I've heard horror stories that the contractor will rip out the cable and charge the buyer for any damaged studs/wallboard. The logic is that the contractor usually still owns all the material and can't accept the legal liability if the buyer screws up the work before the job is done.
Engineering contractors are providing special knowledge as their service. The only 'tools' they use on the job are their minds (that the contractor is supplying).
It is a fine line, but to use the electrician example, an electrician who works on site for only one company using company tools and company procedures and doing whatever the company says is NOT a contractor under the law. Yes many companies are doing this these days (2 out of the 3 i've worked for), but it is illegal.
good point, but you will encounter greater force again when you go back 'up' on the other side of center. It all works out about the same, you just would tend to 'glide' more in the center without gaining or losing speed than in the periphery.
You can have an instantaneous just-about-anything if you know your calculus. I didn't say anything about conservation of momentum, I said it doesn't change - subtle difference. The reason is that you are adding mass with a velocity of 0 in the frame of reference. That adds 0 momentum to the system, but mass is increased. I make no claims of conservation since this is an open system (one that is gaining mass without limit.) F=ma is inappropriate in this context. You should be considering using E = 1/2 * m(v^2) or p = m*v
Note that conservation of momentum only applies if no outside forces are applied. That is only appropriate for a closed system like a textbook case of inelastic collision.
The Pisa experiment compared two masses starting at rest. This pendulum would be a moving mass absorbing a stationary mass. There would be a loss of velocity, but no loss/gain in momentum at any instant when a unit of mass is absorbed. So on the next swing up, you have a larger mass with insufficient momentum and velocity to regain the initial (subterrianian) altitude.
Agreed. Except there will be a lawsuit. You don't screw over 119 students and their families without at least a few having the means and motivation to take this to court.
Harvard is walking into a huge legal quagmire. If this was an 'ethical litmus test' as Harvard is purporting implicitly, then they must apply this in a uniform manner. They clearly did not. This is very true of job applications, and probably applies to college admissions as well.
It would be unethical, and illegal, if the students were able to change their admissions status due to a hack. Harvard would also have been negligent if this were to happen, since it would disadvantage many applicants through Harvard's own IT choices. But this wasn't the case. The people who chose to view their admissions information knew the magnitude and nature of what they were doing (minor), and had no intent to do anything but view their own admissions status.
Harvard should have just let the matter go away quietly. Fix the incompetent IT practices allowing for this to happen in the first place, and let it go away. The worst that could happen is that a potential student makes a poor admissions decision based on this tentative data.
NAT isn't a security device by design. But it is about the most effective thing you can do for a low cost. NAT (address translation) has the beneficial side-effect of keeping unsolicited packets away from vulnerable boxes.
If the intruders do not have a method of addressing your PC from the outside world (as in the case of NAT), then they cannot send you 'pings of death.'
They can, however, get to your computer through user-initiated connections, such as weaknesses in browsers or chat clients.
If I bought a new Mercedes, i would have an expectation that the seats would not have holes in them, or that the fuel line not leak. That is a reasonable expectation. If there were a reason to believe otherwise, it is the responsibility of the vendor to disclose such information.
Likewise, when I bought my last laptop, I expected the screen to work as advertised (1024x768 pixels ALL of them working) I got screwed, and was never made aware until after the fact. Whether or not there is an ISO standard to define LCD quality is irrelevant, the LCD should work just like the one on display in the showroom.
I look forward to the class-action lawsuit that is long overdue for this issue.
I think everone is missing the analysis done by the author of this attack (including the author).
The point is that the malicious code is in both versions of the downloaded package. The swappable payload just serves as an activation key. The actual mal-code can be present in both packages 'hidden inside an AES encrpted data section.'
Yes MD5's have collisions, so do any current or future hashes or checksums. The point is it's still very very unlikely that anyone can insert arbitrary and functional code into *someone else's* binary package and get the same checksum.
It is still safe to use MD5 for passwords (remember to salt though), it's still fine for checksumming files.
The issue is whether a signed/hashed piece of code has a hidden bad payload or not. A hash will not protect you from mal-ware if the creator of the package did this intentionally. The trigger could simply be a date check, or your ip address or some other external data. The only effect of making two versions of the same package is that you can deploy the package to everyone without detection after the first package passed review. But like I said, you can pass a tainted package past review if the payload isn't triggered. Finally, the security of a checksum assumes that the package is somehow reviewed by a credible source, and that they catch all possible obfuscated attacks.
Exactly my point. Financing the national debt isn't the same thing as going to get an amortized (mortgage) loan. It is financed by investors, not by some mega bank loan.
True, so many problems are software issues which require the AI to have a complete and introspective understanding of itself and the system.
Hardware repairs can be much more modularized and are easily accomplished today with 10 year old technology. i.e. 'Overtemp detected in server module - send robot 57 to datarack 2632 and hotswap the unit.' The actual failure rate of hardware is low enough to not require much automation though.
There is real contracting, and then there is 'contracting' which sounds like this situation. The latter, 'contracting', is really just working as a temp or through a temp agency who's office you've never set foot in. The 'contract' is basically: "We'll hire you without benefits as long as we can and pay you a few bucks less until we determine that you are docile enough to work for us on our payroll."
Your options are:
- Write your own contract which is truly equitable and see if they bite.
- Take the job as stated above, which I did, and hope for the best.
- Keep looking for work.
Garbage men are allowed to work autonomously. No PHB destroying the pride in thier work. Yes, there is pride in any work done correctly.
Garbage men, god bless them, are likely to come from a lower working class. There is still a lot to look up to from that perspective. Any measure of success is self earned, and not from a good 'ol boy connection. Occasionally, a garbage man's kid might get a merit scholarship to college.
They don't have the problem upper-middle and upper classes have. i.e. to satisfy their soccer-mom wife with expensive houses on the nice side of town and overpriced SUV's, all while saving for their kids' college education and still paying off their own.
That being said, it's nice to be somewhere in the middle.
exactly. And my orignal point was to dispute the poster who said that only diesel engines have the torque available to move a truck. Electric motors have the best torque curves across all speeds compared to combustion engines. Electric trains and trucks are usually direct drive, or fixed gearing to avoid complexity and failure points.
Obviously, the electricity has to come from somewhere. But an electric drive-train opens up many other options: regenerative braking with a battery, dynamic braking which is more reliable and conisitant than mechanical brakes, and constant RPMs on the diesel for efficiency and reduced wear.
Python is a highly flexible framework/environment. In your example you can use the existing list.sort() if it immediately meets your needs in order to get the program written quickly. You can also define your own sort function and plug that in, which I frequently do. If you had a performance issue, you can always drop back down to C to do it.
I think Python is better than just another high-level language. It always seems to be the simplest, best factored way that I can express a problem. I've often thought about how to design a system like the article describes where the program is defined in natural language. But I always come back to the conclusion that natural language is too incomplete to define determinate processes. There is nothing wrong with the common programming structures: loops, conditionals, storage and operators. It really is the best way to represent software. So much so that even hardware is designed using these structures (VHDL).
The error rate that the article attributes conventional programming structures is BS. Errors result from the practical limitations of the languages. For example, in C, doing so much manual memory allocation, poor string handling capabilities, poor exception handling and easily misplaced operators are the usual cause of problems. To knowledgeable programmers, errors have nothing to do with the underlying constructs themselves. All of these issues went away when I did more projects in Python. Python lets me express a problem as I intend it, so much so that I had to push aside my conventional mindset when trying to learn Python. The idea that I can just put data in a list without allocation or pointers was difficult to get used to.
There are always going to be some concessions due to the concrete nature of a real-world programming system. I think the article was too pie-in the sky as far as asking for a new paradigm that they cannot give a concrete working example of. The best systems are like Python that allow a problem to be solved by choosing what level of abstraction I want to handle it at. It's also very easy to test small parts of programs in Python.
Once part of the article caught my attention:
"If programmers have a weak hypothesis about the cause of a failure, any implicit assumptions about what did or did not happen at runtime will go unchecked."
This is a very valid point raised in the article, programming languages should have more facilities built into them besides conventional debuggers. Easy to use tools should be routinely used to better validate input/output and possible conditions in a program on a more complete scale.
As for performance, yeah it kinda sucks. It would be nice if Python had more development spent on tackling performance issues. To do what the article describes, you could use Python syntax to output compiled-machine code programs. Tools like Pyrex do this already.
Interesting thought, but that woudn't be an issue. Those examples of 'slowing the speed of light' aren't really the relativistic speed of light. It has more to do with slowing the propagation of photons in matter. Even ordinary window glass creates this phenomenon to a small degree (this is the cause of refraction.) Also, if your assertion were correct, then the earth would stop spinning on its axis whenever one of these experiments are tried (or the experiment itself would go flying off the earth) to compensate for the acceleration vector created by the rotating earth.
But Capitalism is a zero-sum game at any one point in time. There are finite production possibilities. The economy can only grow at a certain rate based mostly on technological factors. Resources are still scarce and un-evenly distributed in Capitalistic society. There is better, but not perfect, incentive to push the boundaries of productivity and distribution.
Regulation is required even in a 'free' market because natural monopolies form automatically in certain industries. A natural monopoly is still bad though. Proponents of de-regulation who want to introduce competition are actually self defeating in cases where the industry would revert back to an un-regulated monopoly. In this case though, VoIP isn't the monopoly due to the non-exclusive nature of the technology and the relatively low capital requirements, it is the broadband carriers that are to be watched closely.
Yeah, as someone who graduated a year too late to get a decent programming job, I can say that 'entry-level' is a thing of the past. I did get a job, but it is dead-end. Companies are still living in 2002 and think they can get PhD's with 20 years experience for $40k. I see many mid-level job positions with hyped-up requirements that go unfilled for 6 months or more. Only now are workers starting to burn out from being overloaded by this employment gap. The pendulum is about to swing back in a big way.
Slashdot Mirror
I've heard horror stories that the contractor will rip out the cable and charge the buyer for any damaged studs/wallboard. The logic is that the contractor usually still owns all the material and can't accept the legal liability if the buyer screws up the work before the job is done.
Engineering contractors are providing special knowledge as their service. The only 'tools' they use on the job are their minds (that the contractor is supplying).
It is a fine line, but to use the electrician example, an electrician who works on site for only one company using company tools and company procedures and doing whatever the company says is NOT a contractor under the law. Yes many companies are doing this these days (2 out of the 3 i've worked for), but it is illegal.
Looks like you accidentally hit the nail on the head. A true contractor is the CEO of their own firm.
good point, but you will encounter greater force again when you go back 'up' on the other side of center. It all works out about the same, you just would tend to 'glide' more in the center without gaining or losing speed than in the periphery.
You can have an instantaneous just-about-anything if you know your calculus. I didn't say anything about conservation of momentum, I said it doesn't change - subtle difference. The reason is that you are adding mass with a velocity of 0 in the frame of reference. That adds 0 momentum to the system, but mass is increased. I make no claims of conservation since this is an open system (one that is gaining mass without limit.)
F=ma is inappropriate in this context.
You should be considering using E = 1/2 * m(v^2) or p = m*v
Visit this site:
"The law of conservation of momentum states that the total momentum of any group of objects remains the same unless outside forces act on the objects."
Note that conservation of momentum only applies if no outside forces are applied. That is only appropriate for a closed system like a textbook case of inelastic collision.
You are partly insightful.
The Pisa experiment compared two masses starting at rest. This pendulum would be a moving mass absorbing a stationary mass. There would be a loss of velocity, but no loss/gain in momentum at any instant when a unit of mass is absorbed. So on the next swing up, you have a larger mass with insufficient momentum and velocity to regain the initial (subterrianian) altitude.
Agreed. Except there will be a lawsuit. You don't screw over 119 students and their families without at least a few having the means and motivation to take this to court.
Harvard is walking into a huge legal quagmire. If this was an 'ethical litmus test' as Harvard is purporting implicitly, then they must apply this in a uniform manner. They clearly did not. This is very true of job applications, and probably applies to college admissions as well.
It would be unethical, and illegal, if the students were able to change their admissions status due to a hack. Harvard would also have been negligent if this were to happen, since it would disadvantage many applicants through Harvard's own IT choices. But this wasn't the case. The people who chose to view their admissions information knew the magnitude and nature of what they were doing (minor), and had no intent to do anything but view their own admissions status.
Harvard should have just let the matter go away quietly. Fix the incompetent IT practices allowing for this to happen in the first place, and let it go away. The worst that could happen is that a potential student makes a poor admissions decision based on this tentative data.
NAT isn't a security device by design. But it is about the most effective thing you can do for a low cost. NAT (address translation) has the beneficial side-effect of keeping unsolicited packets away from vulnerable boxes.
If the intruders do not have a method of addressing your PC from the outside world (as in the case of NAT), then they cannot send you 'pings of death.'
They can, however, get to your computer through user-initiated connections, such as weaknesses in browsers or chat clients.
If I bought a new Mercedes, i would have an expectation that the seats would not have holes in them, or that the fuel line not leak. That is a reasonable expectation. If there were a reason to believe otherwise, it is the responsibility of the vendor to disclose such information.
Likewise, when I bought my last laptop, I expected the screen to work as advertised (1024x768 pixels ALL of them working) I got screwed, and was never made aware until after the fact. Whether or not there is an ISO standard to define LCD quality is irrelevant, the LCD should work just like the one on display in the showroom.
I look forward to the class-action lawsuit that is long overdue for this issue.
Wouldn't surprise me at all if there were still 16 bit integers in an old creaky database system.
It flew under the weather radar satellites - LITERALLY. There.
I think everone is missing the analysis done by the author of this attack (including the author).
The point is that the malicious code is in both versions of the downloaded package. The swappable payload just serves as an activation key. The actual mal-code can be present in both packages 'hidden inside an AES encrpted data section.'
Yes MD5's have collisions, so do any current or future hashes or checksums. The point is it's still very very unlikely that anyone can insert arbitrary and functional code into *someone else's* binary package and get the same checksum.
It is still safe to use MD5 for passwords (remember to salt though), it's still fine for checksumming files.
The issue is whether a signed/hashed piece of code has a hidden bad payload or not. A hash will not protect you from mal-ware if the creator of the package did this intentionally. The trigger could simply be a date check, or your ip address or some other external data. The only effect of making two versions of the same package is that you can deploy the package to everyone without detection after the first package passed review. But like I said, you can pass a tainted package past review if the payload isn't triggered. Finally, the security of a checksum assumes that the package is somehow reviewed by a credible source, and that they catch all possible obfuscated attacks.
Exactly my point. Financing the national debt isn't the same thing as going to get an amortized (mortgage) loan. It is financed by investors, not by some mega bank loan.
Use: Principle * e ^ (rate*years)
for a compounding interest calculation.
At only 4.5% over 10 years, this is a gross payback of 156%
Excel's formula sounds like simple interest.
Also, college loan rates are very low compared to T-bills or even the prime rate. Your loan may even be subsidised.
True, so many problems are software issues which require the AI to have a complete and introspective understanding of itself and the system.
Hardware repairs can be much more modularized and are easily accomplished today with 10 year old technology. i.e. 'Overtemp detected in server module - send robot 57 to datarack 2632 and hotswap the unit.' The actual failure rate of hardware is low enough to not require much automation though.
Exactly,
There is real contracting, and then there is 'contracting' which sounds like this situation. The latter, 'contracting', is really just working as a temp or through a temp agency who's office you've never set foot in. The 'contract' is basically: "We'll hire you without benefits as long as we can and pay you a few bucks less until we determine that you are docile enough to work for us on our payroll."
Your options are:
- Write your own contract which is truly equitable and see if they bite.
- Take the job as stated above, which I did, and hope for the best.
- Keep looking for work.
Garbage men are allowed to work autonomously. No PHB destroying the pride in thier work. Yes, there is pride in any work done correctly.
Garbage men, god bless them, are likely to come from a lower working class. There is still a lot to look up to from that perspective. Any measure of success is self earned, and not from a good 'ol boy connection. Occasionally, a garbage man's kid might get a merit scholarship to college.
They don't have the problem upper-middle and upper classes have. i.e. to satisfy their soccer-mom wife with expensive houses on the nice side of town and overpriced SUV's, all while saving for their kids' college education and still paying off their own.
That being said, it's nice to be somewhere in the middle.
exactly. And my orignal point was to dispute the poster who said that only diesel engines have the torque available to move a truck. Electric motors have the best torque curves across all speeds compared to combustion engines. Electric trains and trucks are usually direct drive, or fixed gearing to avoid complexity and failure points.
Obviously, the electricity has to come from somewhere. But an electric drive-train opens up many other options: regenerative braking with a battery, dynamic braking which is more reliable and conisitant than mechanical brakes, and constant RPMs on the diesel for efficiency and reduced wear.
The biggest of trucks are electric. What you say?? here: AC solutions for haul trucks
Jeez people... not even 11k signatures yet. weak.
I think Python is better than just another high-level language. It always seems to be the simplest, best factored way that I can express a problem. I've often thought about how to design a system like the article describes where the program is defined in natural language. But I always come back to the conclusion that natural language is too incomplete to define determinate processes. There is nothing wrong with the common programming structures: loops, conditionals, storage and operators. It really is the best way to represent software. So much so that even hardware is designed using these structures (VHDL).
The error rate that the article attributes conventional programming structures is BS. Errors result from the practical limitations of the languages. For example, in C, doing so much manual memory allocation, poor string handling capabilities, poor exception handling and easily misplaced operators are the usual cause of problems. To knowledgeable programmers, errors have nothing to do with the underlying constructs themselves. All of these issues went away when I did more projects in Python. Python lets me express a problem as I intend it, so much so that I had to push aside my conventional mindset when trying to learn Python. The idea that I can just put data in a list without allocation or pointers was difficult to get used to.
There are always going to be some concessions due to the concrete nature of a real-world programming system. I think the article was too pie-in the sky as far as asking for a new paradigm that they cannot give a concrete working example of. The best systems are like Python that allow a problem to be solved by choosing what level of abstraction I want to handle it at. It's also very easy to test small parts of programs in Python.
Once part of the article caught my attention:
This is a very valid point raised in the article, programming languages should have more facilities built into them besides conventional debuggers. Easy to use tools should be routinely used to better validate input/output and possible conditions in a program on a more complete scale.
As for performance, yeah it kinda sucks. It would be nice if Python had more development spent on tackling performance issues. To do what the article describes, you could use Python syntax to output compiled-machine code programs. Tools like Pyrex do this already.
Interesting thought, but that woudn't be an issue. Those examples of 'slowing the speed of light' aren't really the relativistic speed of light. It has more to do with slowing the propagation of photons in matter. Even ordinary window glass creates this phenomenon to a small degree (this is the cause of refraction.) Also, if your assertion were correct, then the earth would stop spinning on its axis whenever one of these experiments are tried (or the experiment itself would go flying off the earth) to compensate for the acceleration vector created by the rotating earth.
Picard went FTL in every episode, maybe you meant non-warp FTL??
I agree mostly with you that communism was bad.
But Capitalism is a zero-sum game at any one point in time. There are finite production possibilities. The economy can only grow at a certain rate based mostly on technological factors. Resources are still scarce and un-evenly distributed in Capitalistic society. There is better, but not perfect, incentive to push the boundaries of productivity and distribution.
Regulation is required even in a 'free' market because natural monopolies form automatically in certain industries. A natural monopoly is still bad though. Proponents of de-regulation who want to introduce competition are actually self defeating in cases where the industry would revert back to an un-regulated monopoly. In this case though, VoIP isn't the monopoly due to the non-exclusive nature of the technology and the relatively low capital requirements, it is the broadband carriers that are to be watched closely.
Yeah, as someone who graduated a year too late to get a decent programming job, I can say that 'entry-level' is a thing of the past. I did get a job, but it is dead-end. Companies are still living in 2002 and think they can get PhD's with 20 years experience for $40k. I see many mid-level job positions with hyped-up requirements that go unfilled for 6 months or more. Only now are workers starting to burn out from being overloaded by this employment gap. The pendulum is about to swing back in a big way.