There are some user-visible changes from rev to rev, like Spotlight (the desktop-grep thing) and Time Machine (versioned filesystem and backup) and fast user switching and iChat and so on. But there are also a ton of developer-visible changes; developers will rely on newer revisions because they have features that make writing programs a lot easier and/or allow writing much better programs; users then have to move to a newer rev in order to run programs which require them.
And, of course, plenty of people don't pay for it. You can more or less get by buying every other OSX upgrade. I hear there are some people still running OS 9.
On the other hand, it does allow each person to prove to themselves that their vote was not tampered with, which is better than we have now.
Vote coercion and vote buying is a big problem with Vodkaneat's proposal but it's a problem we already have with our current system: absentee ballots make vote buying easy again, in a way that wasn't possible with voting-booth systems. Oregon, for example, has moved entirely to vote-by-mail, and other states are moving in the same direction.
This is a good point, since vote coercion is one of the big ways vote fraud has been committed in the past. But note that the system is already broken in that sense: many states are moving towards absentee ballots. Oregon uses absentee ballots entirely these days, and most states no longer require you to demonstrate some unusual circumstance to get an absentee ballot. So all the boss has to do is get you to vote by mail, then watch you fill in and mail the ballot with the boss's choices on it.
What does cost a few thousand dollars is joining the USB-IF, which is necessary to get a unique vendor ID, which is necessary for drivers to be correctly associated with your device. But you can buy small blocks of product IDs from other peoples' vendor space, if you're not planning to make a hundred different gadgets.
I like how the board "[upheld] the standards of confidentiality" by... violating the confidentiality of the board members' personal lives. Yay for double standards!
They can prevent it in part because parts of the GSM protocol are secret (the encryption algorithms, for example) or patent-encumbered. (UMTS is less secret, as I understand it, but it's also much, much more patent-encumbered.)
Or is it that not-getting-out-very-frequently is more common for people with depression? I've even seen studies that suggest a causal link from depression to heart disease.
For consumer products, probably; but this is cutting-edge research. Also, not all problems are perfectly parallelizable; in fact, some problems are hardly parallelizable at all, and two 350GHz chips will not get you the result any faster than one 350GHz chip.
That matches what I've heard. Even a normal healthy eye makes many small involuntary motions (saccades) so the system has to deal with that somehow. Unless they use eyedrops to paralyze the eye muscles, and I assume that those eyedrops would work on nystigmatus as well.
It's also often the case that "dedicated hardware" is just a PC in a fancy case, maybe with some extra fans or a redundant PSU or something. Oh, and some blue LEDs.
There's the Silicon Valley Homebrew Mobile Phone Club. If the comparison with computers is accurate, then in about twenty years we'll start seeing phones that don't suck.
It's a "spallation" source, which means they bombard a heavy nucleus with something (protons in this case) to knock neutrons off. Details at their web site. It looks like the pulse actually contains a wide range of momenta, but since it's very brief (corresponding to the brief proton pulse that produced it) you know the momentum of any given neutron by when it arrives at the target/detector.
It doesn't literally call mmap(), but if what you're wondering is whether the executable files are memory-mapped and demand-paged, the answer is yes (as for pretty much all unixes in the last fifteen years).
In fact, I've had OSX behave in ways that makes me think that large read() calls actually do a mmap-equivelent --- read() followed by a write() to/dev/null doesn't actually ever read the data into memory. Yikes!
Hear, hear. A program should do one thing and do it well. I want my calendaring and my address book and my email client to coöperate, but I don't want them to be the same program. A program doing half-a-dozen unrelated things rarely does any of them well.
Hah, Slashdot's just having to dig farther into the past to find more stuff to repost. C'mon, Linux has been used in "enterprise" situations since the 90s. Maybe the difference is that now management knows about it.
I'm going to be annoyingly pedantic here and point out that the cypherpunk credo is exactly appropriate: information wants to be free. Maybe we don't want it to be free, though, which is sometimes a problem.
The point that "Information wants to be free" is trying to make is that, since information is so easy to duplicate and disseminate, it "wants" (anthropomorphism alert) to be widely available.
Most people misread the phrase, and think it says "Information ought to be free". That's not the same thing at all.
Re:Why not a community based p2p client/network ?
on
Razorback2 Servers Seized
·
· Score: 2, Informative
There are several of these. Freenet (Ian Clarke's Freenet, not the other things called Freenet) is one good example. The problem is actually prety difficult: since you can't have any centralized servers, and you can't necessarily trust all nodes, it's really hard to write something that will do what you want.
There's a tension between privacy and speed; it's hard to get both, since if you keep information private then you can't use that information to optimize the network. BitTorrent manages to be fast in a wide variety of situations, but it doesn't keep any secrets. Freenet manages to be secure in a wide variety of situations, but it can be really sloooow. They take totally different approaches.
Slashdot Mirror
There are some user-visible changes from rev to rev, like Spotlight (the desktop-grep thing) and Time Machine (versioned filesystem and backup) and fast user switching and iChat and so on. But there are also a ton of developer-visible changes; developers will rely on newer revisions because they have features that make writing programs a lot easier and/or allow writing much better programs; users then have to move to a newer rev in order to run programs which require them. And, of course, plenty of people don't pay for it. You can more or less get by buying every other OSX upgrade. I hear there are some people still running OS 9.
On the other hand, it does allow each person to prove to themselves that their vote was not tampered with, which is better than we have now.
Vote coercion and vote buying is a big problem with Vodkaneat's proposal but it's a problem we already have with our current system: absentee ballots make vote buying easy again, in a way that wasn't possible with voting-booth systems. Oregon, for example, has moved entirely to vote-by-mail, and other states are moving in the same direction.
Did you seriously just refer to a slashdot article as a "serious news story"?
This is a good point, since vote coercion is one of the big ways vote fraud has been committed in the past. But note that the system is already broken in that sense: many states are moving towards absentee ballots. Oregon uses absentee ballots entirely these days, and most states no longer require you to demonstrate some unusual circumstance to get an absentee ballot. So all the boss has to do is get you to vote by mail, then watch you fill in and mail the ballot with the boss's choices on it.
- USB Bit Whacker
- Arduino
- (you can buy the above two from SparkFun if you don't feel like DIY)
- LabJack
- various FTDI-based devices from many companies
What does cost a few thousand dollars is joining the USB-IF, which is necessary to get a unique vendor ID, which is necessary for drivers to be correctly associated with your device. But you can buy small blocks of product IDs from other peoples' vendor space, if you're not planning to make a hundred different gadgets.I like how the board "[upheld] the standards of confidentiality" by ... violating the confidentiality of the board members' personal lives. Yay for double standards!
Is that as in, "*kick door down* Freeze! Abe Lincoln!" or as in, "My humps, my lovely lady lumps, SHALL NOT PERISH FROM THE EARTH." ?
Your firewalls are red? My firewalls are black. From soot.
"My name is Linus, and I am your god."
They can prevent it in part because parts of the GSM protocol are secret (the encryption algorithms, for example) or patent-encumbered. (UMTS is less secret, as I understand it, but it's also much, much more patent-encumbered.)
Or is it that not-getting-out-very-frequently is more common for people with depression? I've even seen studies that suggest a causal link from depression to heart disease.
Then, go read Lessig's Code, and other laws of cyberspace.
For consumer products, probably; but this is cutting-edge research. Also, not all problems are perfectly parallelizable; in fact, some problems are hardly parallelizable at all, and two 350GHz chips will not get you the result any faster than one 350GHz chip.
Obviously we need to scale up the Air Force until it has more foot soldiers than the Army.
That matches what I've heard. Even a normal healthy eye makes many small involuntary motions (saccades) so the system has to deal with that somehow. Unless they use eyedrops to paralyze the eye muscles, and I assume that those eyedrops would work on nystigmatus as well.
Why on earth would run run it inside xen? Just run it as a plain old process. Sheesh. It's like we're regressing to the days of VM/{CMS,MVS,etc}.
It's also often the case that "dedicated hardware" is just a PC in a fancy case, maybe with some extra fans or a redundant PSU or something. Oh, and some blue LEDs.
There's the Silicon Valley Homebrew Mobile Phone Club. If the comparison with computers is accurate, then in about twenty years we'll start seeing phones that don't suck.
Froogle shows a bunch of US sellers.
It's a "spallation" source, which means they bombard a heavy nucleus with something (protons in this case) to knock neutrons off. Details at their web site. It looks like the pulse actually contains a wide range of momenta, but since it's very brief (corresponding to the brief proton pulse that produced it) you know the momentum of any given neutron by when it arrives at the target/detector.
It doesn't literally call mmap(), but if what you're wondering is whether the executable files are memory-mapped and demand-paged, the answer is yes (as for pretty much all unixes in the last fifteen years). In fact, I've had OSX behave in ways that makes me think that large read() calls actually do a mmap-equivelent --- read() followed by a write() to /dev/null doesn't actually ever read the data into memory. Yikes!
Hear, hear. A program should do one thing and do it well. I want my calendaring and my address book and my email client to coöperate, but I don't want them to be the same program. A program doing half-a-dozen unrelated things rarely does any of them well.
Hah, Slashdot's just having to dig farther into the past to find more stuff to repost. C'mon, Linux has been used in "enterprise" situations since the 90s. Maybe the difference is that now management knows about it.
I'm going to be annoyingly pedantic here and point out that the cypherpunk credo is exactly appropriate: information wants to be free. Maybe we don't want it to be free, though, which is sometimes a problem.
The point that "Information wants to be free" is trying to make is that, since information is so easy to duplicate and disseminate, it "wants" (anthropomorphism alert) to be widely available.
Most people misread the phrase, and think it says "Information ought to be free". That's not the same thing at all.
There are several of these. Freenet (Ian Clarke's Freenet, not the other things called Freenet) is one good example. The problem is actually prety difficult: since you can't have any centralized servers, and you can't necessarily trust all nodes, it's really hard to write something that will do what you want.
There's a tension between privacy and speed; it's hard to get both, since if you keep information private then you can't use that information to optimize the network. BitTorrent manages to be fast in a wide variety of situations, but it doesn't keep any secrets. Freenet manages to be secure in a wide variety of situations, but it can be really sloooow. They take totally different approaches.