A similar idea is to use a hash tree (Merkle tree) and build a hash of small blocks of data. That way, you can see whether multiple files share common blocks of data and also automatically implement P2P or caches for transfers of large files with automatic hashing of the entire file (for ease of referencing) and individual blocks during a transfer. Gnutella uses Tiger trees for this, but it would be nice if there was an official standard URI format for hash trees so that Google and other search engines could index those as well.
They used to release the full source code to their Java applet that handled encryption/decryption, and provided instructions for building a byte-exact replica of what they distribute.
Theoretically, hushmail can be used in a perfectly secure manner; download the source, check it for back-doors, compile the applet yourself and memorize its hash. Then whenever you use hushmail, just verify that the hash of the downloaded applet is the same as the one you compiled yourself.
Probably hushmail was just feeding a tainted applet to the specific targets of the investigation, otherwise I'm sure some other astute user would have noticed the change in the applet signature. The typical muscle-bound steroid dealer probably doesn't have the time to memorize and compare hashes though...
The radio waves from your cordless or callphone are probably being transmitted across half the neighbourhood, does that mean you have no problem with the government eavesdropping on them ?
Radios that receive in the cellular and portable phone spectrum are not illegal, that should tell you something. Encrypt your communications if you want them to be private. Use an envelope if you want your letters to be private. Put your secret documents in a safe if you don't want your neighbors to see them on the coffee table and read them when they're visiting. Use some common sense.
Don't use them. Exercise some fucking restraint. It shouldn't take legislation for you to respect other people's property and privacy.
It's my property as much as theirs, that's the way the airwaves work and why the FCC even exists. I have no trouble *sharing* with other people, but when their insecure methods of communication impedes my freedom to listen to my own airwaves, then it's gone beyond common sense. It should be everyone's duty to secure their own communications, not to worry about accidentally hearing someone else's.
Why would putting a server up on port 80 be considered public anymore than putting up a wireless access point? I don't see how having a web server is "implied public". Just because I put it there doesn't mean I want everyone to access it. That's a poor example to use.
The Internet is a public network. If you want a private network, you can buy private connections from your ISP that don't have routes to the Internet. Duh.
Yeah this just happens to be a prison but how are you going to feel when someone releases the schematics to the air conditioning system at jrandom fort in your town and proceeds to gas and entire base of people?
If all you need is a diagram of air ducts in order to gas an entire base, then security at that base was already in a total state of failure. Just prevent access to the goddamn ducts in the first place!
In the days of on-demand satellite imagery do you really think knowing where forces are in Iraq is such a big deal? The insurgents already know the Humvees and soldiers lack the proper equipment (armor) because the news media widely reports it. The U.S. military is *in* Iraq, so how secret do you suppose their location and general equipment status really is? In any urban environment I think you have to assume that pretty much all your factual secrets (positions, armaments, etc.) are compromised by civilian infiltrators and spies.
It means nothing more than that. If the government can continue to do as it pleases, national security is not at risk. If the people were to revolt against the U.S.'s foreign entanglements that eventually cause terrorism against the U.S., it might inconvenience the elected leaders and the people pulling their strings. There's plenty of money to be made in wars, and the war profiteers aren't even trying to hide it anymore (see Blackwater and Halliburton).
Heat transfer to the same environment depends directly on surface area, assuming bodies of the same temperature. Until hypothermia sets in, skin temperature is probably going to be pretty stable across individuals.
Don't forget that nerds and geeks often have much more surface area than the standard professor, leading to higher heat loss into the cooler server room.
What are you planning to do for network access? If there's cell phone coverage, it would be neat if the OLPC supported bluetooth/USB connections so the cellphones could be turned into "edge routers" for the mesh. It seems like it's either that or an expensive land line/satellite/wireless link out of the village...
Interestingly, the much-reviewed TrueCrypt engine seems to slow to a crawl if you create a bunch of files (and therefore keys) in a hurry - presumably it has an RNG that actually blocks waiting until it has enough new "really random" bits for each new key. This is a cool idea for a crypto library, but not usable for a general-purpose RNG, which suggests that the system libraries should probably provide *two* RNGs.
Older versions of truecrypt used CBC with IVs generated directly from the sector number, without sector-specific keys or IVs. Newer versions use LRW which also does not use sector specific keys and derives the IV from the sector number in a secure fashion. The entire volume is encrypted with a single key (or two with LRW, and only if you're not cascading ciphers), so I would look elsewhere for your performance issue. It may just be Windows sucking at creating new files, since the file-system layer is independent of TrueCrypt and just uses it as a raw device.
Not to mention the Iraq and Afghan wars were a DIRECT RESULT of the 9/11 attacks. Those countries would not have been invaded if we had not had planes hijacked and hitting military buildings and skyscrapers.
Afghanistan; probably not. I doubt Bush could find it on a map (much less pronounce it) before September 11, 2001.
Iraq, hell yeah. They have oil, and they tried to kill his daddy. Can you see *any* direct link between 9/11 and Iraq? Iraq supposedly had WMD, but that has nothing to do with Islamic terrorism, and they had those before 9/11. Iraq gassed the Kurds before 9/11. Iraq was ignoring foreign policy and the UN and shooting at our planes before 9/11. Their behavior *did* *not* *change* as a result of 9/11. If anything, the war in Afghanistan was used as an excuse for going into Iraq, but I see no direct causal link between 9/11 and the invasion of Iraq. If anything, Iraq was invaded because Afghanistan had already been invaded, and the war was going rather poorly. That's a secondary effect and almost entirely due to the failure in Afghanistan (and the ease of getting the public to agree to war on one more country instead of having to start it from scratch in Iraq). We still haven't caught Bin Laden (although we killed Saddam). What does that tell you about the relationship between 9/11 and the war in Iraq?
9/11 had a resonant impact on our society, and we're still dealing with the effects of it today. So the 9/11 attacks bother me the most, and they should bother you the most as well.
Only because you let it impact you. Three buildings fell down and ~3000 people didn't have to die of cancer, heart disease, or car accidents. Whoop-di-do; collect the insurance money and rebuild. That's the American way, not whining for years on end.
The only effects of 9/11 I'm still dealing with are the stupid airline searches that are less than 50% effective at detecting weapons, the reduction in human and civil rights that affect ~300,000,000 people in the U.S. and countless more in other countries, and having to listen to whiners like you who can't get it through your head that the 9/11 "terrorist" attacks were frankly indistinguishable from murder/arson. The murderers/arsonists are dead, and the people who trained them are dead or on the run at this point. What more can you possibly want besides a fascist police state?
Just curious...how are they going to know you own a gun? Do some states require you to register them or something?
Good question.
I bought all my guns used...cash. The only time I ever had to register any of them, when when I listed them on a concealed carry permit. I didn't know some states required you to register your guns...otherwise, how would they know you owned them?
Every time these "connect desktops to become the fastest computer in the world" articles come up, I have to dust off my Cluster Urban Legends article to clear up the mis-conceptions that abound. I also did a piece on the Linux Magazine site as well that debunks much of the spam-bot supercomputer legend (need to register for that one)
Too bad you're wrong in this case, since protein folding is embarrassingly parallel. How do you think Folding@home works?
Nor does killing a bunch of people including yourself count as a good move, evolutionarily speaking.
My guess is that you're not a direct descendant of a Neanderthal or any number of the other humanoids pushed to extinction by their other humanoid neighbors. Evolution doesn't give a shit about how individuals die; all that matters is that the survivors continue to breed.
You're stating as a matter of fact that Sun "used the product from the patent". This is stretching the truth somewhat. The actual facts of the matter are that NetApp claims Sun have violated their patent (WAFL, etc), and filed suit requesting relief. Sun however disagree and believe they do not violate NetApps' patents - indeed Sun claim, in their counter-suits, that NetApp are violating Suns' patents. However, no-one is violating anyone's patents until either both parties agree they are, or a judge says so.
And if anyone thinks WAFL is "special", it's not. It's basically just the concept of never writing twice what only needed to be written once while keeping complete historical copies (or versions).
If you write a file, you have to store all the data in blocks in the filesystem somewhere. You have to have a structure (metadata) pointing to where those blocks live. If you want to append to the file, you write more blocks and append to the metadata. If you want to overwrite blocks on a file, you write some new blocks and adjust the metadata for the new version of the file to point to the updated data blocks while retaining pointers to the unchanged blocks. As far as I can tell from talking to NetApp people and looking at the patents, that's all WAFL is really about. They may have some random patents on writing new blocks to free space and reorganizing old blocks and their specific use of Reed Solomon coding for data redundancy (basically RAID5/6 without striping which allows arbitrary addition of data or parity disks without rebuilding the entire array), too, but ultimately I don't know of anything truly innovative they've done.
Correct me if I'm wrong, because I'd love to find some actual new ideas in filesystem design that haven't been talked about or implemented for the last 20 or 30 years.
I think the idea of micropayments for individual tasks is an interesting one. The biggest flaw in the system I see is that there is a wide variety of skill levels, worker to worker, which will effect the results you get, and I don't see any way of ensuring that you get someone whose work you can trust.
That's a problem in general. The obvious answer is to use probable redundancy; pay five or ten people to do the same task and pick the best result. To prevent worker collusion (e.g. it's far easier for one worker to perform all ten of your work units) the market (Mechanical Turk, in this instance) needs a way to verify the uniqueness of work results. Ultimately that's impossible without randomly assigning work units, since if workers can always choose which jobs to work on they can easily collude or one person can do everything with multiple accounts. If Amazon had a separate category for slightly higher paying randomized workloads, that would be an efficient solution. People wouldn't be able to pick tasks they liked, but they would get a little more money for the random ones and the employer would get multiple results of which the best (or agreeing) could be chosen.
a) would be physically seperated so you'd have to perform some correlation first to remove the arbitrary time delays from the audio source to the phones and this would remove some of the resolution
Couldn't you synchronize the samples using the phase of a common high amplitude frequency component? Say the first drum beat of the song, which would provide a good synchronization point as well as phase information.
Having submitted several patents through my company, I can attest that they need to be rewritten all the time because patent reviewers are idiots. They take a one sentence claim, pick a two 'big' words out it and do a literature search. If those two words appear in any publication remotely related to the field related to your patent, they mark it as prior art.
Hence the patent attempts for "rotational conveyance devices", whatever those are...
a good leader can go into a room, or on tv, or in front of a news reporter, and say something in such a way that the majority of people reading or hearing what he says find it compelling
That's a function of oratory skill and charisma and plenty of non-leaders are very good at it. They're called actors. You are simply fiddling with the definition of leadership, which is ultimately the ability to get other people to do what the leader wants. That can be done by charismatic speeches, religious fervor, rational arguments, or secret police and gulags.
Perhaps you are using the adjective "good" in an ill-defined way and expecting me to understand what you're thinking. By "good" do you mean morally good, good at public speaking, good at rational arguments, good at religious words, or good at running the secret police? I'm guessing you mean good at public speaking, which is a tiny part of what being a leader is about. A leader's actions are just as important as words, if not more. What would Ghandi have been without non-violent resistance? Would Churchill have been remembered if Britain had been invaded and destroyed? Would Lincoln have a memorial if the South had won?
I'm not going to argue that charisma is useless to leaders; it's very useful but ultimately must be tied to some other substantial leadership skill. Otherwise you get empty figureheads with absolutely no purpose or direction to their actions. I do not think charisma is necessary for good leadership; all that is necessary is respect and trust for the leader, or the despotic alternative of the secret police to keep everyone in line.
Leaders are generic. What kind of leader do you want? Someone who can make good 15 second sound bytes, or one who understands problems and figures out how to fix them and also knows when to leave well enough alone? Exactly how low are your standards, anyway? You're electing the figurehead and executive director of an entire country and you're willing to go with "Well, he does have some leadership ability"? A good leader understands what's going on. I think Bush twiddles his thumbs in the Oval Office waiting for an adviser to make a decision for him or hand him a paper to sign. He's extraneous. Why bother with elections in that case?
Slashdot Mirror
A similar idea is to use a hash tree (Merkle tree) and build a hash of small blocks of data. That way, you can see whether multiple files share common blocks of data and also automatically implement P2P or caches for transfers of large files with automatic hashing of the entire file (for ease of referencing) and individual blocks during a transfer. Gnutella uses Tiger trees for this, but it would be nice if there was an official standard URI format for hash trees so that Google and other search engines could index those as well.
They used to release the full source code to their Java applet that handled encryption/decryption, and provided instructions for building a byte-exact replica of what they distribute.
Theoretically, hushmail can be used in a perfectly secure manner; download the source, check it for back-doors, compile the applet yourself and memorize its hash. Then whenever you use hushmail, just verify that the hash of the downloaded applet is the same as the one you compiled yourself.
Probably hushmail was just feeding a tainted applet to the specific targets of the investigation, otherwise I'm sure some other astute user would have noticed the change in the applet signature. The typical muscle-bound steroid dealer probably doesn't have the time to memorize and compare hashes though...
The radio waves from your cordless or callphone are probably being transmitted across half the neighbourhood, does that mean you have no problem with the government eavesdropping on them ?
Radios that receive in the cellular and portable phone spectrum are not illegal, that should tell you something. Encrypt your communications if you want them to be private. Use an envelope if you want your letters to be private. Put your secret documents in a safe if you don't want your neighbors to see them on the coffee table and read them when they're visiting. Use some common sense.
Don't use them. Exercise some fucking restraint. It shouldn't take legislation for you to respect other people's property and privacy.
It's my property as much as theirs, that's the way the airwaves work and why the FCC even exists. I have no trouble *sharing* with other people, but when their insecure methods of communication impedes my freedom to listen to my own airwaves, then it's gone beyond common sense. It should be everyone's duty to secure their own communications, not to worry about accidentally hearing someone else's.
Wifi spectrum is unregulated, not happy touchy feely.
That's the fucking point; unregulated means my laptop can talk to your access point and you can't do a thing about it except ignore my packets.
Why would putting a server up on port 80 be considered public anymore than putting up a wireless access point? I don't see how having a web server is "implied public". Just because I put it there doesn't mean I want everyone to access it. That's a poor example to use.
The Internet is a public network. If you want a private network, you can buy private connections from your ISP that don't have routes to the Internet. Duh.
Yeah this just happens to be a prison but how are you going to feel when someone releases the schematics to the air conditioning system at jrandom fort in your town and proceeds to gas and entire base of people?
If all you need is a diagram of air ducts in order to gas an entire base, then security at that base was already in a total state of failure. Just prevent access to the goddamn ducts in the first place!
In the days of on-demand satellite imagery do you really think knowing where forces are in Iraq is such a big deal? The insurgents already know the Humvees and soldiers lack the proper equipment (armor) because the news media widely reports it. The U.S. military is *in* Iraq, so how secret do you suppose their location and general equipment status really is? In any urban environment I think you have to assume that pretty much all your factual secrets (positions, armaments, etc.) are compromised by civilian infiltrators and spies.
It means nothing more than that. If the government can continue to do as it pleases, national security is not at risk. If the people were to revolt against the U.S.'s foreign entanglements that eventually cause terrorism against the U.S., it might inconvenience the elected leaders and the people pulling their strings. There's plenty of money to be made in wars, and the war profiteers aren't even trying to hide it anymore (see Blackwater and Halliburton).
Yeah, now the terrorists know exactly where all their prisoners of war (errrr "illegal combatants") are and can go rescue them!
Heat transfer to the same environment depends directly on surface area, assuming bodies of the same temperature. Until hypothermia sets in, skin temperature is probably going to be pretty stable across individuals.
Don't forget that nerds and geeks often have much more surface area than the standard professor, leading to higher heat loss into the cooler server room.
What are you planning to do for network access? If there's cell phone coverage, it would be neat if the OLPC supported bluetooth/USB connections so the cellphones could be turned into "edge routers" for the mesh. It seems like it's either that or an expensive land line/satellite/wireless link out of the village...
Interestingly, the much-reviewed TrueCrypt engine seems to slow to a crawl if you create a bunch of files (and therefore keys) in a hurry - presumably it has an RNG that actually blocks waiting until it has enough new "really random" bits for each new key. This is a cool idea for a crypto library, but not usable for a general-purpose RNG, which suggests that the system libraries should probably provide *two* RNGs.
Older versions of truecrypt used CBC with IVs generated directly from the sector number, without sector-specific keys or IVs. Newer versions use LRW which also does not use sector specific keys and derives the IV from the sector number in a secure fashion. The entire volume is encrypted with a single key (or two with LRW, and only if you're not cascading ciphers), so I would look elsewhere for your performance issue. It may just be Windows sucking at creating new files, since the file-system layer is independent of TrueCrypt and just uses it as a raw device.
Not to mention the Iraq and Afghan wars were a DIRECT RESULT of the 9/11 attacks. Those countries would not have been invaded if we had not had planes hijacked and hitting military buildings and skyscrapers.
Afghanistan; probably not. I doubt Bush could find it on a map (much less pronounce it) before September 11, 2001.
Iraq, hell yeah. They have oil, and they tried to kill his daddy. Can you see *any* direct link between 9/11 and Iraq? Iraq supposedly had WMD, but that has nothing to do with Islamic terrorism, and they had those before 9/11. Iraq gassed the Kurds before 9/11. Iraq was ignoring foreign policy and the UN and shooting at our planes before 9/11. Their behavior *did* *not* *change* as a result of 9/11. If anything, the war in Afghanistan was used as an excuse for going into Iraq, but I see no direct causal link between 9/11 and the invasion of Iraq. If anything, Iraq was invaded because Afghanistan had already been invaded, and the war was going rather poorly. That's a secondary effect and almost entirely due to the failure in Afghanistan (and the ease of getting the public to agree to war on one more country instead of having to start it from scratch in Iraq). We still haven't caught Bin Laden (although we killed Saddam). What does that tell you about the relationship between 9/11 and the war in Iraq?
9/11 had a resonant impact on our society, and we're still dealing with the effects of it today. So the 9/11 attacks bother me the most, and they should bother you the most as well.
Only because you let it impact you. Three buildings fell down and ~3000 people didn't have to die of cancer, heart disease, or car accidents. Whoop-di-do; collect the insurance money and rebuild. That's the American way, not whining for years on end.
The only effects of 9/11 I'm still dealing with are the stupid airline searches that are less than 50% effective at detecting weapons, the reduction in human and civil rights that affect ~300,000,000 people in the U.S. and countless more in other countries, and having to listen to whiners like you who can't get it through your head that the 9/11 "terrorist" attacks were frankly indistinguishable from murder/arson. The murderers/arsonists are dead, and the people who trained them are dead or on the run at this point. What more can you possibly want besides a fascist police state?
Just curious...how are they going to know you own a gun? Do some states require you to register them or something?
Good question.
I bought all my guns used...cash. The only time I ever had to register any of them, when when I listed them on a concealed carry permit. I didn't know some states required you to register your guns...otherwise, how would they know you owned them?
Well, NOW they know...
Every time these "connect desktops to become the fastest computer in the world" articles come up, I have to dust off my Cluster Urban Legends article to clear up the mis-conceptions that abound. I also did a piece on the Linux Magazine site as well that debunks much of the spam-bot supercomputer legend (need to register for that one)
Too bad you're wrong in this case, since protein folding is embarrassingly parallel. How do you think Folding@home works?
Nor does killing a bunch of people including yourself count as a good move, evolutionarily speaking.
My guess is that you're not a direct descendant of a Neanderthal or any number of the other humanoids pushed to extinction by their other humanoid neighbors. Evolution doesn't give a shit about how individuals die; all that matters is that the survivors continue to breed.
You're stating as a matter of fact that Sun "used the product from the patent". This is stretching the truth somewhat. The actual facts of the matter are that NetApp claims Sun have violated their patent (WAFL, etc), and filed suit requesting relief. Sun however disagree and believe they do not violate NetApps' patents - indeed Sun claim, in their counter-suits, that NetApp are violating Suns' patents. However, no-one is violating anyone's patents until either both parties agree they are, or a judge says so.
And if anyone thinks WAFL is "special", it's not. It's basically just the concept of never writing twice what only needed to be written once while keeping complete historical copies (or versions).
If you write a file, you have to store all the data in blocks in the filesystem somewhere. You have to have a structure (metadata) pointing to where those blocks live. If you want to append to the file, you write more blocks and append to the metadata. If you want to overwrite blocks on a file, you write some new blocks and adjust the metadata for the new version of the file to point to the updated data blocks while retaining pointers to the unchanged blocks. As far as I can tell from talking to NetApp people and looking at the patents, that's all WAFL is really about. They may have some random patents on writing new blocks to free space and reorganizing old blocks and their specific use of Reed Solomon coding for data redundancy (basically RAID5/6 without striping which allows arbitrary addition of data or parity disks without rebuilding the entire array), too, but ultimately I don't know of anything truly innovative they've done.
Correct me if I'm wrong, because I'd love to find some actual new ideas in filesystem design that haven't been talked about or implemented for the last 20 or 30 years.
I think the idea of micropayments for individual tasks is an interesting one. The biggest flaw in the system I see is that there is a wide variety of skill levels, worker to worker, which will effect the results you get, and I don't see any way of ensuring that you get someone whose work you can trust.
That's a problem in general. The obvious answer is to use probable redundancy; pay five or ten people to do the same task and pick the best result. To prevent worker collusion (e.g. it's far easier for one worker to perform all ten of your work units) the market (Mechanical Turk, in this instance) needs a way to verify the uniqueness of work results. Ultimately that's impossible without randomly assigning work units, since if workers can always choose which jobs to work on they can easily collude or one person can do everything with multiple accounts. If Amazon had a separate category for slightly higher paying randomized workloads, that would be an efficient solution. People wouldn't be able to pick tasks they liked, but they would get a little more money for the random ones and the employer would get multiple results of which the best (or agreeing) could be chosen.
a) would be physically seperated so you'd have to perform some correlation first to remove the arbitrary time delays from the audio source to the phones and this would remove some of the resolution
Couldn't you synchronize the samples using the phase of a common high amplitude frequency component? Say the first drum beat of the song, which would provide a good synchronization point as well as phase information.
Having submitted several patents through my company, I can attest that they need to be rewritten all the time because patent reviewers are idiots. They take a one sentence claim, pick a two 'big' words out it and do a literature search. If those two words appear in any publication remotely related to the field related to your patent, they mark it as prior art.
Hence the patent attempts for "rotational conveyance devices", whatever those are...
a good leader can go into a room, or on tv, or in front of a news reporter, and say something in such a way that the majority of people reading or hearing what he says find it compelling
That's a function of oratory skill and charisma and plenty of non-leaders are very good at it. They're called actors. You are simply fiddling with the definition of leadership, which is ultimately the ability to get other people to do what the leader wants. That can be done by charismatic speeches, religious fervor, rational arguments, or secret police and gulags.
Perhaps you are using the adjective "good" in an ill-defined way and expecting me to understand what you're thinking. By "good" do you mean morally good, good at public speaking, good at rational arguments, good at religious words, or good at running the secret police? I'm guessing you mean good at public speaking, which is a tiny part of what being a leader is about. A leader's actions are just as important as words, if not more. What would Ghandi have been without non-violent resistance? Would Churchill have been remembered if Britain had been invaded and destroyed? Would Lincoln have a memorial if the South had won?
I'm not going to argue that charisma is useless to leaders; it's very useful but ultimately must be tied to some other substantial leadership skill. Otherwise you get empty figureheads with absolutely no purpose or direction to their actions. I do not think charisma is necessary for good leadership; all that is necessary is respect and trust for the leader, or the despotic alternative of the secret police to keep everyone in line.
Get over it, whining like a bitch about why you think you were right doesn't make you right.
You must be new here.
Leaders are generic. What kind of leader do you want? Someone who can make good 15 second sound bytes, or one who understands problems and figures out how to fix them and also knows when to leave well enough alone? Exactly how low are your standards, anyway? You're electing the figurehead and executive director of an entire country and you're willing to go with "Well, he does have some leadership ability"? A good leader understands what's going on. I think Bush twiddles his thumbs in the Oval Office waiting for an adviser to make a decision for him or hand him a paper to sign. He's extraneous. Why bother with elections in that case?