The GPL is not an EULA. Read it. It does not apply if you only run the software, or if you use the software in the ways normally allowed by fair use (making backup copies, copying to RAM for execution, that sort of stuff). On the contrary, an EULA purports to apply as soon as you click ok during the installation.
The GPL gives you rights that you do not have under copyright/fair use. It does not take away any of the rights that you have under copyright/fair use.
Now prices can remain the same while profits go up...
Right, that is how the semiconductor world typically works. This has kept computer prices at the same level since the seventies and ensured the current situation where there are only a few thousand computers in the world. Imagine if prices on IC's had been allowed to fall - there might have been a computer in front of me now. Or maybe someone would even have managed to create a digital watch!
Have we really come to the point where even Slashdotters don't know the difference between silicon and silicone? It is no wonder that the perfectly good word niggardly is lost forever.
Buying SCO's Unix IP and going after Linux with that would most likely result in more antitrust attention at Microsoft. It is much more convenient for them that someone else is doing the suing.
Basically I would end up having to trust a set of experts. The number of people who can reverse engineer at the required level is simply too small, at least in a country without advanced IC factories. Besides, hardware bugs are obviously hard to find. I think that a sufficiently smart engineer could put in a backdoor that had a good chance of either slipping through testing or looking like an honest error if found.
Anyway, what do you mean by "chief executive"? The Danish government does not have a president. Perhaps you would suggest that the Queen is the chief executive. Elections are hardly a problem there though... Or take the German government. The president has so little power that few outside Germany even know his name.
How do I, as a voter, verify those machines? Will I be allowed to take them apart after the election? Will I be instructed in the use of an electron microscope when I need to verify the workings of the chips?
By the way, there are plenty of democracies without presidents.
If you have hardware smart enough to cryptographically ensure that it is binaries compiled from a certain source that is running, how do you trust that the hardware is running the software faithfully?
And as to the thrown away votes, as long as there are only two outcomes of a vote, you will always have a worst-case of half the votes thrown away. They didn't get any representation, no matter which system.
I think you should strike the "closed-source" part. Even if you have the source, you cannot prove that this source corresponds to the binaries running on the machine.
What is really needed is a system where everyone can prove that the system works. The easiest way to do that is manually counted paper ballots.
It will always be a problem when a very small number of voters have a very large influence on the outcome. All the people who vote for any of the losers in a US presidential election had their vote thrown away. With a proportional voting system, a small error is not so disastrous. Of course, you cannot elect a president with a proportional voting system.
The filesystem thought it was using UTF-8 filenames. That is what the specification says it should use. However the unfortunate poster has used ISO-8859-1 (or -15) file names. Therefore he now has a file system that does not conform to the standard, and of course he wants to do something about it.
The RF-to-digital block is implemented as a card that covers the frequency range from 30 MHz to 2.5Gz
Later:
The RF card contains several antenna ports. The active port can be selected through software to enable use of different antennas for different bands, or multiple can be activated for applications that exploit diversity.
It seems that they have the analog part pretty much handled.
These things are the radio equivalent of winmodems - cheap frontend with an already-available processor. They have the potential to get really really cheap, just like winmodems. And like winmodems, in the beginning the processing power needed is annoyingly high. These days noone care that 2% of the CPU is spent on the winmodem. One day, noone will care that 10% of the CPU is spent on software radio. Instead they will love how they can get access to WLAN, bluetooth, and cellular with one simple and cheap device. Oh and listen to radio, whether analog or digital.
1Gbps ethernet is specified for 100m on CAT5e-cable. You cannot necessarily expect available cables to follow the specification; there are plenty of out-of-spec USB and PATA-cables.
You may be used to to ultrafast connections, but for many people a slow permanent connection is very useful. I should know, I used to have a 33kbps permanent circuit a few years ago. Just enforce caching and avoid p2p. Schedule heavy transfers between 1am and 5am. In any case, such a connection is much more useful than a metered connection like ISDN. With a metered connection you cannot check your email all the time, you cannot be on messaging services or chat much, and when browsing you have to keep an eye on the clock. Waiting is not that much of a problem when you are not metered for the time you wait...
Best of all, when the money starts flowing they only have to upgrade one connection. If they do bandwidth allocation right, they will have a very low latency (unlike POTS or ISDN).
Why is it a joke? It is a fairly common thing to do. The only thing stopping it from working in this case is the distances involved. With a fiber switch it is probably the best option available to them.
A monkey for each natural number is enough. The set of possible strings is clearly countably infinite; most easily seen if you encode the strings in your favourite computer encoding. *poof* all strings are assigned natural numbers and hence there can only be a countable infinity of them.
The problem is that apparently noone signed a paper with the rules for patent disclosure at JEDEC. The rules were supposedly agreed to verbally. This does not make the rules any less valid, but it is a problem to prove that the rules existed. When it suddenly turns out that the chairman himself did not follow the rules, the existence of the rules is in serious doubt.
No matter what happens, it is obvious that what Rambus did was very sleazy. It is not, however, obvious that they were sleazy in an illegal way. It is not even obvious that they were the sleaziest company of the bunch in JEDEC.
It all reminds me of a bunch of card sharks cheating and winning the shirts off of everyone else. Yet another clueless player shows up, and proceeds to win a fortune from them by cheating in a new and inventive way. Then the sharks call the police, complaining about the cheating...
Now, do you think it's reasonable that the average Slashdot reader (Americans) should have to know ALL that to understand this little blurb?
Absolutely. I think you are underestimating the people visiting Slashdot. And even if the average Slashdotter is as, uh, educationally challenged as you imply, this article may have made them either understand a little more about the world or scared them away. Both cases leave Slashdot with a more educated audience.
I will be on this planet. My family will be on this planet. My friends will be on this planet. Tell me again why I should care whether a few hundred people on Mars stay alive?
A Linux client was promised even before Neverwinter Nights was released. Whether it was planned or not is rather irrelevant for me; it was promised. Fortunately (?) I have been burned often enough to never again buy software on promises.
Multicast is putting the intelligence into the network. It requires big smart routers that keep track of who gets which streams. This is the opposite of what the Internet is about. The Internet solution is to replicate at the endpoints: If you want to get a stream, fetch it off of the nearest person who already listens to that stream. Basically P2P but streaming instead of moving files about.
Incidentally, multicast works just as well (or rather, horribly) with IPv4 as with IPv6. P2P-streaming on the other hand requires that it is possible to connect to all the participants in the stream. IPv4 with NAT makes that a hopeless proposition. Let us hope that IPv6 happens and that laws against P2P-streaming do not get passed.
I must admit that I don't think "bad programmers afraid to be fired because their code has a security flaw" as a large enough demographic to be worth worrying about. Good programmers make mistakes, and a little extra safety may lengthen the time between discovery (and hopefully disclosure) and exploit. Bad programmers will find a way to screw up either way, but probably not any worse than without the extra safety.
Physical and virtual exploits are very different. As an example, if someone breaks your window and burgles your house, you rarely blame the makers of the window. Exploiting the weakness tends to be risky for the attacker, and the cost of fixing the weakness (unbreakable windows) is prohibitive in most cases.
In contrast, software problems are often easy to exploit anonymously and the cost of fixing them once found is usually very low.
It is highly unlikely that RF in the power range of cell phones will crash a plane. Imagine that it sometimes interferes with sensors (such as smoke sensors) and makes pilots believe that there is a problem with the plane. This could mean unscheduled landings, and those are very expensive. Yet terrorists might find causing an unscheduled landing oddly unfulfilling. Banning cell phones on planes seem to be a reasonable course of action in this case.
Drivers have a skewed view of the cost of driving dangerously. Humans are not very good at intuition about probabilities (if they were, lotteries would be a lot less popular.) Also, if humans perceived all risks the same way, many people would have fear of driving the same way they have fear of heights. The thing to be afraid of (high speed impact with with immobile object) is about the same whether you are in a car or at the top of a cliff.
Whether an additional safety feature will make people even more confident is hard to say and more of a psychology question than an economics question.
Slashdot Mirror
The GPL gives you rights that you do not have under copyright/fair use. It does not take away any of the rights that you have under copyright/fair use.
Right, that is how the semiconductor world typically works. This has kept computer prices at the same level since the seventies and ensured the current situation where there are only a few thousand computers in the world. Imagine if prices on IC's had been allowed to fall - there might have been a computer in front of me now. Or maybe someone would even have managed to create a digital watch!
Have we really come to the point where even Slashdotters don't know the difference between silicon and silicone? It is no wonder that the perfectly good word niggardly is lost forever.
Buying SCO's Unix IP and going after Linux with that would most likely result in more antitrust attention at Microsoft. It is much more convenient for them that someone else is doing the suing.
Basically I would end up having to trust a set of experts. The number of people who can reverse engineer at the required level is simply too small, at least in a country without advanced IC factories. Besides, hardware bugs are obviously hard to find. I think that a sufficiently smart engineer could put in a backdoor that had a good chance of either slipping through testing or looking like an honest error if found.
Anyway, what do you mean by "chief executive"? The Danish government does not have a president. Perhaps you would suggest that the Queen is the chief executive. Elections are hardly a problem there though... Or take the German government. The president has so little power that few outside Germany even know his name.
How do I, as a voter, verify those machines? Will I be allowed to take them apart after the election? Will I be instructed in the use of an electron microscope when I need to verify the workings of the chips?
By the way, there are plenty of democracies without presidents.
If you have hardware smart enough to cryptographically ensure that it is binaries compiled from a certain source that is running, how do you trust that the hardware is running the software faithfully?
And as to the thrown away votes, as long as there are only two outcomes of a vote, you will always have a worst-case of half the votes thrown away. They didn't get any representation, no matter which system.
I think you should strike the "closed-source" part. Even if you have the source, you cannot prove that this source corresponds to the binaries running on the machine.
What is really needed is a system where everyone can prove that the system works. The easiest way to do that is manually counted paper ballots.
It will always be a problem when a very small number of voters have a very large influence on the outcome. All the people who vote for any of the losers in a US presidential election had their vote thrown away. With a proportional voting system, a small error is not so disastrous. Of course, you cannot elect a president with a proportional voting system.
The filesystem thought it was using UTF-8 filenames. That is what the specification says it should use. However the unfortunate poster has used ISO-8859-1 (or -15) file names. Therefore he now has a file system that does not conform to the standard, and of course he wants to do something about it.
1Gbps ethernet is specified for 100m on CAT5e-cable. You cannot necessarily expect available cables to follow the specification; there are plenty of out-of-spec USB and PATA-cables.
Best of all, when the money starts flowing they only have to upgrade one connection. If they do bandwidth allocation right, they will have a very low latency (unlike POTS or ISDN).
Why is it a joke? It is a fairly common thing to do. The only thing stopping it from working in this case is the distances involved. With a fiber switch it is probably the best option available to them.
Ok, you are right about infinite strings. Fortunately Shakespeare only managed to come up with a finite string.
A monkey for each natural number is enough. The set of possible strings is clearly countably infinite; most easily seen if you encode the strings in your favourite computer encoding. *poof* all strings are assigned natural numbers and hence there can only be a countable infinity of them.
Fragmentation is a problem on hard drives because seek time is so high. This is a non-problem with solid-state storage. So just don't defragment.
No matter what happens, it is obvious that what Rambus did was very sleazy. It is not, however, obvious that they were sleazy in an illegal way. It is not even obvious that they were the sleaziest company of the bunch in JEDEC.
It all reminds me of a bunch of card sharks cheating and winning the shirts off of everyone else. Yet another clueless player shows up, and proceeds to win a fortune from them by cheating in a new and inventive way. Then the sharks call the police, complaining about the cheating...
The whole thing makes me sad.
Absolutely. I think you are underestimating the people visiting Slashdot. And even if the average Slashdotter is as, uh, educationally challenged as you imply, this article may have made them either understand a little more about the world or scared them away. Both cases leave Slashdot with a more educated audience.
I will be on this planet. My family will be on this planet. My friends will be on this planet. Tell me again why I should care whether a few hundred people on Mars stay alive?
A Linux client was promised even before Neverwinter Nights was released. Whether it was planned or not is rather irrelevant for me; it was promised. Fortunately (?) I have been burned often enough to never again buy software on promises.
Incidentally, multicast works just as well (or rather, horribly) with IPv4 as with IPv6. P2P-streaming on the other hand requires that it is possible to connect to all the participants in the stream. IPv4 with NAT makes that a hopeless proposition. Let us hope that IPv6 happens and that laws against P2P-streaming do not get passed.
I must admit that I don't think "bad programmers afraid to be fired because their code has a security flaw" as a large enough demographic to be worth worrying about. Good programmers make mistakes, and a little extra safety may lengthen the time between discovery (and hopefully disclosure) and exploit. Bad programmers will find a way to screw up either way, but probably not any worse than without the extra safety.
In contrast, software problems are often easy to exploit anonymously and the cost of fixing them once found is usually very low.
It is highly unlikely that RF in the power range of cell phones will crash a plane. Imagine that it sometimes interferes with sensors (such as smoke sensors) and makes pilots believe that there is a problem with the plane. This could mean unscheduled landings, and those are very expensive. Yet terrorists might find causing an unscheduled landing oddly unfulfilling. Banning cell phones on planes seem to be a reasonable course of action in this case.
Whether an additional safety feature will make people even more confident is hard to say and more of a psychology question than an economics question.