Professor: James, what's happened? It seems like your studies have become disorganized - and your test scores have taken a turn for the worse. Is there anything I need to know?
Me: I'm having.. trouble sleeping, Professor. It.. it's hardest to watch for.. these.. these headcrabs.. especially at night.
Professor: *leans closer* I know; scary little bastards.
I largely agree with you - and it's a shame this incredibly simple fact eludes so many, especially in those in technical communities.
IMHO, computer crime is an attractive hobby for many that convince themselves of the impossibility of being traced and the difficulties associated with enforcement.
For the most part this is the root of the problem: enforcement is lacking, especially overseas (as the Anonymous Coward points out). Herein one can refer to the book of Real Life under "cases were poor enforcement results in no change in behavior" (take prohibition for example).
From either point of view, the least amount of blame falls on the frequently uneducated consumer who paid for a product that was prone to being hacked. A little more blame on system administrators over their heads. Perhaps even more blame on the vendors producing the software (bucky128 makes a good point. Have you ever heard of the Chevrolet Corvair? What about all those wireless-routers shipping with WEP-disabled?)... but the most blame should always, always, always be placed on the criminals.
Unfortunately, this presents a difficult situation for everyone (blame almost becomes a commodity). Due to poor cooperation by "the authorities" and lax penalties, more and more "responsibility" is being placed on the vendors by the consumers (which are being told that computer crime is waning, when in reality it is increasing due to increased press coverage).
I love how so many people get caught up in the security arms race - jumping about from one standard to another. Perhaps the vendors producing software designed with security in mind will sell more products (Microsoft is clearly feeling the bite - *points to a delayed Longhorn*), but the real overhaul isn't needed in the code. It's need in the law - and mostly law overseas (though things like bounties - recently offered by vendors like Microsoft - can curb this without legal changes {most of this crime is for "masturbatory gratification" and that mentality follows "the more people watching, the better"}). Implementing changes in that will require decades, and until then the 'net will remain a Wild-Wild-West of sorts.
It's only human to wonder where it will fail. The fulfillment of the law is, as the anonymous coward stated so eloquently, psychological. When that point comes, the industry won't want to let it go: already IBM has changed it on their site from every 18 months to every two years. In fact, IBM is changing all sorts of benchmarks as they are beginning to attempt to shift the spotlight towards benchmarks like chip features, away from core statistics (i.e. "dropping the GHz").
We won't get too picayunish with atoms: that point will never be reached for pragmatic reasons: for instance, should this trend continue without new advances in materials or cooling technology, your computer will sound like a hairdryer in about a decade.
Moore's Law will have to evolve, at least in definition, to quantify something else. That, or perish; and soon.
Assume the IC Mr. Moore was describing above is 0.1 by 0.1 inches. According to Moore's law, when would the size of a transistor become the same size as an atom of silicon? Solution: Find the approximate volume of one silicon atom. Start by reviewing basic chemistry, and in particular Avogadro's number (6.02214199 * 1023). Also find the atomic weight of Silicon (Si), which is about 28. This means that 6.02214199 * 1023 atoms of silicon will weigh 28 grams. Now we have to figure out the volume of 28 grams of silicon. Searching on the web for density of silicon produced this page which says 28 grams of silicon will occupy 0.000012 cubic meters of space. That must mean one atom of silicon will occupy about 0.000012 / 6.02214199 * 1023 = 20 * 10^-30 cubic meters, or a cube about 270 * 10-12 on a side. An IC which is.1 by.1 inches is about 2.54 * 10-3 meters on a side. If we measure that in widths of a Si atom, that would be 2.54 * 10-3 / 270 * 10-12 = 10 million atoms on a side. The surface of the chip would contain about 10 million squared atoms, or 100 * 1012 atoms. So now we need to know how many times you need to double the number 200 to get up to 100 * 1012, we can solve that by the equation 200*2x=100*1012. That leads us to an x of about 39. This means it will take about 78 years before the size of a transistor matches the size of an atom. 1965 + 78 = 2043. So in about 40 years from now.
Slashdot Mirror
Professor: James, what's happened? It seems like your studies have become disorganized - and your test scores have taken a turn for the worse. Is there anything I need to know?
.. it's hardest to watch for .. these.. these headcrabs.. especially at night.
Me: I'm having.. trouble sleeping, Professor. It
Professor: *leans closer* I know; scary little bastards.
Aliens vs. Predator. Hands-down.
I largely agree with you - and it's a shame this incredibly simple fact eludes so many, especially in those in technical communities.
IMHO, computer crime is an attractive hobby for many that convince themselves of the impossibility of being traced and the difficulties associated with enforcement.
For the most part this is the root of the problem: enforcement is lacking, especially overseas (as the Anonymous Coward points out). Herein one can refer to the book of Real Life under "cases were poor enforcement results in no change in behavior" (take prohibition for example).
From either point of view, the least amount of blame falls on the frequently uneducated consumer who paid for a product that was prone to being hacked. A little more blame on system administrators over their heads. Perhaps even more blame on the vendors producing the software (bucky128 makes a good point. Have you ever heard of the Chevrolet Corvair? What about all those wireless-routers shipping with WEP-disabled?)... but the most blame should always, always, always be placed on the criminals.
Unfortunately, this presents a difficult situation for everyone (blame almost becomes a commodity). Due to poor cooperation by "the authorities" and lax penalties, more and more "responsibility" is being placed on the vendors by the consumers (which are being told that computer crime is waning, when in reality it is increasing due to increased press coverage).
I love how so many people get caught up in the security arms race - jumping about from one standard to another. Perhaps the vendors producing software designed with security in mind will sell more products (Microsoft is clearly feeling the bite - *points to a delayed Longhorn*), but the real overhaul isn't needed in the code. It's need in the law - and mostly law overseas (though things like bounties - recently offered by vendors like Microsoft - can curb this without legal changes {most of this crime is for "masturbatory gratification" and that mentality follows "the more people watching, the better"}). Implementing changes in that will require decades, and until then the 'net will remain a Wild-Wild-West of sorts.
Europe must be very patient.
It's only human to wonder where it will fail. The fulfillment of the law is, as the anonymous coward stated so eloquently, psychological. When that point comes, the industry won't want to let it go: already IBM has changed it on their site from every 18 months to every two years. In fact, IBM is changing all sorts of benchmarks as they are beginning to attempt to shift the spotlight towards benchmarks like chip features, away from core statistics (i.e. "dropping the GHz").
We won't get too picayunish with atoms: that point will never be reached for pragmatic reasons: for instance, should this trend continue without new advances in materials or cooling technology, your computer will sound like a hairdryer in about a decade.
Moore's Law will have to evolve, at least in definition, to quantify something else. That, or perish; and soon.
Assume the IC Mr. Moore was describing above is 0.1 by 0.1 inches. According to Moore's law, when would the size of a transistor become the same size as an atom of silicon? Solution: Find the approximate volume of one silicon atom. Start by reviewing basic chemistry, and in particular Avogadro's number (6.02214199 * 1023). Also find the atomic weight of Silicon (Si), which is about 28. This means that 6.02214199 * 1023 atoms of silicon will weigh 28 grams. Now we have to figure out the volume of 28 grams of silicon. Searching on the web for density of silicon produced this page which says 28 grams of silicon will occupy 0.000012 cubic meters of space. That must mean one atom of silicon will occupy about 0.000012 / 6.02214199 * 1023 = 20 * 10^-30 cubic meters, or a cube about 270 * 10-12 on a side. An IC which is .1 by .1 inches is about 2.54 * 10-3 meters on a side. If we measure that in widths of a Si atom, that would be 2.54 * 10-3 / 270 * 10-12 = 10 million atoms on a side. The surface of the chip would contain about 10 million squared atoms, or 100 * 1012 atoms. So now we need to know how many times you need to double the number 200 to get up to 100 * 1012, we can solve that by the equation 200*2x=100*1012. That leads us to an x of about 39. This means it will take about 78 years before the size of a transistor matches the size of an atom. 1965 + 78 = 2043. So in about 40 years from now.