Oil only has value because it is in high demand and is a scarce resource.
A new method of generating huge amounts of cheap energy will cause energy prices to fall. Which translates to less $$$ for energy companies (unless they manage increase their profit margin without getting undercutted [cartel, anyone?] ).
Please remind me again, why Exxon, Shell, Mobil, and other OIL COMPANIES are going to be interested in making this happen?
In terms of goverment, energy is also a national securtity issue. Ties between countries are often heavily focused on energy trade. What is Saudi Arabia going to sell to the world if the world doesn't need (much) oil? Sand!?!? ROTFL!
This guy should be going to electric companies instead, not oil companies.
This is a neat idea but I don't think it will happen for a long, long time. It could be a nightmare to maintain such a utility as the various components of Linux evolve over time. This would also be a massive project, given the long list of popularly used Windows & Linux programs (and their multiple verisions in use [Office 2000, XP, 2003, etc, etc ).
If course if you are doing Windows-Windows, you can easily movie all your user settings by just burning your %userprofile% directory to a CD. (It has your user's registry, My Documents, Application settings, etc... )
1. I sure hope you said astronomical, rather than astrological, or the astronomers will shoot you when they find out.
Nah... Astronomers don't need guns. They will just point the telescope at the sun (in daytime) and place the reporter under the eyepiece. Just like an ant under a magnifying glass....
I think its the letter of the law that confuses people.
If 1000 people camped out in the middle of a public road in front of the entrance to a company, would they be breaking a crime by not allowing people to enter/exit? In essence, they would be executing a "denial of service" attack to the companies road.
Or what if a few 18-wheelers decided to park in the middle of an interstate to block it. This is also a DOS attack.
What if 1 million people concertedly & simultaneously dialed 911 for "testing purposes" once a month. This would also be a DOS attack.
In each case different laws might be violated but the principle is the same: resources are being purposely mis-used in order to deprive others of them.
Now a question arises: is the Internet a public utility or just a privately owned network? IANAL, but the latter would seem to make the Britisher's offsense a Civil one, not a Criminal one.
You make some good points... For some types of calculations/applications, generating formal proofs isn't too difficult.
I have noticed that under some models, actually constructing the proofs (correctly) is more difficult than writing a correct program. Thus, the people who can do such proofs easily are probably less likely to screw up a small routine. (although such proofs may be highly valuable in regression tests).
I don't know much about LISP, but my limit experience with formal techniques has shown to me that it is difficult to implement an *efficient* correct program that is *simple* to prove correctness. That is if the proof is easy (due to the language), then efficiency is often lacking. Conversely, languages that allow for specializated techniques often are less amenable to formal proofs (e.g. assembly). Functional languages sometimes allow for elegant programs but at the expense of (unnecessary) data copying.
There is also another verification program that is seldom addressed. Pretend you wrote a function to calculate the arithmatic average of a list of numbers. You formally proved that your function is correct and that the result is indeed an arithmetic average. These proofs do not prevent the programmer from improperly applying this function in their application. For example, if the list of numbers was a list of rates, then the harmonic mean might be a more appropriate average. In other applications, the geometric mean might be more appropriate (e.g. average of a list of gain factors). Even for the same set of numbers, different means may be appropriate under different circumstances.
No formal language will ever be able to ensure the semantic correctness of programs. For this reason, I feel some types of program errors and security breaches will be inevitable. I see formal techniques a helpful crutch, but not an be-all, end-all solution.
The parent post does make a good point: There does need to a change in the way people write code.
Has anyone browsed the Changelog for the most recent Linux kernel? I recall seeing a networking bug that existed since 2.1.x!!!! It is quite scary software has progressed to where it is now almost completely without formal verification!
I wonder, does anyone even *need* an SSN to do much 'identity theft'? Sure, it is often demanded on forms, etc... But of those that ask for SSN, how many organizations actually verify it or use it in a matter that would implicitly verify it? I think someone's identity could be well-tarnished just with a name and address.
After all, banks in the US long ago stopped checking signatures on checks to see if they actually match... They basically will cash any check with a scribble on the signature line. I suspect the handling of SSNs might be similar in some circumstances.
Slashdot Mirror
Oil only has value because it is in high demand and is a scarce resource.
A new method of generating huge amounts of cheap energy will cause energy prices to fall. Which translates to less $$$ for energy companies (unless they manage increase their profit margin without getting undercutted [cartel, anyone?] ).
Please remind me again, why Exxon, Shell, Mobil, and other OIL COMPANIES are going to be interested in making this happen?
In terms of goverment, energy is also a national securtity issue. Ties between countries are often heavily focused on energy trade. What is Saudi Arabia going to sell to the world if the world doesn't need (much) oil? Sand!?!? ROTFL!
This guy should be going to electric companies instead, not oil companies.
This is a neat idea but I don't think it will happen for a long, long time. It could be a nightmare to maintain such a utility as the various components of Linux evolve over time. This would also be a massive project, given the long list of popularly used Windows & Linux programs (and their multiple verisions in use [Office 2000, XP, 2003, etc, etc ). If course if you are doing Windows-Windows, you can easily movie all your user settings by just burning your %userprofile% directory to a CD. (It has your user's registry, My Documents, Application settings, etc... )
WHY would they fire a woman in IT?????!?!??! Why!?!? The article just didn't state how many women were working at Novell. Only the men got fired.
Or perhaps the managers are typically female... In which case, 20% of the male employees probably made passes at them.
Nah... Astronomers don't need guns. They will just point the telescope at the sun (in daytime) and place the reporter under the eyepiece. Just like an ant under a magnifying glass....
I think its the letter of the law that confuses people.
If 1000 people camped out in the middle of a public road in front of the entrance to a company, would they be breaking a crime by not allowing people to enter/exit? In essence, they would be executing a "denial of service" attack to the companies road.
Or what if a few 18-wheelers decided to park in the middle of an interstate to block it. This is also a DOS attack.
What if 1 million people concertedly & simultaneously dialed 911 for "testing purposes" once a month. This would also be a DOS attack.
In each case different laws might be violated but the principle is the same: resources are being purposely mis-used in order to deprive others of them.
Now a question arises: is the Internet a public utility or just a privately owned network? IANAL, but the latter would seem to make the Britisher's offsense a Civil one, not a Criminal one.
You make some good points... For some types of calculations/applications, generating formal proofs isn't too difficult.
I have noticed that under some models, actually constructing the proofs (correctly) is more difficult than writing a correct program. Thus, the people who can do such proofs easily are probably less likely to screw up a small routine. (although such proofs may be highly valuable in regression tests).
I don't know much about LISP, but my limit experience with formal techniques has shown to me that it is difficult to implement an *efficient* correct program that is *simple* to prove correctness. That is if the proof is easy (due to the language), then efficiency is often lacking. Conversely, languages that allow for specializated techniques often are less amenable to formal proofs (e.g. assembly). Functional languages sometimes allow for elegant programs but at the expense of (unnecessary) data copying.
There is also another verification program that is seldom addressed. Pretend you wrote a function to calculate the arithmatic average of a list of numbers. You formally proved that your function is correct and that the result is indeed an arithmetic average. These proofs do not prevent the programmer from improperly applying this function in their application. For example, if the list of numbers was a list of rates, then the harmonic mean might be a more appropriate average. In other applications, the geometric mean might be more appropriate (e.g. average of a list of gain factors). Even for the same set of numbers, different means may be appropriate under different circumstances.
No formal language will ever be able to ensure the semantic correctness of programs. For this reason, I feel some types of program errors and security breaches will be inevitable. I see formal techniques a helpful crutch, but not an be-all, end-all solution.
The parent post does make a good point: There does need to a change in the way people write code.
Has anyone browsed the Changelog for the most recent Linux kernel? I recall seeing a networking bug that existed since 2.1.x!!!! It is quite scary software has progressed to where it is now almost completely without formal verification!
I wonder, does anyone even *need* an SSN to do much 'identity theft'? Sure, it is often demanded on forms, etc... But of those that ask for SSN, how many organizations actually verify it or use it in a matter that would implicitly verify it? I think someone's identity could be well-tarnished just with a name and address.
After all, banks in the US long ago stopped checking signatures on checks to see if they actually match... They basically will cash any check with a scribble on the signature line. I suspect the handling of SSNs might be similar in some circumstances.