I never liked the Macs and their frilly user interface. Being a Unix geek, I just wanted a set of Unix-like (or better tools).
You know, I'm the same way. However, I recently bought a PowerMac, and it really is a wonderful machine. A lot of the standard UNIX apps are even better on OS X than on Linux. Emacs, in particular, is miles ahead, supporting an interface that actually blends in with the Aqua UI, and sports anti-aliased fonts and a Mac-style top menubar. The only caveat is that the default terminal app could be a little bit better.
Ack, you're right. The maximum velocities are 2.5% and 3.6% of the speed of light, respectively. Ironically, I'm an aerospace student, so math realy shouldn't scare me...
There is nothing unphysical about traveling between the planets of a solar system in days. The distance from the Sun to Pluto is about 0.22 light days. If your maximum acceleration and decceleration is 1g, you can travel to Pluto in roughly 18 days, with a maximum velocity of 2% of the speed of light. If you can tolerate acceleration at 2g, you can get there in roughly 12.5 days, with a maximum velocity of 1.8% the speed of light. All of these are entirely within the realm of possibility, not only within our current physics, but with, from a physics theory point of view, fairly reasonable technologies.
Re:look at it from a new perspective
on
The H-1B Swindle
·
· Score: 1
I'm sure youd expect that the lion's share goes to owners and managers, but do you have proof? For every high-profile company that can afford to just sit on a 20% cost savings, there are a dozen others doing battle with actual competitors that are forced to pass it on to the consumer.
It is not a philosophical one at all. Science is a method. It's the method you learned in school: propose a hypothesis, devise an experiment to gather evidence to determine if that hypothesis is true. Various proponents of intelligent design either do not follow that method at all, or do so shoddily.
The philosophical question of whether there is a creator is bigger than science. From a philosophical standpoint, the intelligent design people are no more right or wrong than the scientists. The problem with intelligent design is that its religion trying to ride the coat-tails of science. The success of science in advancing civilization has given it a credibility that religion no longer possesses. Intelligent design is just an example of religion trying to give itself the credible veneer of science without following its rigorous practices.
Philosophers also cannot assure me whether I really exist or not. I find the field as stimulating as the next person, but honestly, does anybody believe that philosophy has any bearing on the real world? It is a wonderfully pure field, but in that purity, necessarily makes itself useless for practical purposes.
And yet most Americans support some form of evolution as a valid explanation of how we got here.
Unfortunately, nearly half do not. You could get into subtlety about evolution being a theory, but I doubt most Americans have given it that much thought. They simply hear two different things, one from science and one from their church, and choose to trust the church instead. That disturbs me.
True. Interesting, though. Most people cannot actually explain evolutionary theory
And I find that rather dissapointing. Of course, I find the same thing to be true of religion as well. Catholics are always entertaining in this respect --- many seem to be quite fervent, yet don't realize that their religion is a complex and subtle one that requires some study to truely understand. Not being Catholic, but having studied Catholicism, I find that most Catholics are unable to explain to me why the believe what they do, or even the process through which the Church determines and enforces what they should believe.
People need to question everything, religious beliefs as well as scientific precepts.
Undoubtedly. However, let's speak at a practical level. I do not hold great hope that most people will one day become enlightened and learn to be rigorous about their beliefs. The sad truth is that most people will remain blind. Given that, I'd much rather they blindly follow science than blindly follow religion, for the sake of the continued prosperity of our society.
Both have been proven wrong in many instances.
Here's the way I see it. Science is couched in the provable. It does not claim to have absolute truth, just some useful likeness of it. Religion is couched in belief. Organized religion, anyway, does claim to have absolute truth. Science can be wrong and retain its credibility. Religion cannot.
This does not mean that all life is a product of evolution.
The fact that things evolve is not the scientific evidence that is used to support the theory of evolution. The facts that are used to support evolution, well they do imply that all life is a product of evolution.
As you may have guessed, I do believe in the bible and creation. That does not mean that I do not accept scientific fact.
How much do you believe the Bible? If you believe it in its entirety, then you cannot claim that you accept scientific fact, because the Bible contradicts many such scientific facts. If you only believe in the Bible partially, then you may claim to accept scientific facts, but then realize that the "pick and choose" varient of Christianity that seems so popular today is rather at odds with what the Bible says.
Ultimately, I don't really care what people choose to believe. Being an engineer, my only interest is the pragmatic. I don't oppose the efforts of the religious to teach creationism as science, etc, because I care what children believe --- I oppose it because it will lead to fewer biologists. The way I see it, Jesus doesn't pay the bills, science does. The only purpose of public education is to create a productive workforce. Teaching religion doesn't do that, teaching science does.
"Moral values are ailing in this country. We need to do more to pull people in to Christianity." Sound familiar?
Yeah sure, and aside from the "pull people to Christianity" bit, I have no problem with trying to address to moral values problem in the country (as long as its not the government trying to do it). However, unlike science, moral values are not the bedrock of our economy and prosperity as a nation...
I think that's an overly optimistic viewpoint. Science is ailing in this country. The high-profile crusades against it from the creationists is just the tip of the iceberg. Far more important is the fact that science just plain isn't held in high regard, at a cultural level, and not enough Americans are persuing careers in the various scientific fields. On top of that is all the snake-oil masquerading as science, and the fact that the general public really has no idea of what is and is not science. Of course, that is nothing new, but it is something that universal education was supposed to fix. Well, in that case, universal education has failed. It is not at all surprising to see why, though. In the vast majority of class rooms in the US, science is taught not as a set of principles and methods, but as a loosly-connected facts. Students are not taught how to think scientifically, but are mearly forced to learn tidbits of information that may as well have just been pitching statistics for all the good they do.
Um, that struct from the Solaris code isn't portable to an Alpha-based Cray (which are ILP64 machines). It also assumes alignment rules all over the place. On top of all that, it'd take me maybe a day to come up with a Lisp macro that let you specify a byte-accurate "data packet" format, without all the unportable assumptions regarding the size and alignment of ints and chars.
Actually, I'm expressing horror at the consequences of interpreting a non-machine-native IL. Adequate performance entails implementing a full virtual machine with runtime code generation.
Or a install-time compiler. Or a regular compiler with signed binaries.
It's not the programmer I'm worried about. It's the end user I'm worried about; at the end of the day, he's the one taking the risk of running code on his machine. This system design leaves him vulnerable to local DoS attacks.
How so? With the compile-at-install model, the only attack vector is defeating the protections in the trusted compiler. I don't think that's particularly easier that defeating the protections in a 'trusted' kernel.
Provide control over the runtime environment, with the explicit possibility of no runtime (which C does and Lisp does not)
Define "runtime". C++ has a runtime. C++ can and has been used for writing high-performance microkernels. Lisp can be used so it requires no bigger a runtime than C++. As long as you don't try to use EVAL in the kernel, you're set. All it requires is a GC, and I find your criteria for excluding the GC capricious and aribtrary --- one that only has merit in the specific case of RT kernels.
We are discusing high-level languages in general which perform field reordering optimizations;
If, as you just said, you "stick to machine words" what does field reordering have to do with anything?
As discussed, C forbids this, so it's already a little bit easier to talk to the hardware.
But it says nothing about alignment, so even though its more predictable, its still completely useless. To put things into concrete terms, you can't use a struct like this to talk to hardware:
struct foo {
int rega;
char regb;
int regc; };
Well, you can, but its completely unportable. Sure, at least you know that 'regc' is after 'regb', but you still don't know where 'regc' is.
Your grasp of statistics is marginal at best. The FBI doesn't do 250m investigations every year, so that success rate is hardly 99.xx%. The FBI had 34,451 criminal investigations last year. That's around a 0.6% failure rate. Good enough for government work, probably, but let's recall that Apple is dealing with a class-action lawsuit because the iPod Nano's acrylic has a failure rate of 0.1%...
The number of murders in the US last year was 16137. Out of 295,734,134 people, that's a rate of 0.0054%, a figure that is more or less negligable. There is absolutely no point in getting worked up over murders with those kind of numbers.
An agent can make an honest mistake in the performance of a job (just like we all can, right?)
I dunno. Ford's engineers make a mistake, people die, and the company is sued for tens of millions of dollars. Merck's chemists make a mistake, people die, and the company is sued for tens of millions of dollars. FBI's agents make a mistake, and what exactly happens?
The "we're all human and we make mistakes" line wrt. the present administration is really wearing thin. Man, I'd love to be in a line of work I could fuck up that much and still keep my job...
I realize the two situations are quantitatively (if not qualitatively) different, but I wonder if there were apologists like you for the KGB?
The "you never hear the good stuff that happens" line is completely useless, because its not verifiable. You just assume good stuff happens, but since you have no proof, you can just as easily assume that it doesn't. Which side you choose is arbitrary, but I'm inclined to choose the latter, since at least I have the evidence of history to suggest its correctness.
Of course it wouldn't do. The fact that the agency investigates the violations doesn't change the fact that the violations shouldn't be happening in the first place. Creating a problem, then trying to fix it isn't as good as not creating a problem in the first place.
The Athlon64's adds and muls are pipelined, so its still 1 per cycle. If you look at the pipeline of a GPU (dozens of stages), you'll probably find that add and mul takes even longer.
Right. If you're sticking to machine words, then who *cares* that C's rules for laying out structs is more predictable than what the Lisp optimizer might do to a struct? You can use a read/write oriented machine-word module in Lisp as easily as you can do the same in C!
And researchers who want to build a general-purpose operating system that can be used in all of these scenarios are what? Shit out of luck?
I think you've missed the point entirely. The whole point of this line of discussion was that such researchers were barking up the wrong tree. The microkernel itself becomes useless (and all the massive contortions required to support it), when the language itself is safe. The kernel ceases to exist as a seperate entity and becomes a seperate (though special, in that it can use unsafe operations) library. You stop worrying about protection domains, context switches, interprocedural communication, drivers in userspace vs kernelspace, how to implement paging, etc. Protection becomes fine-grained (per object), context switches become cheap, protection-domains become unnecessary, interprocedural communication becomes wicked fast (shared memory), and userspace vs kernelspace ceases to have any meaning. All you give up (in user code) is: pointers and manual memory management.
For hard-real time, CPU-intensive uses, the second concession might be a deal-breaker. In that case, the microkernel has some semblence of relevance. However, for the vast majority of systems which don't need hard-RT, the concessions are minimal compared to the gains.
Slashdot Mirror
Yes, it's called "every other institute in existence".
I never liked the Macs and their frilly user interface. Being a Unix geek, I just wanted a set of Unix-like (or better tools).
You know, I'm the same way. However, I recently bought a PowerMac, and it really is a wonderful machine. A lot of the standard UNIX apps are even better on OS X than on Linux. Emacs, in particular, is miles ahead, supporting an interface that actually blends in with the Aqua UI, and sports anti-aliased fonts and a Mac-style top menubar. The only caveat is that the default terminal app could be a little bit better.
Ack, you're right. The maximum velocities are 2.5% and 3.6% of the speed of light, respectively. Ironically, I'm an aerospace student, so math realy shouldn't scare me...
There is nothing unphysical about traveling between the planets of a solar system in days. The distance from the Sun to Pluto is about 0.22 light days. If your maximum acceleration and decceleration is 1g, you can travel to Pluto in roughly 18 days, with a maximum velocity of 2% of the speed of light. If you can tolerate acceleration at 2g, you can get there in roughly 12.5 days, with a maximum velocity of 1.8% the speed of light. All of these are entirely within the realm of possibility, not only within our current physics, but with, from a physics theory point of view, fairly reasonable technologies.
I'm sure youd expect that the lion's share goes to owners and managers, but do you have proof? For every high-profile company that can afford to just sit on a 20% cost savings, there are a dozen others doing battle with actual competitors that are forced to pass it on to the consumer.
Not for long. The next generation of PC tech will rely heavily on multicore CPUs as well.
It is not a philosophical one at all. Science is a method. It's the method you learned in school: propose a hypothesis, devise an experiment to gather evidence to determine if that hypothesis is true. Various proponents of intelligent design either do not follow that method at all, or do so shoddily.
The philosophical question of whether there is a creator is bigger than science. From a philosophical standpoint, the intelligent design people are no more right or wrong than the scientists. The problem with intelligent design is that its religion trying to ride the coat-tails of science. The success of science in advancing civilization has given it a credibility that religion no longer possesses. Intelligent design is just an example of religion trying to give itself the credible veneer of science without following its rigorous practices.
I dunno. At my brother's highschool, every other kid has an iPod.
Philosophers also cannot assure me whether I really exist or not. I find the field as stimulating as the next person, but honestly, does anybody believe that philosophy has any bearing on the real world? It is a wonderfully pure field, but in that purity, necessarily makes itself useless for practical purposes.
And yet most Americans support some form of evolution as a valid explanation of how we got here.
Unfortunately, nearly half do not. You could get into subtlety about evolution being a theory, but I doubt most Americans have given it that much thought. They simply hear two different things, one from science and one from their church, and choose to trust the church instead. That disturbs me.
True. Interesting, though. Most people cannot actually explain evolutionary theory
And I find that rather dissapointing. Of course, I find the same thing to be true of religion as well. Catholics are always entertaining in this respect --- many seem to be quite fervent, yet don't realize that their religion is a complex and subtle one that requires some study to truely understand. Not being Catholic, but having studied Catholicism, I find that most Catholics are unable to explain to me why the believe what they do, or even the process through which the Church determines and enforces what they should believe.
People need to question everything, religious beliefs as well as scientific precepts.
Undoubtedly. However, let's speak at a practical level. I do not hold great hope that most people will one day become enlightened and learn to be rigorous about their beliefs. The sad truth is that most people will remain blind. Given that, I'd much rather they blindly follow science than blindly follow religion, for the sake of the continued prosperity of our society.
Both have been proven wrong in many instances.
Here's the way I see it. Science is couched in the provable. It does not claim to have absolute truth, just some useful likeness of it. Religion is couched in belief. Organized religion, anyway, does claim to have absolute truth. Science can be wrong and retain its credibility. Religion cannot.
This does not mean that all life is a product of evolution.
The fact that things evolve is not the scientific evidence that is used to support the theory of evolution. The facts that are used to support evolution, well they do imply that all life is a product of evolution.
As you may have guessed, I do believe in the bible and creation. That does not mean that I do not accept scientific fact.
How much do you believe the Bible? If you believe it in its entirety, then you cannot claim that you accept scientific fact, because the Bible contradicts many such scientific facts. If you only believe in the Bible partially, then you may claim to accept scientific facts, but then realize that the "pick and choose" varient of Christianity that seems so popular today is rather at odds with what the Bible says.
Ultimately, I don't really care what people choose to believe. Being an engineer, my only interest is the pragmatic. I don't oppose the efforts of the religious to teach creationism as science, etc, because I care what children believe --- I oppose it because it will lead to fewer biologists. The way I see it, Jesus doesn't pay the bills, science does. The only purpose of public education is to create a productive workforce. Teaching religion doesn't do that, teaching science does.
"Moral values are ailing in this country. We need to do more to pull people in to Christianity." Sound familiar?
Yeah sure, and aside from the "pull people to Christianity" bit, I have no problem with trying to address to moral values problem in the country (as long as its not the government trying to do it). However, unlike science, moral values are not the bedrock of our economy and prosperity as a nation...
I think that's an overly optimistic viewpoint. Science is ailing in this country. The high-profile crusades against it from the creationists is just the tip of the iceberg. Far more important is the fact that science just plain isn't held in high regard, at a cultural level, and not enough Americans are persuing careers in the various scientific fields. On top of that is all the snake-oil masquerading as science, and the fact that the general public really has no idea of what is and is not science. Of course, that is nothing new, but it is something that universal education was supposed to fix. Well, in that case, universal education has failed. It is not at all surprising to see why, though. In the vast majority of class rooms in the US, science is taught not as a set of principles and methods, but as a loosly-connected facts. Students are not taught how to think scientifically, but are mearly forced to learn tidbits of information that may as well have just been pitching statistics for all the good they do.
Um, that struct from the Solaris code isn't portable to an Alpha-based Cray (which are ILP64 machines). It also assumes alignment rules all over the place. On top of all that, it'd take me maybe a day to come up with a Lisp macro that let you specify a byte-accurate "data packet" format, without all the unportable assumptions regarding the size and alignment of ints and chars.
Actually, I'm expressing horror at the consequences of interpreting a non-machine-native IL. Adequate performance entails implementing a full virtual machine with runtime code generation.
Or a install-time compiler. Or a regular compiler with signed binaries.
It's not the programmer I'm worried about. It's the end user I'm worried about; at the end of the day, he's the one taking the risk of running code on his machine. This system design leaves him vulnerable to local DoS attacks.
How so? With the compile-at-install model, the only attack vector is defeating the protections in the trusted compiler. I don't think that's particularly easier that defeating the protections in a 'trusted' kernel.
Provide control over the runtime environment, with the explicit possibility of no runtime (which C does and Lisp does not)
Define "runtime". C++ has a runtime. C++ can and has been used for writing high-performance microkernels. Lisp can be used so it requires no bigger a runtime than C++. As long as you don't try to use EVAL in the kernel, you're set. All it requires is a GC, and I find your criteria for excluding the GC capricious and aribtrary --- one that only has merit in the specific case of RT kernels.
We are discusing high-level languages in general which perform field reordering optimizations;
If, as you just said, you "stick to machine words" what does field reordering have to do with anything?
As discussed, C forbids this, so it's already a little bit easier to talk to the hardware.
But it says nothing about alignment, so even though its more predictable, its still completely useless. To put things into concrete terms, you can't use a struct like this to talk to hardware:
struct foo
{
int rega;
char regb;
int regc;
};
Well, you can, but its completely unportable. Sure, at least you know that 'regc' is after 'regb', but you still don't know where 'regc' is.
Your grasp of statistics is marginal at best. The FBI doesn't do 250m investigations every year, so that success rate is hardly 99.xx%. The FBI had 34,451 criminal investigations last year. That's around a 0.6% failure rate. Good enough for government work, probably, but let's recall that Apple is dealing with a class-action lawsuit because the iPod Nano's acrylic has a failure rate of 0.1%...
The number of murders in the US last year was 16137. Out of 295,734,134 people, that's a rate of 0.0054%, a figure that is more or less negligable. There is absolutely no point in getting worked up over murders with those kind of numbers.
An agent can make an honest mistake in the performance of a job (just like we all can, right?)
I dunno. Ford's engineers make a mistake, people die, and the company is sued for tens of millions of dollars. Merck's chemists make a mistake, people die, and the company is sued for tens of millions of dollars. FBI's agents make a mistake, and what exactly happens?
The "we're all human and we make mistakes" line wrt. the present administration is really wearing thin. Man, I'd love to be in a line of work I could fuck up that much and still keep my job...
I realize the two situations are quantitatively (if not qualitatively) different, but I wonder if there were apologists like you for the KGB?
The "you never hear the good stuff that happens" line is completely useless, because its not verifiable. You just assume good stuff happens, but since you have no proof, you can just as easily assume that it doesn't. Which side you choose is arbitrary, but I'm inclined to choose the latter, since at least I have the evidence of history to suggest its correctness.
Of course it wouldn't do. The fact that the agency investigates the violations doesn't change the fact that the violations shouldn't be happening in the first place. Creating a problem, then trying to fix it isn't as good as not creating a problem in the first place.
Actually, the GPUs are 128-bit, with 32-bit floats.
The Athlon64's adds and muls are pipelined, so its still 1 per cycle. If you look at the pipeline of a GPU (dozens of stages), you'll probably find that add and mul takes even longer.
So? That's running at 490 mHz core 1300 mhz memory. This Radeon is running at 1000Mhz core and 2000Mhz memory.
Right, which is why you stick to machine words.
Right. If you're sticking to machine words, then who *cares* that C's rules for laying out structs is more predictable than what the Lisp optimizer might do to a struct? You can use a read/write oriented machine-word module in Lisp as easily as you can do the same in C!
And researchers who want to build a general-purpose operating system that can be used in all of these scenarios are what? Shit out of luck?
I think you've missed the point entirely. The whole point of this line of discussion was that such researchers were barking up the wrong tree. The microkernel itself becomes useless (and all the massive contortions required to support it), when the language itself is safe. The kernel ceases to exist as a seperate entity and becomes a seperate (though special, in that it can use unsafe operations) library. You stop worrying about protection domains, context switches, interprocedural communication, drivers in userspace vs kernelspace, how to implement paging, etc. Protection becomes fine-grained (per object), context switches become cheap, protection-domains become unnecessary, interprocedural communication becomes wicked fast (shared memory), and userspace vs kernelspace ceases to have any meaning. All you give up (in user code) is: pointers and manual memory management.
For hard-real time, CPU-intensive uses, the second concession might be a deal-breaker. In that case, the microkernel has some semblence of relevance. However, for the vast majority of systems which don't need hard-RT, the concessions are minimal compared to the gains.