Ah, but if you're shipping networking gear, the telecom industry is broke, so the only major source of customers is the military. You're going to design to fit the requirements of the paying customers, and as keeping two designs going is more expensive, there will be more and more ipv6-capable gear sold to everyone if the military demands it.
People forget that as late as the early 80s, the US military bought 50% or more of all electronics purchased in the US.
This is already being done. Here are some examples:
IBM and Rice University have both licensed patents having to do with register allocation, so that GCC can use them.
IBM has licensed its RCU patent, which is used by the Linux kernel (this is a case where SCO is claiming ownership of the technology even though IBM owns the patent!)
Raph Levien, of Advogato and Ghostscript fame, has licensed a whole series of patents he holds with respect to printing technology for use in GPLed code.
In all cases, these patents are free to use by any GPLed software, but not by non-copylefted free software.
What you say already works for the WiFi network. The problem is that it doesn't work for the cellular network, because it is a proprietary, nonstandard network, much like the old French Minitel system, designed to ding you with a bunch of charges for very limited functionality.
What should happen is that your WiFi and cellular networks are both TCP/IP, and your OS knows how to set up and take down routes depending on where you are (as you say, based on signal strength). Verizon does not have to "manage your experience", they just need to enable it, meaning that they give you a straight TCP/IP network ("Internet dial tone", as some of the digirati have called it) and let you run whatever app you want.
There's no question that RMS resents the fact that Linus gets more credit than he deserves. Just the same, people forget history: Linux only took off because of the AT&T/BSD lawsuit of the early 90s; in 1992, Linux was way behind the BSDs. If Linux (the kernel) comes under severe legal attack (that is, if SCO actually manages to legally impede the use of the Linux kernel), it's completely proper to point out that the GNU folks have a backup plan.
However, the realistic backup plan is not the Hurd, which as far as I'm concerned was a huge botch, which after more than ten years hasn't reached the level of usability that the Linux kernel had after two. It's Debian GNU/BSD. With a bit more work, anyone using this system won't be able to tell that they aren't using a Debian GNU/Linux system, because every single application will work the same way.
You misunderstand what RMS means when he says that
the GPL only adds rights, it does not take any rights away. RMS is absolutely correct in a legal sense. That's because if he just wrote "Copyright FSF, all rights reserved", the default situation is that of copyright law, which says that no one but the copyright holder may make copies or derivative works. If you have a legal copy you can run it, make one backup, study how it works, and so forth. These are the kinds of things that shrinkwrap licenses sometimes restrict, but copyright law does not restrict.
The reason it's often called a tax is that paying that $50 is not optional in most cases (especially for laptops), even if you don't want to use the OS at all.
As you say, it might not be worth the hassle to most people to figure out how not to pay that $50.
Intel tunes heavily to score well on important benchmakrks, while gcc doesn't. So, while on average icc generates faster code, the ratio is much smaller than SPEC ratings would suggest, and there are some anomalies like the one you mention, where one compiler or the other generates really slow code.
Censorship occurs when someone else tells you that you can't say, or see, or read something. There is no censorship when you make a personal decision not to watch a commercial, or a nude scene, or whatever else you don't care for.
The most well-known filesystem-based version control system is Clearcase. Essentially, the user sets up a view, which has a config spec. The config spec specifies what versions of each directory and file the user wants to see. Checkins and checkouts are done as in CVS, except that directories and file renames are tracked as well.
Lessig proposes to solve this by technical standards, backed by law: the bandwidth has no central authority, but there are power limitations and protocols for the devices so that everyone can "talk".
No, you didn't recently read that. What you read was that the real effective data rate over "g" is more like 20 Mbps, and that in a network with a lot of "b" devices it might be more like ten. The raw speed is still 54 Mbps, the real speed is less because of the signaling overhead. There is similar overhead in 802.11b, you only get about 5 Mbps, and if there are any Bluetooth devices around you'll get a lot less because of interference.
802.11g is designed to interoperate with 802.11b, although the presence of "b" users in the same area does slow "g" down. Still, everyone is confusing effective rate (say, 20 Mbits/sec actually transmitted) with theoretical peak rate (54 Mbit/sec). "b" users are not getting 11 Mbits/sec; if they are lucky they are getting 5, and if they are surfing the web through DSL or cable modem they aren't even getting 2. When lots of people are using the same access point, the bottleneck isn't
In a year or two, most folks will ditch their "b" equipment for "g" and it won't matter.
Google hasn't exactly patented the algorithm for all uses, and no court has determined that the code infringes the patent, and software patents aren't valid in most countries, so it's not clear whether or not there is any compatibility.
It would seem that anyone who uses the code to build a search engine would be infringing, but even that is something that lawyers can argue about.
You'd lose the case, because about $50 is what MS charges companies like Dell per copy when they buy in bulk. Their lawyers would say that this is the fair value, and the rest is markup for the retail distribution system. Pretty high markup, but hardly unprecedented.
In any case, if Microsoft can be accused of dumping for charging $50 for their software, couldn't a case be made against Red Hat for providing free ISO downloads? Isn't that dumping? It costs more than $0 to provide that service, so they are clearly dumping it.
But SCO isn't alleging that Linux is a pirated copy. The analogy would be to rap artists using samples without authorization. If they could find 4,700 stolen lines, then 0.1% of Linux would be copied, and that's like 0.2 seconds of a three-minute song.
They allege "hundreds of lines", and have shown 80, though they haven't offered any evidence about the direction of copying.
Let's not lose context here.
The Linux kernel is roughly 4.7 million lines of code. If the duplicated code is in a device driver, it could be because the same person contributed a device driver for the same device to both OSes.
There can be any number of other explanations.
In any given area (other than, say, Manhattan) there will be many unused television channels. Just allow wireless networking equipment to use those channels. Any of several techniques could be used to avoid interference: detect a TV broadcast signal, or just use a database ("I'm in Phoenix, therefore channels x, y, and z are free").
Slashdot Mirror
Ah, but if you're shipping networking gear, the telecom industry is broke, so the only major source of customers is the military. You're going to design to fit the requirements of the paying customers, and as keeping two designs going is more expensive, there will be more and more ipv6-capable gear sold to everyone if the military demands it.
People forget that as late as the early 80s, the US military bought 50% or more of all electronics purchased in the US.
The Carter administration tried to get the country converted to metric, but Reagan killed it.
This is already being done. Here are some examples:
In all cases, these patents are free to use by any GPLed software, but not by non-copylefted free software.
What you say already works for the WiFi network. The problem is that it doesn't work for the cellular network, because it is a proprietary, nonstandard network, much like the old French Minitel system, designed to ding you with a bunch of charges for very limited functionality.
What should happen is that your WiFi and cellular networks are both TCP/IP, and your OS knows how to set up and take down routes depending on where you are (as you say, based on signal strength). Verizon does not have to "manage your experience", they just need to enable it, meaning that they give you a straight TCP/IP network ("Internet dial tone", as some of the digirati have called it) and let you run whatever app you want.
There's no question that RMS resents the fact that Linus gets more credit than he deserves. Just the same, people forget history: Linux only took off because of the AT&T/BSD lawsuit of the early 90s; in 1992, Linux was way behind the BSDs. If Linux (the kernel) comes under severe legal attack (that is, if SCO actually manages to legally impede the use of the Linux kernel), it's completely proper to point out that the GNU folks have a backup plan.
However, the realistic backup plan is not the Hurd, which as far as I'm concerned was a huge botch, which after more than ten years hasn't reached the level of usability that the Linux kernel had after two. It's Debian GNU/BSD. With a bit more work, anyone using this system won't be able to tell that they aren't using a Debian GNU/Linux system, because every single application will work the same way.
You misunderstand what RMS means when he says that the GPL only adds rights, it does not take any rights away. RMS is absolutely correct in a legal sense. That's because if he just wrote "Copyright FSF, all rights reserved", the default situation is that of copyright law, which says that no one but the copyright holder may make copies or derivative works. If you have a legal copy you can run it, make one backup, study how it works, and so forth. These are the kinds of things that shrinkwrap licenses sometimes restrict, but copyright law does not restrict.
The reason it's often called a tax is that paying that $50 is not optional in most cases (especially for laptops), even if you don't want to use the OS at all.
As you say, it might not be worth the hassle to most people to figure out how not to pay that $50.
Intel tunes heavily to score well on important benchmakrks, while gcc doesn't. So, while on average icc generates faster code, the ratio is much smaller than SPEC ratings would suggest, and there are some anomalies like the one you mention, where one compiler or the other generates really slow code.
In most of Europe, the patent doesn't cover software implementations, only hardware, if I understand the rules correctly.
Right, but as most nations don't recognize software patents, that doesn't matter as much as you might think.
Censorship occurs when someone else tells you that you can't say, or see, or read something. There is no censorship when you make a personal decision not to watch a commercial, or a nude scene, or whatever else you don't care for.
The US could not mandate right of reply. However, the US could mandate that people can't bring libel actions if they are given a right of reply.
The most well-known filesystem-based version control system is Clearcase. Essentially, the user sets up a view, which has a config spec. The config spec specifies what versions of each directory and file the user wants to see. Checkins and checkouts are done as in CVS, except that directories and file renames are tracked as well.
Lessig proposes to solve this by technical standards, backed by law: the bandwidth has no central authority, but there are power limitations and protocols for the devices so that everyone can "talk".
No, you didn't recently read that. What you read was that the real effective data rate over "g" is more like 20 Mbps, and that in a network with a lot of "b" devices it might be more like ten. The raw speed is still 54 Mbps, the real speed is less because of the signaling overhead. There is similar overhead in 802.11b, you only get about 5 Mbps, and if there are any Bluetooth devices around you'll get a lot less because of interference.
When lots of people are using the same access point, the bottleneck isn't the wireless rate, it's the connection from the base station to the network.
802.11g is designed to interoperate with 802.11b, although the presence of "b" users in the same area does slow "g" down. Still, everyone is confusing effective rate (say, 20 Mbits/sec actually transmitted) with theoretical peak rate (54 Mbit/sec). "b" users are not getting 11 Mbits/sec; if they are lucky they are getting 5, and if they are surfing the web through DSL or cable modem they aren't even getting 2. When lots of people are using the same access point, the bottleneck isn't
In a year or two, most folks will ditch their "b" equipment for "g" and it won't matter.
Google hasn't exactly patented the algorithm for all uses, and no court has determined that the code infringes the patent, and software patents aren't valid in most countries, so it's not clear whether or not there is any compatibility.
It would seem that anyone who uses the code to build a search engine would be infringing, but even that is something that lawyers can argue about.
If and when apple wins, GNU will stand for "GNU's Now Unix".
You'd lose the case, because about $50 is what MS charges companies like Dell per copy when they buy in bulk. Their lawyers would say that this is the fair value, and the rest is markup for the retail distribution system. Pretty high markup, but hardly unprecedented.
In any case, if Microsoft can be accused of dumping for charging $50 for their software, couldn't a case be made against Red Hat for providing free ISO downloads? Isn't that dumping? It costs more than $0 to provide that service, so they are clearly dumping it.
But SCO isn't alleging that Linux is a pirated copy. The analogy would be to rap artists using samples without authorization. If they could find 4,700 stolen lines, then 0.1% of Linux would be copied, and that's like 0.2 seconds of a three-minute song. They allege "hundreds of lines", and have shown 80, though they haven't offered any evidence about the direction of copying.
Let's not lose context here. The Linux kernel is roughly 4.7 million lines of code. If the duplicated code is in a device driver, it could be because the same person contributed a device driver for the same device to both OSes. There can be any number of other explanations.
No, Eric did not coin the term "open source", Brett. He was at the meeting where the term was suggested, but it wasn't him.
Nice round numbers that are powers of ten are not "barriers".
In any given area (other than, say, Manhattan) there will be many unused television channels. Just allow wireless networking equipment to use those channels. Any of several techniques could be used to avoid interference: detect a TV broadcast signal, or just use a database ("I'm in Phoenix, therefore channels x, y, and z are free").