd be in a lot of hot water themselves.
I certainly wouldn't use my own home or cell phone to implicate a politician or other person who might be able to gain access to the 911 call records on their own authority (or via coercion,) but if you're making an anonymous tip about a more common crime like domestic abuse or whatever, you're generally still pretty safe even if the 911 operator could theoretically identify you since they wouldn't be legally allowed to release that information.
I'm not sure how this follows. If all of their lines are registered as "Comcast," then what do I care which particular one they use to call me with? Sure if I'm told the actual number then I'd in principle be able to tell when two Comcast techs call me using different lines but um.. so what? I still don't see how that would change anything significant.
What unmasking all callers would do is prevent people from hiding their numbers, which has certain privacy implications for individuals in addition to making it harder for autodialers and spam callers and the similar to get away with their annoyances.