OK, I'll bite. I actually took a full-quarter class that focused on the decision to drop the atomic bomb, taught in a small group setting and lead by one of the leading scholars on Japan, Barton Bernstein. His information is here: http://www.stanford.edu/group/STS/bernstein.shtml. We read about 1,000 pages of recently declassified documents, as well as several secondary interpretations from widely respected historians, so hopefully what I have to say has some intelligence behind it.
I think you have been thinking about this in the wrong frame of mind - you are assuming that leaders were actively assesing Japanese potential lives lost versus American invasion casualties. I won't get into the ethical debate about whether its better to kill civilians versus military personnel because they are preparesd to die (largely because I think its a pointless debate), but there is the problem that the US leaders did not expect the Atomic Bomb to make Japan surrender. They merely viewed it as another weapon in the arsenel of the US, to be used at will, without concern for Japanese casualties. One of the most telling documents we came across was a memo from then Secretary of War Stimson to Admiral Forrestal the day after the first bomb was dropped, advising to continue procurement and armament, as if nothing had happened in the war. There wasn't even a hint of, 'well, let's wait and see'. There were no demobilization plans drawn up before Japan surrendered, because no one expected them to. Furthermore, from what documents exist about the Japanese leaders state of mind, it appears that the invasion of the Soviets after the first bomb but before Nagasaki bomb contributed more to their surrender than the bombs.
If you want to get into a numbers argument, it turns out that the Japanese actually had significantly less troops and materiel available to defend against a Japanese invasion that American intelligence had predicted (on the order of 1/2 as much). Furthermore, American intelligence had predicted between 180,000-230,000 casualties from both the invasion of Kyshu and the Tokyo plain (both operations). Casualties, in military terms, means killed or wounded. Assuming the average, consistent with other battles in the Pacific, of about a 1 to 10 rate of killed to wounded, or even being charitable and assuming 1 to 5, this could mean only as high as 40,000 deaths for both invasions. This compares to 250,000 killed outright from both bombs. Hardly an even count. Even more interesting, after the war, a survey was done by the Air Force to asses the effectiveness of bombing. Although you have to somewhat suspect the conclusion because it is sort of self-serving, they estimated that the Air Force's plan to bomb eight key rail tunnels (it turns out that for whatever reason, there were some serious chokepoints in the Japanese rail system), they could have shut down the ability of the Japanese to transport the fall harvest from the countryside to the cities, leaving the Japanese literally starving in the streets, and probably would have forced surrender no later than mid to late October, which is earlier than the proposed November date of the invasion of Kyshu.
So to cover a little more about the mindset of the US leaders, President Truman has no recorded order to use the bombs on Hiroshima and Nagasaki - rather, he sent an order directing the use of Atomic Bombs 'as available'. He didn't pay particular attention to their use, and didn't think about them heavily except in the context of how they would change the post-war landscape. I personally suspect that he didn't really know how powerful/deadly they would be, but there really isn't records to prove that either way. Nor did most of the generals and admirals of the time, as noted in the memo between Forrestal and Stimson (and others) point out. It was a continuation of an acceptable policy by all sides, which became more accepted as the war went on - that it was morally acceptable to kill civilians. The Germans did it with the bombing of London and the V-2's. The Soviets displayed no compunction against killing Poles, Germans, and even their own citizens. We firebombed Tokyo one night in May and killed 40,000 civilians, with very little apparent military benefit. We also used the Atomic Bombs. It was a mindset that was very popular in the US - even after Japan surrendered, 20% of Americans said we should have kept dropping bombs in Japan until we blew it off the face of the earth. It was a fact of the war. I'm also not going to get into the debate about whether it is acceptable to judge history by today's standards, but suffice to say it should serve as a warning to us now. I hope this clears some things up.
Re:Interesting Lecture.
on
CPRM Lecture
·
· Score: 1
Ok, after actually ** attending in person ** this lecture, I think most of the posters are missing what he was saying. There are more than 16 keys, and you could not compromise the sceme by finding 16 of them. They are using this scema where there is a tree-based key system, with each leaf being a key. Clearly, this is more than sixteen for any sizeable tree (although he didn't actually specify the tree size). Really, the best way to break this system is with the above mentioned attack on the serial number (but serial numbers are 2^64, so finding a matching serial number or one close would be nearly impossible, as serial numbers are incrementable).
The thing I'm most sceptical about is the C4's consotium's invalidation of media. The real war will be as hackers create a program with a hacked key. C4 invalidates the key in all new media. Hacker compromises another key, creates new program, etc. The issue is that Joe Blow on the street won't have the energy to constantly be looking for hack players. It will be a serious pain in the ass to hack, and only for people with a lot of time on their hands. By and large, Hollywood/RIAA will get their copy protection money on this (how many hackers do you suppose pay for music anyway?)
The other interesting thing is the guy from IBM was constantly referring to the 'rogue hacker in Norway' as an example of people who would want to get around the CPRM system. This is an obvious attempt to marginalize the people who actually want to use their paid-for media in legitimate ways (i.e. ways that are not against the law). It will be interesting to see what happens..
While several people have noted that it should be easy to recompile and run many of the standard UNIX utilities, I'm very curious as to how OS X's new file architecture (using packages...noted on/. a while ago) will affect UNIX programs. It seems Apple's trying to standardize how applications are stored in the file system, and my guess is a simple recompile doesn't really do that. But maybe not? Also, will there renaming of certain directories cause problems? Like/usr to/User? Seems like this should break many applications....but maybe not?
I largely agree with him on this one. As a fellow Stanford student, I've talked to quite a few students who had serious problems in the advanced C class porting their code from architecture X (usually WINTEL) to the platform to be tested on, Sun workstations. There is a large potential for things to go very wrong when trying to port
Issues such as big-endianess versus little-endianess become much less academic issues and things that look like they *should* work stop, it is about the most frustrating experience you can have in Computer Science. My hat goes off to people who work professionally at porting games, but at this early stage of your CS career, the agony isn't worth it.
Are you sure?? I'm really quite positive that most if not all of the contests that distributed.net has won were sponsored by RSA. Check this link out: http://www.distributed.net/rc5/ Maybe distributed.net is posting bogus news, but I higly doubt it:)
There's probably more truth to that than you might suspect. Encryption using PGP is inherently more complex. Since it requires two keys, as opposed to one key in traditional crypto, the math gets a lot more complicated. You can't just use a reversible function.
While there haven't been any real structural attacks to PGP, up until this, it is theoretically more likely that structural attacks will work against PGP than standard crypto. Perhaps the NSA has already found a way? Also, traditional PGP uses the RSA encryption algorithm, which, if you follow Distributed.net, gets brute-forced regularly. If you really are scared of the government reading your email, then I doubt PGP will put your fears to rest.
His distinction is interesting, because OS reasearch is moving in the opposite direction of his definition.
There has been some serious work done on stripping out most of the components in a modern OS, and having each application 'bringing their own' This concept is called the XO microkernel
What this means is that OS'es of the future may have even less functionality and the applications will contain libraries that implement more and more of the OS themselves - for the obvious efficency payoffs. Standard libraries that can be added into applications easily will have to be written of course, but the speedup gains are several orders of magnitude (like x40, using Apache as the benchmark)
This of course completely negates his point, but then again, it's mostly semantics anyway. He just chose to define it in a way that most people don't.
I agree that the current technological implementation leaves a lot to be desired, but on the other hand, consider that pin-touch technology is trying to agument human touch, by providing essentially sensory overload, while Immersion seeks to simulate. This is the important distinction. The mouse seeks to use vibrations to simulate texture and 'roughness'. This is most accurately done through vibrations. While I'm not an expert at haptics, I do know some people that are intimately involved in the subject, and vibrations can be expressed in mathmatical terms - via waves and other known quantities. It is much harder to model general textures in any other way (such as creating a grid maybe) and is basically impossible to do in any general way.
Actually, vibration is probably best. While doing some programming using Immersion technology (used in the Logitech Mouse), there are demonstrations about increased accuracy using their technology that are quite impressive. While this was using the Wingman FF mouse - which is significantly different, because the Wingman supports directional forces while the iFeel Logitech does not - the idea is the same. Most of the effects humans feel can be mathmatically modeled best with vibrations. The math involved in haptics (essentially the study of force feedback) is very complicated, but suffice to say that doing a pin-grid array doesn't really add much to the sensation of what you feel. Pin-grid can be very helpful, but are poor at simulating reality. I don't proposed that Force Feedback, especially in their current implementations, are the end all solution, but the direction that they are trying to move them in is not agumentation, but simulation of the human touch.
Rendering farms (who are the only ones who have the $ to buy Maya anyway) are not going to convert en masse to using Maya. I think it's impact will be minimal at best. Alias|Wavefront announced Maya for the Macintosh OS X will be ready roughly at roughly the same time, and I would suspect the creative professionals are more likely to be familiar with a Mac than with Linux. On top of that, Linux has not been a hit with IT people to the same extent NT has and on top of that, the Linux user base is not very graphically oriented. I really see the impact of this announcement to be very limited for Linux.
Slashdot Mirror
OK, I'll bite. I actually took a full-quarter class that focused on the decision to drop the atomic bomb, taught in a small group setting and lead by one of the leading scholars on Japan, Barton Bernstein. His information is here: http://www.stanford.edu/group/STS/bernstein.shtml. We read about 1,000 pages of recently declassified documents, as well as several secondary interpretations from widely respected historians, so hopefully what I have to say has some intelligence behind it.
I think you have been thinking about this in the wrong frame of mind - you are assuming that leaders were actively assesing Japanese potential lives lost versus American invasion casualties. I won't get into the ethical debate about whether its better to kill civilians versus military personnel because they are preparesd to die (largely because I think its a pointless debate), but there is the problem that the US leaders did not expect the Atomic Bomb to make Japan surrender. They merely viewed it as another weapon in the arsenel of the US, to be used at will, without concern for Japanese casualties. One of the most telling documents we came across was a memo from then Secretary of War Stimson to Admiral Forrestal the day after the first bomb was dropped, advising to continue procurement and armament, as if nothing had happened in the war. There wasn't even a hint of, 'well, let's wait and see'. There were no demobilization plans drawn up before Japan surrendered, because no one expected them to. Furthermore, from what documents exist about the Japanese leaders state of mind, it appears that the invasion of the Soviets after the first bomb but before Nagasaki bomb contributed more to their surrender than the bombs.
If you want to get into a numbers argument, it turns out that the Japanese actually had significantly less troops and materiel available to defend against a Japanese invasion that American intelligence had predicted (on the order of 1/2 as much). Furthermore, American intelligence had predicted between 180,000-230,000 casualties from both the invasion of Kyshu and the Tokyo plain (both operations). Casualties, in military terms, means killed or wounded. Assuming the average, consistent with other battles in the Pacific, of about a 1 to 10 rate of killed to wounded, or even being charitable and assuming 1 to 5, this could mean only as high as 40,000 deaths for both invasions. This compares to 250,000 killed outright from both bombs. Hardly an even count. Even more interesting, after the war, a survey was done by the Air Force to asses the effectiveness of bombing. Although you have to somewhat suspect the conclusion because it is sort of self-serving, they estimated that the Air Force's plan to bomb eight key rail tunnels (it turns out that for whatever reason, there were some serious chokepoints in the Japanese rail system), they could have shut down the ability of the Japanese to transport the fall harvest from the countryside to the cities, leaving the Japanese literally starving in the streets, and probably would have forced surrender no later than mid to late October, which is earlier than the proposed November date of the invasion of Kyshu.
So to cover a little more about the mindset of the US leaders, President Truman has no recorded order to use the bombs on Hiroshima and Nagasaki - rather, he sent an order directing the use of Atomic Bombs 'as available'. He didn't pay particular attention to their use, and didn't think about them heavily except in the context of how they would change the post-war landscape. I personally suspect that he didn't really know how powerful/deadly they would be, but there really isn't records to prove that either way. Nor did most of the generals and admirals of the time, as noted in the memo between Forrestal and Stimson (and others) point out. It was a continuation of an acceptable policy by all sides, which became more accepted as the war went on - that it was morally acceptable to kill civilians. The Germans did it with the bombing of London and the V-2's. The Soviets displayed no compunction against killing Poles, Germans, and even their own citizens. We firebombed Tokyo one night in May and killed 40,000 civilians, with very little apparent military benefit. We also used the Atomic Bombs. It was a mindset that was very popular in the US - even after Japan surrendered, 20% of Americans said we should have kept dropping bombs in Japan until we blew it off the face of the earth. It was a fact of the war. I'm also not going to get into the debate about whether it is acceptable to judge history by today's standards, but suffice to say it should serve as a warning to us now. I hope this clears some things up.
Ok, after actually ** attending in person ** this lecture, I think most of the posters are missing what he was saying. There are more than 16 keys, and you could not compromise the sceme by finding 16 of them. They are using this scema where there is a tree-based key system, with each leaf being a key. Clearly, this is more than sixteen for any sizeable tree (although he didn't actually specify the tree size). Really, the best way to break this system is with the above mentioned attack on the serial number (but serial numbers are 2^64, so finding a matching serial number or one close would be nearly impossible, as serial numbers are incrementable). The thing I'm most sceptical about is the C4's consotium's invalidation of media. The real war will be as hackers create a program with a hacked key. C4 invalidates the key in all new media. Hacker compromises another key, creates new program, etc. The issue is that Joe Blow on the street won't have the energy to constantly be looking for hack players. It will be a serious pain in the ass to hack, and only for people with a lot of time on their hands. By and large, Hollywood/RIAA will get their copy protection money on this (how many hackers do you suppose pay for music anyway?) The other interesting thing is the guy from IBM was constantly referring to the 'rogue hacker in Norway' as an example of people who would want to get around the CPRM system. This is an obvious attempt to marginalize the people who actually want to use their paid-for media in legitimate ways (i.e. ways that are not against the law). It will be interesting to see what happens..
While several people have noted that it should be easy to recompile and run many of the standard UNIX utilities, I'm very curious as to how OS X's new file architecture (using packages...noted on /. a while ago) will affect UNIX programs. It seems Apple's trying to standardize how applications are stored in the file system, and my guess is a simple recompile doesn't really do that. But maybe not? Also, will there renaming of certain directories cause problems? Like /usr to /User? Seems like this should break many applications....but maybe not?
I largely agree with him on this one. As a fellow Stanford student, I've talked to quite a few students who had serious problems in the advanced C class porting their code from architecture X (usually WINTEL) to the platform to be tested on, Sun workstations. There is a large potential for things to go very wrong when trying to port
Issues such as big-endianess versus little-endianess become much less academic issues and things that look like they *should* work stop, it is about the most frustrating experience you can have in Computer Science. My hat goes off to people who work professionally at porting games, but at this early stage of your CS career, the agony isn't worth it.
Are you sure?? I'm really quite positive that most if not all of the contests that distributed.net has won were sponsored by RSA. Check this link out: http://www.distributed.net/rc5/ Maybe distributed.net is posting bogus news, but I higly doubt it :)
There's probably more truth to that than you might suspect. Encryption using PGP is inherently more complex. Since it requires two keys, as opposed to one key in traditional crypto, the math gets a lot more complicated. You can't just use a reversible function.
While there haven't been any real structural attacks to PGP, up until this, it is theoretically more likely that structural attacks will work against PGP than standard crypto. Perhaps the NSA has already found a way? Also, traditional PGP uses the RSA encryption algorithm, which, if you follow Distributed.net, gets brute-forced regularly. If you really are scared of the government reading your email, then I doubt PGP will put your fears to rest.
His distinction is interesting, because OS reasearch is moving in the opposite direction of his definition.
There has been some serious work done on stripping out most of the components in a modern OS, and having each application 'bringing their own' This concept is called the XO microkernel
What this means is that OS'es of the future may have even less functionality and the applications will contain libraries that implement more and more of the OS themselves - for the obvious efficency payoffs. Standard libraries that can be added into applications easily will have to be written of course, but the speedup gains are several orders of magnitude (like x40, using Apache as the benchmark)
This of course completely negates his point, but then again, it's mostly semantics anyway. He just chose to define it in a way that most people don't.
I agree that the current technological implementation leaves a lot to be desired, but on the other hand, consider that pin-touch technology is trying to agument human touch, by providing essentially sensory overload, while Immersion seeks to simulate. This is the important distinction. The mouse seeks to use vibrations to simulate texture and 'roughness'. This is most accurately done through vibrations. While I'm not an expert at haptics, I do know some people that are intimately involved in the subject, and vibrations can be expressed in mathmatical terms - via waves and other known quantities. It is much harder to model general textures in any other way (such as creating a grid maybe) and is basically impossible to do in any general way.
Actually, vibration is probably best. While doing some programming using Immersion technology (used in the Logitech Mouse), there are demonstrations about increased accuracy using their technology that are quite impressive. While this was using the Wingman FF mouse - which is significantly different, because the Wingman supports directional forces while the iFeel Logitech does not - the idea is the same. Most of the effects humans feel can be mathmatically modeled best with vibrations. The math involved in haptics (essentially the study of force feedback) is very complicated, but suffice to say that doing a pin-grid array doesn't really add much to the sensation of what you feel. Pin-grid can be very helpful, but are poor at simulating reality. I don't proposed that Force Feedback, especially in their current implementations, are the end all solution, but the direction that they are trying to move them in is not agumentation, but simulation of the human touch.
Rendering farms (who are the only ones who have the $ to buy Maya anyway) are not going to convert en masse to using Maya. I think it's impact will be minimal at best. Alias|Wavefront announced Maya for the Macintosh OS X will be ready roughly at roughly the same time, and I would suspect the creative professionals are more likely to be familiar with a Mac than with Linux. On top of that, Linux has not been a hit with IT people to the same extent NT has and on top of that, the Linux user base is not very graphically oriented. I really see the impact of this announcement to be very limited for Linux.