"Based on a true story" means "based on the title of a book that you might recognize." If you don't know that, you should be kept in a home for the mentally insufficient, for your own safety.
'Based on' should be on Wikipedia's list of weasel words. True story:
A guy was bitten by a spider.
Based on this true story:
A guy was bitten by a spider and turned into spiderman!
You can recognise the true story from the film, but that's not why you watch the film...
Really? I remember mocking the one person I knew who got a Pentium 4 instead of an Athlon or a (by then, much cheaper, Pentium 3). We all knew that it could have 140 instructions in flight at a time, so needed heroic work from the branch predictor to have more than a vanishingly small chance of keeping it full and coming close to the theoretical throughput.
The thing that emerged later was the reason for the problem. It takes around 5 years to get a CPU to production, so you need to make a guess at what the available process technology in five years will be. For once, Intel got it wrong. They expected to be able to launch the P4 at 4GHz and scale it up to 10GHz within roughly the same thermal envelope. At that speed, even if you can only keep the pipeline half full on average, then it's still a very fast chip, and getting to 10GHz with a shorter pipeline is probably not feasible. At 2GHz, it's an embarrassment.
That isn't a Nobel Prize, it's the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel (which is, like the Nobels, selected by the Royal Swedish Academy of Sciences and announced at the same time as the Nobels, but it is not a Nobel prize).
Since it's copyrighted, but not patented (or patentable), you could redesign Linux or *BSD APIs that are similar in principle and operation as the POSIX APIs, but not a complete copy.
Which would have meant that porting code from UNIX to Linux/*NIX would have required shim layers. Worse, it would mean that every program that uses these APIs would be a derived work of UNIX and the UNIX license could prohibit the use of such shim layers.
Just because it affects your pet project does not make it legal.
It's not about my 'pet project', it's about the entire computing landscape. I'd actually be quite happy with a ruling in favour of Oracle: it would completely destroy a large segment of the US software industry and promote investment nearer me...
It wouldn't be the first time that lawyers from one company have been arguing opposite things at the same time in different cases. If the other side is alert, then they'll enter court transcripts from the other case into their testimony.
Yes and it's essential for a thriving industry. Imagine where Linux and *BSD would be, for example, if The Open Group had been able to enforce copyright on the POSIX APIs. There'd be no *NIX systems, no WINE, no drop-in replacements for any library when the author stops supporting it..
As someone commented, ask what happened to Physics
Physics is still on the gravy train created by the Manhattan Project. Each of the large magnets for the SSC (which was never finished) cost more than the US government has spent on computer science research in total, ever.
There are also some issues in the way science is done. It's increasingly difficult for a lone researcher to have significant impact, instead you need teams of people. This includes PhD students working on focussed parts of the problem and postdocs coordinating them and working on broader things, and generally one tenured faculty member (or senior postdoc) overseeing the entire thing. This means that you need more PhD students than postdocs, more postdocs than faculty. That's fine, except that there's the expectation that a lot of the PhD students will become postdocs and a lot of the postdocs will become faculty. That's where it starts to become unsustainable.
If, for $2m, you're promising a softcore that can run on an FPGA, then you're asking way more than the research is worth. Other groups have done this already with smaller research grants. If you're aiming to produce something that can be fabbed and run at a reasonable speed (even without the supporting software stack), then you're asking over an order of magnitude too little to be taken seriously. You'd blow that budget in the first prototyping run.
The 280km/h figure for the TGV is including stops - the peak cruising speed is a fair bit higher. Averaging 120km/h isn't bad, but it's still about what I'd expect from cheap '80s rolling stock.
The problem with helicopters is that maintaining your altitude in one consumes a lot of energy. Ideally you want something more like some of the gyrocopter designs that can take off like a helicopter but fly like an autogyro or a fixed-wing aircraft.
Don't make the mistake of thinking that just because we can make small drones we can make human-sized carriers equally easily. If physics worked that way, then we'd have flying suits based on flies or sparrows that could be powered purely by human muscles.
125km/h is much slower than the TGV. The British Intercity trains can run at 125 miles per hour, which is 60% faster than that. Even our crappy commuter trains have a cruising speed of around 120km/h (at least, that's what my phone's GPS told me last time I was on one). The TGV held the record for the fastest average speed for a conventionally scheduled train in 2007, at 280km/h (which would have managed the grandparent's journey in just under 2 hours).
Add to that, the low-density is mostly true when you get away from the ground. Aircraft density around airports is already sufficiently high that you need a very alert team of air traffic controllers to ensure that there aren't collisions and stories of near misses make the news every year. Flying cars would want to land in places that are convenient for people, so there would be similar density issues, no air traffic control (unless you require them to have an automated landing system that coordinates with other nearby vehicles) and far greater possibilities of killing people when they fall out of the sky (there's a reason why airports don't have buildings right next to the runways...).
I don't use any of his software, but I do pay attention to the politics surrounding it. The loudest and most obnoxious person in that entire community is Lennart. He has a history of making technically questionable decisions and then, when questioned, refusing to reply with technical justifications but instead complaining that everyone hates him and that some people just don't like progress.
Now, as a FreeBSD developer, I'm quite happy with his work: He's driven more people to adopt FreeBSD than anyone else that I'm aware of, but I do find it amusing when he complains that the problem with the community is other people.
It's trivial to add a keylogger to any OS. In X, you don't even need special privileges to log every event, but if you're the one shipping the kernel and display server then you already need to have a few places where every keystroke passes through a bit of your code for other reasons (e.g. trapping control-alt-delete / control-alt-backspace combinations).
purpose of testing is to collect data about the system itself and how it operates in end user environments; this is collecting information about the end users themselves rather than just the machine
How long does it take a user to find the correct button to dismiss a dialog? How many users use keyboard navigation rather than the mouse to navigate dialogs? How many times do the people who do use keyboard navigation hit tab without typing doing anything that would modify the field? All of these things require a keylogger (or a camera pointed at the screen) to find out and give valuable data when designing a UI. You'd hope that there's something client side that filters out anything that might be a password and aggregates some of the data, but it's all information about the end user that you need to collect to do a good job at UI design.
You can't collect it about all end users from production code without making the product unusable, so you put out testing releases that do collect it.
True, but that tends to be the best you can do in HCI testing. Users won't do the same things with a camera pointed at them as they'll do in private, but you hope that they'll do enough that's the same that you get useful results.
What are the cons for the consumer? As I understand it, Comcast and Time Warner cable are already in almost entirely geographically distinct regions, and neither has plans to enter the other's territory. At the moment, they get to pretend that their not a monopoly, because if you take the US as a whole there is competition, even though in any given area there isn't. Post merger, they will have bigger economies of scale and no way to hide from the fact that their a monopoly.
It's an early test program. The entire reason that it exists is to see how people use it, whether the UI decisions make sense, and what the designers overlooked. It is not intended for normal use and it is not intended for production environments.
I gave up part way through season 2. I think it's the only adaptation of a book where they've managed to tell the story more slowly than the novels. The casting was well done, but they really needed to work on the pace.
I enjoyed Roger Ramjet when I was 6. I saw them again about a year ago and was astonished at how much I'd missed as a child. Definitely designed for parents to watch with their children.
Mary Beard has written an excellent article about the pros and cons of standardised testing (unfortunately, I have the version in her book, so don't have a link to the original handy). The problem with standardised testing is that you need a mark scheme that can be applied evenly by a load of different examiners. This effectively limits you to 'this essay must make 5 of these points:' sorts of thing, rather than 'this essay must demonstrate knowledge of the subject and make coherent arguments'. For example, in a university essay I'd have no problem with making 'the student must tell me something that I didn't tell them first' a requirement for the top grade, because I expect good students to read additional material beyond the lecture notes. It's impossible to set that in a national exam, not least because the examiner has no idea what the teachers told the students in set of scripts that they're marking. This means that in a standardised exam, you can test whether the student can memorise things and, for science subjects, whether they can apply knowledge to problems with well-known solutions. You can't easily test creativity or reasoning ability.
There is a small voice of reason that keeps reminding me that I do value Democracy and the president SHOULD act on what the people want
There's a balance. If the elected representatives are going to vote exactly how the people feel right now then you may as well save the money, fire them all, and have direct democracy. The point of representative democracy is to elect people who have similar world views to you, who will vote in the same way that you would if you had time to study all of the issues.
Someone recently pointed out to me that you can buy 'vegetarian water', which has all microorganisms filtered out so that you don't accidentally consume them. My immediate reaction was to wonder if we can persuade the people who buy this to stop breathing, as air also contains microorganisms, many of which are members of the animal kingdom...
Slashdot Mirror
"Based on a true story" means "based on the title of a book that you might recognize." If you don't know that, you should be kept in a home for the mentally insufficient, for your own safety.
'Based on' should be on Wikipedia's list of weasel words. True story:
A guy was bitten by a spider.
Based on this true story:
A guy was bitten by a spider and turned into spiderman!
You can recognise the true story from the film, but that's not why you watch the film...
Really? I remember mocking the one person I knew who got a Pentium 4 instead of an Athlon or a (by then, much cheaper, Pentium 3). We all knew that it could have 140 instructions in flight at a time, so needed heroic work from the branch predictor to have more than a vanishingly small chance of keeping it full and coming close to the theoretical throughput.
The thing that emerged later was the reason for the problem. It takes around 5 years to get a CPU to production, so you need to make a guess at what the available process technology in five years will be. For once, Intel got it wrong. They expected to be able to launch the P4 at 4GHz and scale it up to 10GHz within roughly the same thermal envelope. At that speed, even if you can only keep the pipeline half full on average, then it's still a very fast chip, and getting to 10GHz with a shorter pipeline is probably not feasible. At 2GHz, it's an embarrassment.
That isn't a Nobel Prize, it's the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel (which is, like the Nobels, selected by the Royal Swedish Academy of Sciences and announced at the same time as the Nobels, but it is not a Nobel prize).
Since it's copyrighted, but not patented (or patentable), you could redesign Linux or *BSD APIs that are similar in principle and operation as the POSIX APIs, but not a complete copy.
Which would have meant that porting code from UNIX to Linux/*NIX would have required shim layers. Worse, it would mean that every program that uses these APIs would be a derived work of UNIX and the UNIX license could prohibit the use of such shim layers.
Just because it affects your pet project does not make it legal.
It's not about my 'pet project', it's about the entire computing landscape. I'd actually be quite happy with a ruling in favour of Oracle: it would completely destroy a large segment of the US software industry and promote investment nearer me...
It wouldn't be the first time that lawyers from one company have been arguing opposite things at the same time in different cases. If the other side is alert, then they'll enter court transcripts from the other case into their testimony.
Yes and it's essential for a thriving industry. Imagine where Linux and *BSD would be, for example, if The Open Group had been able to enforce copyright on the POSIX APIs. There'd be no *NIX systems, no WINE, no drop-in replacements for any library when the author stops supporting it..
As someone commented, ask what happened to Physics
Physics is still on the gravy train created by the Manhattan Project. Each of the large magnets for the SSC (which was never finished) cost more than the US government has spent on computer science research in total, ever.
There are also some issues in the way science is done. It's increasingly difficult for a lone researcher to have significant impact, instead you need teams of people. This includes PhD students working on focussed parts of the problem and postdocs coordinating them and working on broader things, and generally one tenured faculty member (or senior postdoc) overseeing the entire thing. This means that you need more PhD students than postdocs, more postdocs than faculty. That's fine, except that there's the expectation that a lot of the PhD students will become postdocs and a lot of the postdocs will become faculty. That's where it starts to become unsustainable.
If, for $2m, you're promising a softcore that can run on an FPGA, then you're asking way more than the research is worth. Other groups have done this already with smaller research grants. If you're aiming to produce something that can be fabbed and run at a reasonable speed (even without the supporting software stack), then you're asking over an order of magnitude too little to be taken seriously. You'd blow that budget in the first prototyping run.
The 280km/h figure for the TGV is including stops - the peak cruising speed is a fair bit higher. Averaging 120km/h isn't bad, but it's still about what I'd expect from cheap '80s rolling stock.
The problem with helicopters is that maintaining your altitude in one consumes a lot of energy. Ideally you want something more like some of the gyrocopter designs that can take off like a helicopter but fly like an autogyro or a fixed-wing aircraft.
Don't make the mistake of thinking that just because we can make small drones we can make human-sized carriers equally easily. If physics worked that way, then we'd have flying suits based on flies or sparrows that could be powered purely by human muscles.
125km/h is much slower than the TGV. The British Intercity trains can run at 125 miles per hour, which is 60% faster than that. Even our crappy commuter trains have a cruising speed of around 120km/h (at least, that's what my phone's GPS told me last time I was on one). The TGV held the record for the fastest average speed for a conventionally scheduled train in 2007, at 280km/h (which would have managed the grandparent's journey in just under 2 hours).
Add to that, the low-density is mostly true when you get away from the ground. Aircraft density around airports is already sufficiently high that you need a very alert team of air traffic controllers to ensure that there aren't collisions and stories of near misses make the news every year. Flying cars would want to land in places that are convenient for people, so there would be similar density issues, no air traffic control (unless you require them to have an automated landing system that coordinates with other nearby vehicles) and far greater possibilities of killing people when they fall out of the sky (there's a reason why airports don't have buildings right next to the runways...).
Because a regular investor is investing their money
Spoken like someone who has no idea how the banking system works.
I don't use any of his software, but I do pay attention to the politics surrounding it. The loudest and most obnoxious person in that entire community is Lennart. He has a history of making technically questionable decisions and then, when questioned, refusing to reply with technical justifications but instead complaining that everyone hates him and that some people just don't like progress.
Now, as a FreeBSD developer, I'm quite happy with his work: He's driven more people to adopt FreeBSD than anyone else that I'm aware of, but I do find it amusing when he complains that the problem with the community is other people.
It's trivial to add a keylogger to any OS. In X, you don't even need special privileges to log every event, but if you're the one shipping the kernel and display server then you already need to have a few places where every keystroke passes through a bit of your code for other reasons (e.g. trapping control-alt-delete / control-alt-backspace combinations).
purpose of testing is to collect data about the system itself and how it operates in end user environments; this is collecting information about the end users themselves rather than just the machine
How long does it take a user to find the correct button to dismiss a dialog? How many users use keyboard navigation rather than the mouse to navigate dialogs? How many times do the people who do use keyboard navigation hit tab without typing doing anything that would modify the field? All of these things require a keylogger (or a camera pointed at the screen) to find out and give valuable data when designing a UI. You'd hope that there's something client side that filters out anything that might be a password and aggregates some of the data, but it's all information about the end user that you need to collect to do a good job at UI design.
You can't collect it about all end users from production code without making the product unusable, so you put out testing releases that do collect it.
True, but that tends to be the best you can do in HCI testing. Users won't do the same things with a camera pointed at them as they'll do in private, but you hope that they'll do enough that's the same that you get useful results.
What are the cons for the consumer? As I understand it, Comcast and Time Warner cable are already in almost entirely geographically distinct regions, and neither has plans to enter the other's territory. At the moment, they get to pretend that their not a monopoly, because if you take the US as a whole there is competition, even though in any given area there isn't. Post merger, they will have bigger economies of scale and no way to hide from the fact that their a monopoly.
It's an early test program. The entire reason that it exists is to see how people use it, whether the UI decisions make sense, and what the designers overlooked. It is not intended for normal use and it is not intended for production environments.
I gave up part way through season 2. I think it's the only adaptation of a book where they've managed to tell the story more slowly than the novels. The casting was well done, but they really needed to work on the pace.
I enjoyed Roger Ramjet when I was 6. I saw them again about a year ago and was astonished at how much I'd missed as a child. Definitely designed for parents to watch with their children.
Mary Beard has written an excellent article about the pros and cons of standardised testing (unfortunately, I have the version in her book, so don't have a link to the original handy). The problem with standardised testing is that you need a mark scheme that can be applied evenly by a load of different examiners. This effectively limits you to 'this essay must make 5 of these points:' sorts of thing, rather than 'this essay must demonstrate knowledge of the subject and make coherent arguments'. For example, in a university essay I'd have no problem with making 'the student must tell me something that I didn't tell them first' a requirement for the top grade, because I expect good students to read additional material beyond the lecture notes. It's impossible to set that in a national exam, not least because the examiner has no idea what the teachers told the students in set of scripts that they're marking. This means that in a standardised exam, you can test whether the student can memorise things and, for science subjects, whether they can apply knowledge to problems with well-known solutions. You can't easily test creativity or reasoning ability.
There is a small voice of reason that keeps reminding me that I do value Democracy and the president SHOULD act on what the people want
There's a balance. If the elected representatives are going to vote exactly how the people feel right now then you may as well save the money, fire them all, and have direct democracy. The point of representative democracy is to elect people who have similar world views to you, who will vote in the same way that you would if you had time to study all of the issues.
Someone recently pointed out to me that you can buy 'vegetarian water', which has all microorganisms filtered out so that you don't accidentally consume them. My immediate reaction was to wonder if we can persuade the people who buy this to stop breathing, as air also contains microorganisms, many of which are members of the animal kingdom...