So? Cache like google. No url rewriting, or at least no tag rewriting. That way, your banner will be still be viewed as much as possible, but/. takes the burden for the pages. As long as your ad-server can handle it, you get all the impressions you want...
The big problem is legal. It's mirroring without permission, which may or may not be illegal (I'm not touching that one with a ten-foot pole).
I'm writing this as I watch the mpg, a few points I've noticed:
I really love the fact that the average bandwidth has come so far that we can just send around 300M mpeg's of these things. YAY!
The FBI guy doesn't really seem that technically competent, for instance one of the slides mentioned not storing the "re:" and "subject:" lines... Hmm, did I miss an rfc? I don't recall the re: header... Also, he didn't know email addresses were case-insensitive, no biggie but you'd think someone close to carnivore would know that.
Pity they couldn't have gone a bit higher res, non of the computer screens where readable.
Does anybody know an mpeg-player for linux that shows a time-index in the controls (apart from mtv). Gtv is great, but only gives framenumber, which isn't even relative to the start of the mpeg but to the point where you started playing it.
For some reason he seemed to be trying to avoid the 'only gather addresses' mode because he thought there would be a lot of disagreement about how much it could gather. Isn't bringing that into the open the entire idea of this thingy?
a large part seems to consist of explaining how a packet filter works. Fun, but not what I wanted to see...
They've demonstrated how carnivore can filter to very specific rules, so not to collect anyone elses traffic. I haven't however heard how many of the court orders allowing carnivore use so far have specified any of these filters, or which ones....
All irrelevant data is discarded, and sealed and locked away. Why? Why not just log that data from from was discarded, and wipe it, giving better chances of privacy for those not involved?
In the demo they have a webinterface is used. This uses a form, which implies a cgi (could be javascript, but unlikely), which implies a webserver. That implies a webserver is running locally. Seems to be a) slightly overkill b) a security risk if the webserver is malconfigured, ie somebody could connect to it remotely.
On the public: This is a really cool bunch of geeks, I love em. I want to take em home...
As far as sniffers/parsers go, carnivore/coolminer seems pretty nice stuff. I want! Could the person who got into microsoft please nick this stuff from the fbi and put it on geocities? A port to linux/bsd (X11) would be nice too.
Hmm, there's one guy doing a feature-request. Weird.:-)
I really get the feeling the fbi guy is just saying what everybody wants to hear...
The mpeg just ends somewhere in the middle of a discussion. Very irritating. (It could be the end, but it doesn't look like it)
True, the banshee isn't the hottest thing out there for 3d (but I does fine for me). However, it has the added bonus over the voodoo2 that 2d accel is pretty impressive. (Don't know how it does against the voodoo 3 2d-wise...)
It doesn't matter if it's US english or brittish english. They're just different dialects, Anyone who can understand one can understand the other, so you can use whichever one you want online. Personally I (try to) use brittish english, and that isn't any problem online. However, dutch would be significantly more tricky..
you might want to try a banshee. I've got one in my p2-233, and love it. Things like Q3 and (more importantly) UT rune perfectly... I'm using AGP, but iirc there's also a pci version (the pci is a diamond card iirc).
I wouldn't have a problem with that, but when it goes to a chemical analysis I admit it's a bit less nice. I'll agree that it's a slippery slope. However, I still don't have a problem with monitoring computers at work. You have to go to the bathroom at work (well, more or less). However you don't have to use the comp for things you want to keep private at work...
No, they don't have the right to do that. They do, however, have the right to take snapshots of the toilets, even when you are using them, as long as the photo's don't have you on them.
I assume nothing's changed. It was said reiserfs wasn't going into 2.4.0, but there's nothing against getting a big patch into 2.4.1 if 2.4.0 turns out to be good and stable.
Why not make it mouseover+click shopping? First you have to put your mouse over the buy button, which makes it change colour. Then, you have to click to confirm your order. Ok, it's not lynx/w3m friendly, but apart from that. (BTW, does anybody even _want_ 1-click shopping???)
Nope, He's quite correct. He says: 'In an average week, how many people are struck by lightning for the second time?'
This means that they don't need to have been hit by lightening for the first time in the same week. It doesn't matter if you do it by week or by lifetime.
The original poster of the toplevel comment claimed that you had a higher chance of getting hit by lightening twice than you had of winning the lottery (once). If this where true, more people would get hit by lightening for the second time (but not necessarily twice on that day) than win the lottery.
My bad, I was just trying to state that a known-plaintext attack is possible, though you don't have a lot of plaintext. After reading the parent post I really don't have the faintest idea why I mentioned that, must have been one of those days:-).
Encrypted swap shouldn't be a problem under linux,
you can do it just like regular encrypted disks. losetup it encrypted to a loop device, and then 'swapon' it.
You might want to get your crypto straight. 1024/2048 bits is the kind of keysizes you use for public key crypto. For encrypting a partition you use symetric crypto, for which anything between 128 and 256 bits keys are enough.
Encrypting seperate files is not really ideal, since the apps using it have to support it. After all, there's no way for the kernel to ask a user for a password from the kernel, you don't know if the user is doing X or whatever, so It'd have to be a param to open().
Secondly, a minor point: If you were going to implement it, it would probably be better off as an attribute instead of a mode, see man chattr for a little more info.
Ehm, there is a header: large parts of your first block from the encrypted disk are the same (ext2 headers). I too use an encrypted/home btw, also via loopback. It's fast, but definately noticable if you do large copies or try compiling something. However, if you do single DES on a fast comp it shouldn't be noticable.
And kill a lot on their cpu. I'm not saying it's a bad idea to use -z, it depends on their setup.. But z3 is going to kill a lot of cpu time, 2 or 1 are probably better for most setups.
Not necessarily, depends if it does the standard ps/2 interface thingy. At a minimum, all the info should be readable/writeable from/dev/mouse (or/dev/thingy) as a non-root user (if you have +rw).
After all, somebody might want to use mouse textures in gpm. (yes, gpm runs as root, but that's not the point).
Hmm, I may have formulated that a bit unclearly. It's not about any one specific pixel. It's just that you can have a 3d image with a dot at (0,0,10) and in the same image a dot at (0,0,11). However, you cannot have (0,0,green) and (0,0,pink) in the same image.
Nope, it's not about where one specific object can be. The problem is: a pixel can be at (0,0,10), and another pixel can be at (0,0,11), at the same time. However, you cannot have a pixel at (0,0,blue) and (0,0,red) at the same time.
Slashdot Mirror
They seem to have fixed the validation, I didn't have any troubles (and didn't fill in second address line).
So? Cache like google. No url rewriting, or at least no tag rewriting. That way, your banner will be still be viewed as much as possible, but /. takes the burden for the pages. As long as your ad-server can handle it, you get all the impressions you want...
The big problem is legal. It's mirroring without permission, which may or may not be illegal (I'm not touching that one with a ten-foot pole).
True, the banshee isn't the hottest thing out there for 3d (but I does fine for me). However, it has the added bonus over the voodoo2 that 2d accel is pretty impressive. (Don't know how it does against the voodoo 3 2d-wise...)
It doesn't matter if it's US english or brittish english. They're just different dialects, Anyone who can understand one can understand the other, so you can use whichever one you want online. Personally I (try to) use brittish english, and that isn't any problem online. However, dutch would be significantly more tricky..
you might want to try a banshee. I've got one in my p2-233, and love it. Things like Q3 and (more importantly) UT rune perfectly... I'm using AGP, but iirc there's also a pci version (the pci is a diamond card iirc).
p2/233 64mb here, and doing fine. (not that I don't dream of a dual 1ghz athlon, but the needs nor the funds are really there... :-) ).
I wouldn't have a problem with that, but when it goes to a chemical analysis I admit it's a bit less nice. I'll agree that it's a slippery slope. However, I still don't have a problem with monitoring computers at work. You have to go to the bathroom at work (well, more or less). However you don't have to use the comp for things you want to keep private at work...
No, they don't have the right to do that. They do, however, have the right to take snapshots of the toilets, even when you are using them, as long as the photo's don't have you on them.
More like 'the dim side of the web' :-)
I assume nothing's changed. It was said reiserfs wasn't going into 2.4.0, but there's nothing against getting a big patch into 2.4.1 if 2.4.0 turns out to be good and stable.
Patches appear, patches get used, patches die out.
:-)
forks appear, forks get used, forks die out.
Seems like natural selection to me. Anyways, where you trying to imply linus isn't natural?
Why not make it mouseover+click shopping? First you have to put your mouse over the buy button, which makes it change colour. Then, you have to click to confirm your order. Ok, it's not lynx/w3m friendly, but apart from that. (BTW, does anybody even _want_ 1-click shopping???)
Nope, He's quite correct. He says: 'In an average week, how many people are struck by lightning for the second time?'
This means that they don't need to have been hit by lightening for the first time in the same week. It doesn't matter if you do it by week or by lifetime.
The original poster of the toplevel comment claimed that you had a higher chance of getting hit by lightening twice than you had of winning the lottery (once). If this where true, more people would get hit by lightening for the second time (but not necessarily twice on that day) than win the lottery.
My bad, I was just trying to state that a known-plaintext attack is possible, though you don't have a lot of plaintext. After reading the parent post I really don't have the faintest idea why I mentioned that, must have been one of those days :-).
Encrypted swap shouldn't be a problem under linux, you can do it just like regular encrypted disks. losetup it encrypted to a loop device, and then 'swapon' it.
You might want to get your crypto straight. 1024/2048 bits is the kind of keysizes you use for public key crypto. For encrypting a partition you use symetric crypto, for which anything between 128 and 256 bits keys are enough.
Encrypting seperate files is not really ideal, since the apps using it have to support it. After all, there's no way for the kernel to ask a user for a password from the kernel, you don't know if the user is doing X or whatever, so It'd have to be a param to open(). Secondly, a minor point: If you were going to implement it, it would probably be better off as an attribute instead of a mode, see man chattr for a little more info.
Ehm, there is a header: large parts of your first block from the encrypted disk are the same (ext2 headers). I too use an encrypted /home btw, also via loopback. It's fast, but definately noticable if you do large copies or try compiling something. However, if you do single DES on a fast comp it shouldn't be noticable.
And kill a lot on their cpu. I'm not saying it's a bad idea to use -z, it depends on their setup.. But z3 is going to kill a lot of cpu time, 2 or 1 are probably better for most setups.
Not necessarily, depends if it does the standard ps/2 interface thingy. At a minimum, all the info should be readable/writeable from /dev/mouse (or /dev/thingy) as a non-root user (if you have +rw).
After all, somebody might want to use mouse textures in gpm. (yes, gpm runs as root, but that's not the point).
I'm not sure you get a million, but this is indeed 1 of the 7 problems the clay institute has put up a pretty big (financial) reward for...
Hmm, I may have formulated that a bit unclearly. It's not about any one specific pixel. It's just that you can have a 3d image with a dot at (0,0,10) and in the same image a dot at (0,0,11). However, you cannot have (0,0,green) and (0,0,pink) in the same image.
Nope, it's not about where one specific object can be. The problem is: a pixel can be at (0,0,10), and another pixel can be at (0,0,11), at the same time. However, you cannot have a pixel at (0,0,blue) and (0,0,red) at the same time.
Heck, have you ever noticed all those people with no tea? _that_ is how much it's influenced us.
Colour cannot be seen as a dimension, unless a pixel can be both blue and red at the same time...
Time otoh....