Sheridan is a well respected British expert particularly on wrecked ships, who occasionally chimes in to solve odd seaborne mysteries like the myriad washed up, dead dolphins in '98, a curious lot of whom sported blast holes in their necks. Google him and you'll grant Sheridan more respect.
The first two google hits are for other people, but the third one is this. It appears to be a story about Leo leading a team of divers that claimed in the press that they found the long-lost plane wreck of Amy Johnson. However when questioned directly by officials in London, they said they hadn't actually found anything. That was in 2003 and nothing additional seems to have happened. So, while I was honestly trying to find out more about Leo, the first hit seeems to indicate he likes to make bold exagerations when speaking to the press. This isn't helping my confidence...
Nothing on the Navy page you point to dilutes Sheridan's account of a classified program. That is to say, the existence of open source and/or declassified programs and material says nothing about whether a classified program exists.
Actually it has lots of interesting tidbits, such as this: "Why have there been so many rumors about the NMMP over the years? Several decades of classification of the program's true missions of mine-hunting and swimmer defense, led to media speculation and animal activist charges of dolphins used as offensive weapons, speculation and charges that could not be countered with facts due to that classification. Additionally, fantasy is often times more interesting than reality. With declassification of the missions of the program in the early 1990s, the Navy has repeatedly and openly discussed those missions, but rumors are not easily forgotten, and there are those who continue to actively promote them."
So it was classified, but it isn't anymore. Now, they could easily be lying, but I don't know why I should trust Leo any more than the Navy. In fact I don't trust either in the absence of any proof, which argues for the Null Hypothesis.
SZ, have you ever heard of a straw man?
Have you ever heard of failure of proof of the negative is not proof of the positive? Why *must* it be true that dead dolphins killed by explosions were part of a secret Navy program gone awry? If the dead dolphins had "US NAVY" stamped on them, that would be one thing, but instead we simply had evidence that they were killed in the same unusual way. That raises the following questions:
- Why would the Navy test something in the Mediterranean unless they were sure it would work? - If the dolphins could not be allowed to fall into the wrong hands, why didn't they keep them in holding tanks during transit? (like the ones they used for the Katrina dolphin rescue) - Why use dolphins to guard a fleet in transit, when there is no risk from divers? Navy ships can travel at 30 knots, so you don't have to worry about anyone swimming up to your vessel. - Why not design the device to fall off of the dolphin into the sea rather than explode and kill the dolphin? It's not like the dolphins are going to swim up and sell their secrets to a foreign goverment. - Why couldn't these belong to a foreign government's navy? After years of information about the US program, and the relatively low budget required, why wouldn't navies test this idea? - Why use dolphins at all for offensive operations, when the Navy's published research indicates that sea lions are far better for this purpose? (Dolphins are only used for mine hunting now, security and "force protection" against divers is provided by sea lions)
Here's an alternative: Dolphins get trapped in fishing nets all the time in that sea (see here). Maybe some fishermen were bored and hated dolphins enough to kill all the ones landing in their nets with explosives. Or maybe they were part of a secret ritu
Marine biologists were baffled but Leo Sheridan proposed the only explanation that has not yet been dismissed. "I am convinced that these were dolphins trained by the US Navy and that something went badly wrong,
I am convinced it was done by aliens hiding on the Canary islands... can't dismiss that either. See, it's easy to say something is the result of some secret project: Since all the evidence you would need to prove your case is secret, of course you can't provide any evidence.
In fact it was 1989 when the U.S. Navy began its classified Cetacean Intelligence Mission.
Well, if you trot over to the program's official web page, you can see they have been training toothed whales since 1962. And once again, how does he know the specifics of something supposedly top secret (but with an official web page, of course)? Maybe it began in 1987, and it's even more advanced!
Speculation is fun, but when you do it too much and for too long you simply start seeing patterns that aren't really there. You start believing anything that fits your pattern, even when far simpler explanations fit equally well. Occam's Razor goes out the window. I wonder what Sheridan thinks of the movie A Beautiful Mind.
There have been stories about the marine mammal program before, and regular fights with animal rights groups. It is no longer classified though, so anyone can go find out plenty of information at the project's official website. You can also check out their FAQ. It pretty clearly states that dolphins and sea lions are only used for marking and tagging, and that they are not used offensively since they can't really distinguish friendly forces and foes. It seems some people still refuse to give up on speculation however.
Anyway, I seriously doubt that dolphins are being used with poison darts, since the Navy seems to prefer using sea lions now (They don't need storage pools, work better in tight areas like harbors and piers, and tolerate more varying temperatures). And even if there *were* poison dart weilding dolphins, why on earth would they be left armed while at a training facility during a storm?
In general countries do not look kindly on people in the illegal drug import/export business (in Marc's case, he sold pot seeds to the US from Canada). If you want to be an activist, you should generally do it in your own country, as things get far too murky otherwise. Tell me, how does Canada look at those who import cocaine from South America? Maybe they are just "activists supporting freedom" too?
Even someone importing/exporting only alchohol without doing the proper paperwork would get into legal trouble in most countries. I see no reason why this should be different. If you want to end a law, protest it; Don't start a business that ignores it.
Well not only that, but according to the Inquirer, it's cost per die, not cost per chip. That means that you have to add the cost of all the failed chips. You can only sell the ones that work, and yield is never that high for processors using the latest process. Also, the die price may not include packaging the chip (i.e. adding all the pins, heat spreader, etc. Seems cheap but it actually adds a noticable amount to the overall cost.
Finally, average is just a stupid metric to begin with. Intel probably sells a lot more $59 Celeron Ds than they do $1005 Pentium 4EEs. I'm sure the production cost is quite different, yet with the average metric it looks good on paper to sell more cheap chips. TFA quotes average die cost and a near maximum price, where average price would make a lot more sense. I doubt that its anywhere near $600 on average given the number of low-end OEM chips Intel sells.
Collisions resulting from false positives will be exceedingly rare when using a large cryptographic hash. However if you are really worried about that, just compare the object contents after a collision is found. 10^6 is a lot smaller than 2^256 ~= 10^77, so in other words it won't happen in our lifetimes.
Hashing is really not that advanced an approach, and is commonly applied all over the place in scalable systems. How do you think P2P systems scale the way they do? They are indexing a *lot* of content... millions of users and billions of files isn't that far off for the bigger ones.
Indexed object storage scales just fine if you're willing to pay for it. Want to scale out to a farm of 256 servers? Use the first 8 bits of the hash to index the server, and the remaining bits stored on each server.
Only time will tell whether we gave the Iraqis the gift of freedom or replaced hardship with hell (civil war). Here's to hoping it is the former, but it certainly seems like it was an unwise gamble.
Yes. I've read a lot more too. You should read it more closely.
Modern nuke design does NOT use the design you gave as your example
So what design did you mean when you said:...keep a critical mass amount of it separated... break the separator...
That's the little boy design, which is the only one that works simply, but doesn't work well with Plutonium (unless that Plutonium is more pure than that used in modern nuclear weapons).
Modern nuke design does NOT use the design you gave as your example, and is instead based around increasing the efficiency by using reflectors and high compression via conventional explosive (a la "Fat Man")
Exactly, and those are the ones which are hard to produce, thus aiding in safety. You need to machine a very accurate sphere, measure it even more accurately, and offset the explosive timings to compress it into an almost perfect sphere. If you mess up, it won't work. By work, I mean have anywhere near the expected yield, most likely it'll create a very dirty bomb with the yield of large conventional explosives.
Hell, you could take two 10 kilo chunks of quality weapons grade plutonium, put one on the ground and suspend the other above it by a rope, then sever the rope with a cherry bomb and that would most likely work.
It would have to be higher quality Plutonium than what is used in current nuclear weapons, which is to say it'd be pretty difficult to obtain. In that case you'd still need a neutron source to get any sort of yield that would justify going to the trouble of making a nuclear weapon. If however you are satisfied with simply having a meltdown that realeases a bunch of radiation, then a really simple design like dropping two hemispheres will work.
Rockets are really simple in principle, and anyone can build one. Making a missile that can hit a target accurately several hundred miles away is a different issue entirely. Nuclear weapons are also simple in principle, but its a vast oversimplification compared to actually building one.
This is important though. MySQL can only redeem itself now by issuing another press release which says that partnering with SCO "made them feel dirty."
Up until the time comes when someone has to disarm one.
By which you mean a civilian, since a military person could get access to the manuals if needed. A civilian having to disarm an armed nuclear weapon is pretty much the definition of "unlikely". Most weapons are pretty much impossible to stop once armed anyway, unless by disarm you meant dismantle. Modern bombs tend very much toward trying to be safe, since having one blow up your own people is undesirable. Thus if you want, you could dismantle one with a sledge hammer. The worst you'll face is some radiation from the materials. For more details see this.
Actually you give the perfect example of something that won't work. A gun type nuclear bomb requires U235 since Plutonium would pre-detonate (creating a very dirty but not so strong explosion). This little detail can be found on the Internet, but I expect (and hope) that many other such details are not so easily available.
You're more likely to fuck up and blow yourself up at the wrong time...
This is exactly the kind of failure we want, rather than a succesful detonation at the target location. A partial detonation is still probably big enough to be noticed, but not nearly as destuctive as a successful detonation.
The designs described by that page were state of the art in the 1950, and not detailed enough to build a bomb that will work with any certainty*. If you google around you'll find changes that have been made since then, but still many details are missing (which is good). What you don't find are the detailed designs for Russian "suitcase" nukes or the minaturized fission weapons used as triggers for modern fusion warheads. The article is a far cry from what the government has in its classified documents. The priciples and layout are there, but many details are missing. Critical measurements, information on the some of the materials, and timing delays are missing or unclear.
It's kind of like how a lot of software patents work, they give you the overview but skip the meat that would allow you to actually recreate the thing without a lot of testing. That's what nuclear safety more or less depends on now... keeping the details that are determined by testing secret. If I gave you a sketch of a 4-cycle piston engine, you could recreate one, but only after a lot of testing and several failures. Thanks to the NPT testing bans, we are now very good at noticing such tests.
(*) Of course, the "little boy" design is a concern due to its obviousness, but thankfully it requires a lot of U235 which is harder to obtain than Plutonium. I'd still rather not have the schematics released though, since there is zero benefit to society.
I for one don't want the US releasing the designs for nuclear weapons, even if they are 30 years old. I also don't care who JFK slept with, though I'm sure People magazine would want to know. There has to be some form of benefit to offset the damage.
There are, of course, a lot of things with little damage, but those tend to be of little benefit too. On the other hand, if something is bad enough (such as Nixon's watergate coverup), I expect it'll get leaked anyway. Whistleblowers have the strong motivation of fame.
Well it depends; Some WWII stuff would be fine to release after 1945, while some other things would have to wait until the end of the Cold War. The latter was of indeterminate length right up until the end, making a purely time-based declassification difficult. Reviewing something for release takes a substantial effort since you have to figure out again why something was classified in the first place, and you don't want to make a mistake.
Something that might help is to label classified information with a short list of reasons or keywords. Some examples could be: "Important to keep from Germans during the war" or "Cannot be revealed to Soviet Union". If we tended to use fairly stock phrases, we could simply evaluate those to determine when things get released. With the obvious caching effects this shouldn't be that hard.
Yeah but then you'd have to constantly divert power to the deflector shields. It'd be better if you could do it in a localized fashion, such as for only forward or rear deflector shields. However in a bind you might still have to cut power from life support, leading to undue tension between the captain and engineering.
Slashdot Mirror
Sheridan is a well respected British expert particularly on wrecked ships, who occasionally chimes in to solve odd seaborne mysteries like the myriad washed up, dead dolphins in '98, a curious lot of whom sported blast holes in their necks. Google him and you'll grant Sheridan more respect.
The first two google hits are for other people, but the third one is this. It appears to be a story about Leo leading a team of divers that claimed in the press that they found the long-lost plane wreck of Amy Johnson. However when questioned directly by officials in London, they said they hadn't actually found anything. That was in 2003 and nothing additional seems to have happened. So, while I was honestly trying to find out more about Leo, the first hit seeems to indicate he likes to make bold exagerations when speaking to the press. This isn't helping my confidence...
Nothing on the Navy page you point to dilutes Sheridan's account of a classified program. That is to say, the existence of open source and/or declassified programs and material says nothing about whether a classified program exists.
Actually it has lots of interesting tidbits, such as this: "Why have there been so many rumors about the NMMP over the years? Several decades of classification of the program's true missions of mine-hunting and swimmer defense, led to media speculation and animal activist charges of dolphins used as offensive weapons, speculation and charges that could not be countered with facts due to that classification. Additionally, fantasy is often times more interesting than reality. With declassification of the missions of the program in the early 1990s, the Navy has repeatedly and openly discussed those missions, but rumors are not easily forgotten, and there are those who continue to actively promote them."
So it was classified, but it isn't anymore. Now, they could easily be lying, but I don't know why I should trust Leo any more than the Navy. In fact I don't trust either in the absence of any proof, which argues for the Null Hypothesis.
SZ, have you ever heard of a straw man?
Have you ever heard of failure of proof of the negative is not proof of the positive? Why *must* it be true that dead dolphins killed by explosions were part of a secret Navy program gone awry? If the dead dolphins had "US NAVY" stamped on them, that would be one thing, but instead we simply had evidence that they were killed in the same unusual way. That raises the following questions:
- Why would the Navy test something in the Mediterranean unless they were sure it would work?
- If the dolphins could not be allowed to fall into the wrong hands, why didn't they keep them in holding tanks during transit? (like the ones they used for the Katrina dolphin rescue)
- Why use dolphins to guard a fleet in transit, when there is no risk from divers? Navy ships can travel at 30 knots, so you don't have to worry about anyone swimming up to your vessel.
- Why not design the device to fall off of the dolphin into the sea rather than explode and kill the dolphin? It's not like the dolphins are going to swim up and sell their secrets to a foreign goverment.
- Why couldn't these belong to a foreign government's navy? After years of information about the US program, and the relatively low budget required, why wouldn't navies test this idea?
- Why use dolphins at all for offensive operations, when the Navy's published research indicates that sea lions are far better for this purpose? (Dolphins are only used for mine hunting now, security and "force protection" against divers is provided by sea lions)
Here's an alternative: Dolphins get trapped in fishing nets all the time in that sea (see here). Maybe some fishermen were bored and hated dolphins enough to kill all the ones landing in their nets with explosives. Or maybe they were part of a secret ritu
My god, what a kook this Sheridan guy is...
Marine biologists were baffled but Leo Sheridan proposed the only explanation that has not yet been dismissed. "I am convinced that these were dolphins trained by the US Navy and that something went badly wrong,
I am convinced it was done by aliens hiding on the Canary islands... can't dismiss that either. See, it's easy to say something is the result of some secret project: Since all the evidence you would need to prove your case is secret, of course you can't provide any evidence.
In fact it was 1989 when the U.S. Navy began its classified Cetacean Intelligence Mission.
Well, if you trot over to the program's official web page, you can see they have been training toothed whales since 1962. And once again, how does he know the specifics of something supposedly top secret (but with an official web page, of course)? Maybe it began in 1987, and it's even more advanced!
Speculation is fun, but when you do it too much and for too long you simply start seeing patterns that aren't really there. You start believing anything that fits your pattern, even when far simpler explanations fit equally well. Occam's Razor goes out the window. I wonder what Sheridan thinks of the movie A Beautiful Mind.
There have been stories about the marine mammal program before, and regular fights with animal rights groups. It is no longer classified though, so anyone can go find out plenty of information at the project's official website. You can also check out their FAQ. It pretty clearly states that dolphins and sea lions are only used for marking and tagging, and that they are not used offensively since they can't really distinguish friendly forces and foes. It seems some people still refuse to give up on speculation however.
Anyway, I seriously doubt that dolphins are being used with poison darts, since the Navy seems to prefer using sea lions now (They don't need storage pools, work better in tight areas like harbors and piers, and tolerate more varying temperatures). And even if there *were* poison dart weilding dolphins, why on earth would they be left armed while at a training facility during a storm?
Does Symantec know customers who did?
Of course not, as that would be admitting their products aren't perfect.
the goal is not to compete head to head with the proposed Adobe/Macromedia merger but to turn developers into designers.
And if there's anything the world needed, its more programmer art.
Maybe then I could go a day without the constant prattle about how Windows is so horrible...
Then stop emailing us Word documents and complaining when you can't play standard video formats in Media Player.
No kidding... the GP looks like it was written by one of those random text generators.
In general countries do not look kindly on people in the illegal drug import/export business (in Marc's case, he sold pot seeds to the US from Canada). If you want to be an activist, you should generally do it in your own country, as things get far too murky otherwise. Tell me, how does Canada look at those who import cocaine from South America? Maybe they are just "activists supporting freedom" too?
Even someone importing/exporting only alchohol without doing the proper paperwork would get into legal trouble in most countries. I see no reason why this should be different. If you want to end a law, protest it; Don't start a business that ignores it.
Well not only that, but according to the Inquirer, it's cost per die, not cost per chip. That means that you have to add the cost of all the failed chips. You can only sell the ones that work, and yield is never that high for processors using the latest process. Also, the die price may not include packaging the chip (i.e. adding all the pins, heat spreader, etc. Seems cheap but it actually adds a noticable amount to the overall cost.
Finally, average is just a stupid metric to begin with. Intel probably sells a lot more $59 Celeron Ds than they do $1005 Pentium 4EEs. I'm sure the production cost is quite different, yet with the average metric it looks good on paper to sell more cheap chips. TFA quotes average die cost and a near maximum price, where average price would make a lot more sense. I doubt that its anywhere near $600 on average given the number of low-end OEM chips Intel sells.
I'd have had a TIVO years ago, but they insist on charging me for the box AND charging me a monthly fee. Sorry, no. Pick one.
Why didn't you buy a Tivo with the lifetime subscription then? I'm not sure you can buy that anymore, but it was offered for quite some time.
Collisions resulting from false positives will be exceedingly rare when using a large cryptographic hash. However if you are really worried about that, just compare the object contents after a collision is found. 10^6 is a lot smaller than 2^256 ~= 10^77, so in other words it won't happen in our lifetimes.
Hashing is really not that advanced an approach, and is commonly applied all over the place in scalable systems. How do you think P2P systems scale the way they do? They are indexing a *lot* of content... millions of users and billions of files isn't that far off for the bigger ones.
Indexed object storage scales just fine if you're willing to pay for it. Want to scale out to a farm of 256 servers? Use the first 8 bits of the hash to index the server, and the remaining bits stored on each server.
Only time will tell whether we gave the Iraqis the gift of freedom or replaced hardship with hell (civil war). Here's to hoping it is the former, but it certainly seems like it was an unwise gamble.
Except that searching the entire mail system for duplicate msgids to hard link to every time a message is received is not exactly a scalable solution!
You can do this efficiently by naming objects by their content hash, as in git.
Sure, but was the "new" Afghan government elected? I feel the elections made a big difference in terms of legitamacy.
Did you even read what you cited?
...keep a critical mass amount of it separated ... break the separator...
Yes. I've read a lot more too. You should read it more closely.
Modern nuke design does NOT use the design you gave as your example
So what design did you mean when you said:
That's the little boy design, which is the only one that works simply, but doesn't work well with Plutonium (unless that Plutonium is more pure than that used in modern nuclear weapons).
Modern nuke design does NOT use the design you gave as your example, and is instead based around increasing the efficiency by using reflectors and high compression via conventional explosive (a la "Fat Man")
Exactly, and those are the ones which are hard to produce, thus aiding in safety. You need to machine a very accurate sphere, measure it even more accurately, and offset the explosive timings to compress it into an almost perfect sphere. If you mess up, it won't work. By work, I mean have anywhere near the expected yield, most likely it'll create a very dirty bomb with the yield of large conventional explosives.
Hell, you could take two 10 kilo chunks of quality weapons grade plutonium, put one on the ground and suspend the other above it by a rope, then sever the rope with a cherry bomb and that would most likely work.
It would have to be higher quality Plutonium than what is used in current nuclear weapons, which is to say it'd be pretty difficult to obtain. In that case you'd still need a neutron source to get any sort of yield that would justify going to the trouble of making a nuclear weapon. If however you are satisfied with simply having a meltdown that realeases a bunch of radiation, then a really simple design like dropping two hemispheres will work.
Rockets are really simple in principle, and anyone can build one. Making a missile that can hit a target accurately several hundred miles away is a different issue entirely. Nuclear weapons are also simple in principle, but its a vast oversimplification compared to actually building one.
This is important though. MySQL can only redeem itself now by issuing another press release which says that partnering with SCO "made them feel dirty."
Up until the time comes when someone has to disarm one.
By which you mean a civilian, since a military person could get access to the manuals if needed. A civilian having to disarm an armed nuclear weapon is pretty much the definition of "unlikely". Most weapons are pretty much impossible to stop once armed anyway, unless by disarm you meant dismantle. Modern bombs tend very much toward trying to be safe, since having one blow up your own people is undesirable. Thus if you want, you could dismantle one with a sledge hammer. The worst you'll face is some radiation from the materials. For more details see this.
Actually you give the perfect example of something that won't work. A gun type nuclear bomb requires U235 since Plutonium would pre-detonate (creating a very dirty but not so strong explosion). This little detail can be found on the Internet, but I expect (and hope) that many other such details are not so easily available.
You're more likely to fuck up and blow yourself up at the wrong time...
This is exactly the kind of failure we want, rather than a succesful detonation at the target location. A partial detonation is still probably big enough to be noticed, but not nearly as destuctive as a successful detonation.
The designs described by that page were state of the art in the 1950, and not detailed enough to build a bomb that will work with any certainty*. If you google around you'll find changes that have been made since then, but still many details are missing (which is good). What you don't find are the detailed designs for Russian "suitcase" nukes or the minaturized fission weapons used as triggers for modern fusion warheads. The article is a far cry from what the government has in its classified documents. The priciples and layout are there, but many details are missing. Critical measurements, information on the some of the materials, and timing delays are missing or unclear.
It's kind of like how a lot of software patents work, they give you the overview but skip the meat that would allow you to actually recreate the thing without a lot of testing. That's what nuclear safety more or less depends on now... keeping the details that are determined by testing secret. If I gave you a sketch of a 4-cycle piston engine, you could recreate one, but only after a lot of testing and several failures. Thanks to the NPT testing bans, we are now very good at noticing such tests.
(*) Of course, the "little boy" design is a concern due to its obviousness, but thankfully it requires a lot of U235 which is harder to obtain than Plutonium. I'd still rather not have the schematics released though, since there is zero benefit to society.
I for one don't want the US releasing the designs for nuclear weapons, even if they are 30 years old. I also don't care who JFK slept with, though I'm sure People magazine would want to know. There has to be some form of benefit to offset the damage.
There are, of course, a lot of things with little damage, but those tend to be of little benefit too. On the other hand, if something is bad enough (such as Nixon's watergate coverup), I expect it'll get leaked anyway. Whistleblowers have the strong motivation of fame.
Well it depends; Some WWII stuff would be fine to release after 1945, while some other things would have to wait until the end of the Cold War. The latter was of indeterminate length right up until the end, making a purely time-based declassification difficult. Reviewing something for release takes a substantial effort since you have to figure out again why something was classified in the first place, and you don't want to make a mistake.
Something that might help is to label classified information with a short list of reasons or keywords. Some examples could be: "Important to keep from Germans during the war" or "Cannot be revealed to Soviet Union". If we tended to use fairly stock phrases, we could simply evaluate those to determine when things get released. With the obvious caching effects this shouldn't be that hard.
Wow, that's a great idea. Quite helpful to have the fire seasons out of sync.
I thought it was DOA.
Yeah but then you'd have to constantly divert power to the deflector shields. It'd be better if you could do it in a localized fashion, such as for only forward or rear deflector shields. However in a bind you might still have to cut power from life support, leading to undue tension between the captain and engineering.
In other words, people who will run into most current automatic doors anyway?