If MS puts DRM on it, people won't be able to run their old software anymore. And, AFAIK, software vendors aren't in any deals with MS to sign their software just yet.
How does the government expect to track *all* email? The support isn't in the underlying protocols, and nobody wants to throw away 20+ years of developing email servers just because the government wants a cut. And what about local email? I get emails from my crond daily - would those get taxed, too?
Yes, but it was to correct a number of ebuilds not linking libpthread (including rsync and samba). I was informed via the bug-tracking system that it was due to an old compiler, so I upgraded. The problem wasn't in the toolchain, but in emerge system - the bootstrap worked fine.
What happens if an executable maps in more than 16mb of shared libraries? Will one of these libraries get mapped to NULL? Or will it forgoe security for actually working?
the patch was designed to be as efficient as possible. There's a very minimal (couple of cycles) tracking overhead for every PROT_MMAP system-call, plus there's the 2-3 cycles cost per context-switch.
SIGTERM is a request to quit (roughly equivalnt to windows' default behavior, but dosen't require the message loop to be running). SIGKILL will force it to terminate.
Binary deltas would really be good.
The easiest would be encoding as ascii hex
at the client during commit when the entry is
marked -kb. That would take perhaps 20 minutes
to implement. Of course the storage would be 3x,
but you'd get that back after 2 deltas.
Store it as binary, but convert to hex (or base64) when generating/reading the deltas. Almost as easy to implement.
It doesn't give piece of mind, it gives the illusion of anonymity, which can be even worse.
Freenet is designed to make it virtually impossible to determine what data your node is storing or transmitting. You get plausiable deniability, at least. As for anonymity, are you qualified to assess it? Did you read the protocol spec?
Spaghetti code can work better. Just look at the more successful Corewars programs, especially the evolved ones - they go and modify their code in mid-run, the result being much greater efficiency (and it's more compact, to protect against attacks (or mutations)).
Slashdot Mirror
If MS puts DRM on it, people won't be able to run their old software anymore. And, AFAIK, software vendors aren't in any deals with MS to sign their software just yet.
Bah, this is slashdot. If the moderators don't need a sense of humor, neither do I!
HTTP has supported this for ages. Quit using IE - it's one of the few remaining browsers that dosen't use it.
How does the government expect to track *all* email? The support isn't in the underlying protocols, and nobody wants to throw away 20+ years of developing email servers just because the government wants a cut. And what about local email? I get emails from my crond daily - would those get taxed, too?
Only one problem - tachyons have never been detected.
I did that. It was no help.
Yes, but it was to correct a number of ebuilds not linking libpthread (including rsync and samba). I was informed via the bug-tracking system that it was due to an old compiler, so I upgraded. The problem wasn't in the toolchain, but in emerge system - the bootstrap worked fine.
Gentoo was nice for a while, but things broke when I tried to go from gcc 2.95.x to 3.x. That's what finally convinced me to go to Debian.
Actually, that's a good idea - it would eliminate the losses incurred in photocopying.
On i386 you have to resort to hacks like this to make a readable page of memory non-executable.
Actually, if an app uses an invalid syscall number in Linux, it'll just get back -ENOSYS. Now, calling the wrong syscall, OTOH...
What happens if an executable maps in more than 16mb of shared libraries? Will one of these libraries get mapped to NULL? Or will it forgoe security for actually working?
Overhead:
---------
the patch was designed to be as efficient as possible. There's a very
minimal (couple of cycles) tracking overhead for every PROT_MMAP
system-call, plus there's the 2-3 cycles cost per context-switch.
What would happen if I e-mailed this code to users of a certain brain-dead email client?
SIGTERM is a request to quit (roughly equivalnt to windows' default behavior, but dosen't require the message loop to be running). SIGKILL will force it to terminate.
Are there technical reasons to use ISO9660? Does it have some special error correction, or could I just burn ext2 or something?
What, like this?
If they don't do these tricks, they don't get funding. Simple as that. 'intellectual curiosity' is all well and good, but nobody's paying for it ;)
Spaghetti code can work better. Just look at the more successful Corewars programs, especially the evolved ones - they go and modify their code in mid-run, the result being much greater efficiency (and it's more compact, to protect against attacks (or mutations)).
Actually this could be effective - bacteria are von neumann machines ;)