Right. They would not solve the problem against motivated attackers; however, ordinary drive-by sniffers would not be able to hear much with their plain 802.11b cards.
Don't forget that at 500km distance, due to Earth's curvature you need to put the end points quite high above the ground; otherwise, they won't see each other beyond the horizon.
The point about lack of incentive on the provider's part is what worries me. What if the current 56kbps modem links up to the 1M cable links are sufficient for most people's networking needs? Then there will never be enough demand to justify upgrading the current last-mile infrastructure. Is there a cool new killer-app which would encourage many people to pay for more bandwidth? (More bandwidth here means substantially more, as in gigabits per second over fiber).
Speaking of high resolution monitors, check out this new IBM monitor: http://www.provantage.com/YIBML011.HTM. It's a 22" monster with a resolution of 3840 x 2400!!! That's over 200dpi. At $8k, it's a bargain =)
Wait a second. Her first cat passed away in 2000, and the second one just two years later? Maybe this person thinks the cats are also constructed of Lego bricks, so they don't need feeding, and can be locked up in Lego structures for extraordinary amounts of time:)
My Sinclair was also similarly elegant, not like these complicated, bloated computers we have today. Here is what I liked about it:
- You always knew what was going on inside the computer, because it could only run one program at a time. There was no multitasking in the operating system.
- Programs tended to be small and compact by design-- design of the memory architecture. There were only 16 kilobytes of RAM available so that developers were always on their toes to keep only the essential features in.
- The layout of the file system was based on the extremely elegant, yet powerful design proposed by Alan Turing-- a magnetic tape. Thus there were no complicated directory hierarchies, just individual programs laid out sequentially.
- The user base was very knowledgeable and helpful. Need to solder a memory upgrade onto the motherboard? No problem, there was always a guy in the local radio amateur club willing to help.
I wish today's machines were more like this! I suppose I can find some consolation in the fact that most microcontrollers found in an average PC have the features of the good old Sinclair.
Some x86 motherboards such as this Supermicro board can take up to 16GB of RAM. However, it's not the total amount of RAM that is the problem; rather, the address space of any single process is limited to 4GB because of 32-bit addressing. So in order to use more than 4GB in a single process, you need to use a 64bit architecture. (OK, maybe you don't *need* to, but it is a heck of a lot easier than custom hacks).
Scenario: company Foo making app Bar figures out they cannot survive by selling free software
Slashdot: The great people at Foo, makers of Bar, are going to have to close their doors due to lack of $$$. Remember Bar? Nice app. Too bad, really. Yawn. Allright, where's that new DVD I ordered?
I've been using the expression "just google for X" when referring friends to some site about X that I know shows up high on the hit list. It's funny, google is starting to replace DNS for me-- instead of remembering or bookmarking URLs, I just remember the keyword to google for. For example, the URL http://www.uk.research.att.com/vnc/ is harder to remeber than to just type "vnc" and hit "I'm feeling lucky."
I often wonder how much less productive I would be if google went away tomorrow:) If anyone from google.com is reading this, thank you!
Sure, the story was posted here. However, this is a case where the full disclosure argument starts to apply to scientific breakthroughs-- which come from a community fairly disjoint from the community of developers and engineers. Arguing not to disclose scientific achievements is going to be a hard case indeed =)
I would argue that there *is* a maximum version number-- let's call it v100 to denote 100% security. By "security" I necessarily mean technical security from remote attacks, ignoring various shortcuts using social engineering, physical access, etc. A secure protocol based on public key crypto without bugs would be 99% secure, to account for the remote possibility of finding efficient ways to reconstruct private keys.
Now take an implementation with (presumably) many bugs, such as OpenSSH, that is, say, 50% secure. The sooner the trivial engineering bugs can be fixed, the sooner the implementation will be 99% secure, at which point the only hope of compromising its security is in some unpredictable breakthrough in mathematics. This is why the armor-piercing bullet arms race argument doesn't make sense to me. Sooner or later it is bound to reach a limit which can only be surpassed with revolutionary advances.
I don't think full disclosure is really important in this process; instead, open source is what makes it possible to find bugs early. In fact, open source implies full disclosure-- without bugtraq or whatever, you can imagine downloading each new version of SSH and looking at diffs to deduce what bugs got fixed.
I second that. I remember examining a beefy Dell Precision or something similar, and to my great surprise observing a bunch of copper pipes behind the cooling fan grille on the side. Literally, the kind of copper pipes one finds in the back of a refrigerator. My guess is that there was probably some kind of oil inside, not water, in case it leaks out all over the inside of the notebook.
Unfortunately for whomever made up the original buyout hoax, this announcement hit the air one day too early, preventing the culprit from making a killing on RHAT stock.
And the corollary: always make sure your hoaxes air at the beginning of the week, without any pesky holidays in range!
Wait a second. There is no transmission involved in obtaining geographic coordinates from GPS-- it's based strictly on observation. Your statement that merely using GPS for this purpose can reveal your location to an eavesdropper is incorrect.
It would be possible if the eavesdropper could capture the message you send out, which would presumably contain the observed GPS coordinates. However, you explicitly ruled this case out.
The reason I don't buy your argument is that this is a very decentralized idea. For example, why couldn't Slashdot include lat/long metadata along its users' comments, and let you show discussions in a certain region? I think the likelihood of getting legal trouble is as great as with the curent, possibly incriminating anonymous posts.
A greater problem with this is that it's very hard to prevent posters from spoofing their geographic location. Thus, any fool with an Internet connection could appear to be posting from the local Starbucks, which would dillute the value of the comments similar to spam in usenet newsgroups. I wonder if some sort of moderation would be useful:)
If their point is that they can better reconstruct the original image from non-uniform samples, then it would have been more interesting to see a comparison of reconstructions: in particular, take one MRI images with random pixels missing, and the second MRI images with the same number of missing pixels but arranged in a regular pattern, such as a grid.
However, I suspect their point is that they can reconstruct the original at all with non-uniform sampling. This is useful in cases when it is not feasible to obtain fixed samples.
I haven't had a chance to use the Mac version. I've seen the friend's OSX version-- it looks nice.
I made the comment about stability because I've seen 1. x86 Powerpoint crash and lose data, 2. behave unexpectedly, such as print with different fonts than on screen, or align objects inconsistently. These may be due to problems outside the program-- but they happened. I have not seen this happen with the SO presentation program. As a result, I trust the latter more.
Slashdot Mirror
Right. They would not solve the problem against motivated attackers; however, ordinary drive-by sniffers would not be able to hear much with their plain 802.11b cards.
Don't forget that at 500km distance, due to Earth's curvature you need to put the end points quite high above the ground; otherwise, they won't see each other beyond the horizon.
The point about lack of incentive on the provider's part is what worries me. What if the current 56kbps modem links up to the 1M cable links are sufficient for most people's networking needs? Then there will never be enough demand to justify upgrading the current last-mile infrastructure. Is there a cool new killer-app which would encourage many people to pay for more bandwidth? (More bandwidth here means substantially more, as in gigabits per second over fiber).
Directional antennas would help with all of these problems. Crossed beams don't interfere, and can't be sniffed from the wrong place.
Speaking of high resolution monitors, check out this new IBM monitor: http://www.provantage.com/YIBML011.HTM. It's a 22" monster with a resolution of 3840 x 2400!!! That's over 200dpi. At $8k, it's a bargain =)
Mr. Nunez's email is evillanueva@congreso.gob.pe. Why not send him a short, polite note of praise for his clear and well reasoned letter?
Wait a second. Her first cat passed away in 2000, and the second one just two years later? Maybe this person thinks the cats are also constructed of Lego bricks, so they don't need feeding, and can be locked up in Lego structures for extraordinary amounts of time :)
My favorite metals at the moment are iridium and osmium, because aside from being very inert, they are much denser than gold.
Making jewelry from a light metal like titanium feels like making aluminum trinkets. It seems kind of lame.
My Sinclair was also similarly elegant, not like these complicated, bloated computers we have today. Here is what I liked about it:
- You always knew what was going on inside the computer, because it could only run one program at a time. There was no multitasking in the operating system.
- Programs tended to be small and compact by design-- design of the memory architecture. There were only 16 kilobytes of RAM available so that developers were always on their toes to keep only the essential features in.
- The layout of the file system was based on the extremely elegant, yet powerful design proposed by Alan Turing-- a magnetic tape. Thus there were no complicated directory hierarchies, just individual programs laid out sequentially.
- The user base was very knowledgeable and helpful. Need to solder a memory upgrade onto the motherboard? No problem, there was always a guy in the local radio amateur club willing to help.
I wish today's machines were more like this! I suppose I can find some consolation in the fact that most microcontrollers found in an average PC have the features of the good old Sinclair.
:P
It sounds like that would have been equivalent to emulating a microprocessor. :)
Some x86 motherboards such as this Supermicro board can take up to 16GB of RAM. However, it's not the total amount of RAM that is the problem; rather, the address space of any single process is limited to 4GB because of 32-bit addressing. So in order to use more than 4GB in a single process, you need to use a 64bit architecture. (OK, maybe you don't *need* to, but it is a heck of a lot easier than custom hacks).
Scenario: company Foo making app Bar figures out they cannot survive by selling free software
Slashdot: The great people at Foo, makers of Bar, are going to have to close their doors due to lack of $$$. Remember Bar? Nice app. Too bad, really. Yawn. Allright, where's that new DVD I ordered?
Use foobar/foobar to read the article.
I've been using the expression "just google for X" when referring friends to some site about X that I know shows up high on the hit list. It's funny, google is starting to replace DNS for me-- instead of remembering or bookmarking URLs, I just remember the keyword to google for. For example, the URL http://www.uk.research.att.com/vnc/ is harder to remeber than to just type "vnc" and hit "I'm feeling lucky."
:) If anyone from google.com is reading this, thank you!
I often wonder how much less productive I would be if google went away tomorrow
Sure, the story was posted here. However, this is a case where the full disclosure argument starts to apply to scientific breakthroughs-- which come from a community fairly disjoint from the community of developers and engineers. Arguing not to disclose scientific achievements is going to be a hard case indeed =)
I would argue that there *is* a maximum version number-- let's call it v100 to denote 100% security. By "security" I necessarily mean technical security from remote attacks, ignoring various shortcuts using social engineering, physical access, etc. A secure protocol based on public key crypto without bugs would be 99% secure, to account for the remote possibility of finding efficient ways to reconstruct private keys.
Now take an implementation with (presumably) many bugs, such as OpenSSH, that is, say, 50% secure. The sooner the trivial engineering bugs can be fixed, the sooner the implementation will be 99% secure, at which point the only hope of compromising its security is in some unpredictable breakthrough in mathematics. This is why the armor-piercing bullet arms race argument doesn't make sense to me. Sooner or later it is bound to reach a limit which can only be surpassed with revolutionary advances.
I don't think full disclosure is really important in this process; instead, open source is what makes it possible to find bugs early. In fact, open source implies full disclosure-- without bugtraq or whatever, you can imagine downloading each new version of SSH and looking at diffs to deduce what bugs got fixed.
I second that. I remember examining a beefy Dell Precision or something similar, and to my great surprise observing a bunch of copper pipes behind the cooling fan grille on the side. Literally, the kind of copper pipes one finds in the back of a refrigerator. My guess is that there was probably some kind of oil inside, not water, in case it leaks out all over the inside of the notebook.
Unfortunately for whomever made up the original buyout hoax, this announcement hit the air one day too early, preventing the culprit from making a killing on RHAT stock.
And the corollary: always make sure your hoaxes air at the beginning of the week, without any pesky holidays in range!
Here's an idea: posting messages should cost money, and the cost should be an area integral of the property values.
Wait a second. There is no transmission involved in obtaining geographic coordinates from GPS-- it's based strictly on observation. Your statement that merely using GPS for this purpose can reveal your location to an eavesdropper is incorrect.
It would be possible if the eavesdropper could capture the message you send out, which would presumably contain the observed GPS coordinates. However, you explicitly ruled this case out.
The reason I don't buy your argument is that this is a very decentralized idea. For example, why couldn't Slashdot include lat/long metadata along its users' comments, and let you show discussions in a certain region? I think the likelihood of getting legal trouble is as great as with the curent, possibly incriminating anonymous posts.
:)
A greater problem with this is that it's very hard to prevent posters from spoofing their geographic location. Thus, any fool with an Internet connection could appear to be posting from the local Starbucks, which would dillute the value of the comments similar to spam in usenet newsgroups. I wonder if some sort of moderation would be useful
If their point is that they can better reconstruct the original image from non-uniform samples, then it would have been more interesting to see a comparison of reconstructions: in particular, take one MRI images with random pixels missing, and the second MRI images with the same number of missing pixels but arranged in a regular pattern, such as a grid.
However, I suspect their point is that they can reconstruct the original at all with non-uniform sampling. This is useful in cases when it is not feasible to obtain fixed samples.
I haven't had a chance to use the Mac version. I've seen the friend's OSX version-- it looks nice.
I made the comment about stability because I've seen 1. x86 Powerpoint crash and lose data, 2. behave unexpectedly, such as print with different fonts than on screen, or align objects inconsistently. These may be due to problems outside the program-- but they happened. I have not seen this happen with the SO presentation program. As a result, I trust the latter more.