There are good arguments to use the GPL, even for business-minded folks who disagree with RMS's idea that all software must be free.
The GPL is quite usable for making money. Consider TrollTech, the providers of QT. Their use of the GPL/QPL combo (though the GPL alone would work just as well) got the product of a small Norwegian company known all over the world, and the fact that proprietary developers don't like the GPL means that Troll Tech can sell developer licenses at significant cost.
BSD fans who accuse GPL fans of communism have it backwards: giving work away completely for any purpose, from each according to his ability, to each according to his need, is a hallmark of the BSD license. As Russ Nelson (of OSI) has said, "When I write proprietary software I expect to get paid". By choosing the GPL instead of a non-copyleft license, the author is in a sense insisting on compensation for his work, either in the form of changes that get contributed back, or
with the possibility of selling someone a license to use terms other than the GPL.
What might really harm the BSD folks is the growing practice by holders of software patents to license those patents for use in GPL code, but not in non-copylefted code. IBM is doing a lot of this: they've made some of their patents available to both the Linux kernel and GCC. I would prefer to see no patents at all, but don't see a way of letting the BSD folks use the patents without effectively giving the patent away to all.
They could say, OK, BSD folks, use of the patent is OK as long as full source is provided for the whole app, but did you notice that these conditions are effectively a copyleft, like the GPL (if you use it you have to give the whole source away)?
Why? So far they've gotten every single challenger to back down, without going to court.
Read all about it.. People with a strong case don't need to go to court because the other side caves. Disputes only reach court when both sides think they have a good shot at winning.
The criteria for whether a license is "free software" or "open source" are identical: the Open Source Definition was originally called the Debian Free Software Guidelines. The cases where the OSI says something is an open source license and the FSF says it's a free software license are boundary cases, much like two baseball umpires might rule the same pitch a strike or a ball if it's right on the line. The OSI says the GPL and the LGPL are open source licenses, and the FSF says that the BSD license is a free software license.
It is claimed by the FSF that the fact that SCO shipped the Linux kernel to their customers under the GPL means that they have given their customers, and by extension the general public, license to use everything in it under the GPL, even if some of the code was originally theirs.
Certainly as copyright owner you can give people additional permissions beyond what is stated in the license text, provided that you own the entire code and haven't accepted patches from anyone else.
That's the whole point: this is a technique to catch your staff when they do something that they are not supposed to (like violate the privacy of celebrities). These things don't stop at shits and giggles: if someone gets at a rich man's financial records, that someone can then engage in identity theft, or sale of embarrassing gossip to tabloids, or fanboy/fangirl invasion of a sick person's hospital room, etc.
A woman I knew since grade school got a job in American Express's credit card clearance center in New York in the 80s. Seems some of her colleagues would spend hours every day looking for celebrities' credit card records just so they could gossip
(back then the system was less automated than it is now, and an actual person would need to verify that a charge was valid in many more cases than today). I'm sure that this kind of thing still happens.
Putting in bogus entries with celebrity names would catch such people, as they are only supposed to consult a record when someone tries to use the card in question.
Courts typically craft injunctions in the narrowest possible way. An injunction might forbid the distribution of a Linux kernel containing disputed code, but I can't imagine a judge refusing to allow Linux to be distributed after the disputed code is removed.
Linux is 4.7 million lines of code the last time I checked (2.4.x for some x). The largest piece of code SCO identified to the reviewers as possibly copied was an 80-line segment.
In the case of patent law, not knowing means that damages are 1/3 of what they are if you know (if you know you are infringing a patent, damages triple).
So yes, it is better not to go out actively looking for patents that you might be infringing except in special cases.
The H1B program is not an example of anti-protectionism. Without any trade barriers at all, the employment situation in the US would be like that within the EU boundaries: a programmer from Portugal can get a job in Germany with the same rights as a German worker. Under H1B, an Indian programmer does not have the same rights as a US programmer; he is basically an indentured servant, who must accept any conditions his employer imposes or face immediate deportation.
The argument for H1B is the claim that there is a shortage of skilled technology workers in the US. At present, there is not a shortage, except in very limited cases. However, many companies prefer H1B workers to citizen or permanent resident workers, because they can drive them harder and pay them less, holding the threat of being sent back to India or China in reserve.
Does this mean that Netscape (rather, AOL-Time-Warner) is withdrawing its support? Will they still be providing facilities, network connectivity, etc. or will the Mozilla Foundation have to raise all that on its own? Will Netscape be providing any money to the Mozilla Foundation?
The NY Times has serious flaws (as well as serious virtues). Conservatives usually scream about Jayson Blair; liberals on the other hand are appalled at
Judith Miller, who seemed to be some kind of mole
working for Ahmed Chalabi, the convicted embezzler who wants to run Iraq.
But as every other newspaper and television station in the US basically just reprints and rewrites what the the Times writes, you're pretty much stuck with them. Small newspapers and local TV stations don't have the staff to find non-local news on their own, so they just read the Times and the Washington Post and print what they print.
the article is incorrect about hyperthreading
on
Linux v2.6 Begins Testing
·
· Score: 2, Interesting
It's not true, as the article claims, that making one process look like two doesn't buy you much. The reason is that cache misses are getting more and more expensive: without hyperthreading, a cache miss might cause the processor to wait a hundred cycles.
With hyperthreading, we simply switch to the other process, and pay a far smaller cost.
In particular, the Debian non-us site might have to moved to a country that is not in the European Union (so they can continue to distribute multimedia apps that infringe software patents).
They didn't just distribute in good faith something they got from elsewhere. Their engineers worked on it; they sold it; they explained the implications of the GPL to their customers. They knew what they signed up for. Worse, they used Linux money to buy Unix! That is, they used IPO cash they raised on the basis of being a Linux company, to buy the original Unix code.
In the current Debian vs FSF fights, the issue is that the Debian folks think that the FSF isn't being pure enough. They object to the Gnu Free Documentation License as not being sufficiently free, and are threatening to remove some of the FSF manuals from their distro.
Microsoft didn't develop their kernel from zero with nothing but GNU tools, in a way that no other tools will even work. The Linux kernel is totally dependent on GCC. The only existing C library for Linux is GNU. In the case of Microsoft shipping the GNU toolchain, it's an add-on; in the Linux case, it's a vital component, without which the machine is useless.
I don't understand your argument about the GNU tools not counting. When you write a program, it doesn't touch the kernel directly; it gets all its services from language support libraries (the C and C++ libraries). These are all GNU. Similarly, without the GNU tools you don't have a usable system.
But in answer to some of your other questions: the FSF provided key funding to get the Debian distribution launched. Eben Moglen, the FSF lawyer, has worked behind the scenes to pressure more than a hundred different GPL violators to comply with the license, and has helped many people who didn't assign their copyright to the FSF deal with legal issues (notably the mysql folks).
The Debian folks are working on a distribution that runs the GNU C library and Debian packages on top of a BSD kernel. You won't be able to tell, when using such a system, that you aren't using a Linux-based system, because all the apps will work just the same.
You are right that the FSF tools aren't designed specifically for Linux, they are key infrastructure for all the BSD systems as well, and also Apple's OSX.
There was a time when Red Hat/Cygnus folks were doing the vast majority of work on GCC, but for several years now the release manager and several of the most important contributors work for a company called CodeSourcery. SuSE also employs several vital contributors, and Apple has gotten very active lately (donating support for precompiled headers, which is planned for the 3.4 release). Of course Red Hat still contributes a lot, but they are nowhere near as dominant as they used to be.
It's not about the best time to show ads, it's about the ads. If you're a heterosexual man, aren't you going to watch the beer commercial with the wrestling babes? Over and over?:-)
The problem is that any such scanning will be highly visible, and it will be easy to arrange that there are 1000 honeypot ipv6 addresses for every real address. So anyone who tries to scan will only find fake systems.
Re:Interesting technology
on
RFID Explained
·
· Score: 1
If a tag costs 5 cents and is 1/3 millimeter across, removing it from the product and inserting it into another product is going to cost a lot more than 5 cents (if a $6/hour worker requires 30 seconds to remove the tag and put it in another product, that's a nickel right there).
You want to have vastly more addresses than can ever be used; this will kill scanning attacks by black hats and spammers who just try every network address looking for a victim. Anyone scanning thousands of bogus addresses for every real one will trigger all kinds of alarms.
Slashdot Mirror
There are good arguments to use the GPL, even for business-minded folks who disagree with RMS's idea that all software must be free.
The GPL is quite usable for making money. Consider TrollTech, the providers of QT. Their use of the GPL/QPL combo (though the GPL alone would work just as well) got the product of a small Norwegian company known all over the world, and the fact that proprietary developers don't like the GPL means that Troll Tech can sell developer licenses at significant cost.
BSD fans who accuse GPL fans of communism have it backwards: giving work away completely for any purpose, from each according to his ability, to each according to his need, is a hallmark of the BSD license. As Russ Nelson (of OSI) has said, "When I write proprietary software I expect to get paid". By choosing the GPL instead of a non-copyleft license, the author is in a sense insisting on compensation for his work, either in the form of changes that get contributed back, or with the possibility of selling someone a license to use terms other than the GPL.
What might really harm the BSD folks is the growing practice by holders of software patents to license those patents for use in GPL code, but not in non-copylefted code. IBM is doing a lot of this: they've made some of their patents available to both the Linux kernel and GCC. I would prefer to see no patents at all, but don't see a way of letting the BSD folks use the patents without effectively giving the patent away to all. They could say, OK, BSD folks, use of the patent is OK as long as full source is provided for the whole app, but did you notice that these conditions are effectively a copyleft, like the GPL (if you use it you have to give the whole source away)?
Why? So far they've gotten every single challenger to back down, without going to court.
Read all about it.. People with a strong case don't need to go to court because the other side caves. Disputes only reach court when both sides think they have a good shot at winning.
The criteria for whether a license is "free software" or "open source" are identical: the Open Source Definition was originally called the Debian Free Software Guidelines. The cases where the OSI says something is an open source license and the FSF says it's a free software license are boundary cases, much like two baseball umpires might rule the same pitch a strike or a ball if it's right on the line. The OSI says the GPL and the LGPL are open source licenses, and the FSF says that the BSD license is a free software license.
It is claimed by the FSF that the fact that SCO shipped the Linux kernel to their customers under the GPL means that they have given their customers, and by extension the general public, license to use everything in it under the GPL, even if some of the code was originally theirs.
Certainly as copyright owner you can give people additional permissions beyond what is stated in the license text, provided that you own the entire code and haven't accepted patches from anyone else.
That's the whole point: this is a technique to catch your staff when they do something that they are not supposed to (like violate the privacy of celebrities). These things don't stop at shits and giggles: if someone gets at a rich man's financial records, that someone can then engage in identity theft, or sale of embarrassing gossip to tabloids, or fanboy/fangirl invasion of a sick person's hospital room, etc.
A woman I knew since grade school got a job in American Express's credit card clearance center in New York in the 80s. Seems some of her colleagues would spend hours every day looking for celebrities' credit card records just so they could gossip (back then the system was less automated than it is now, and an actual person would need to verify that a charge was valid in many more cases than today). I'm sure that this kind of thing still happens. Putting in bogus entries with celebrity names would catch such people, as they are only supposed to consult a record when someone tries to use the card in question.
IBM has also licensed some patents covering register allocation so that GCC can use the techniques.
Raph Levien, who currently maintains Ghostscript, has also licensed a number of patents dealing with printing for use in GPLed code.
Courts typically craft injunctions in the narrowest possible way. An injunction might forbid the distribution of a Linux kernel containing disputed code, but I can't imagine a judge refusing to allow Linux to be distributed after the disputed code is removed.
Linux is 4.7 million lines of code the last time I checked (2.4.x for some x). The largest piece of code SCO identified to the reviewers as possibly copied was an 80-line segment.
In the case of patent law, not knowing means that damages are 1/3 of what they are if you know (if you know you are infringing a patent, damages triple). So yes, it is better not to go out actively looking for patents that you might be infringing except in special cases.
The H1B program is not an example of anti-protectionism. Without any trade barriers at all, the employment situation in the US would be like that within the EU boundaries: a programmer from Portugal can get a job in Germany with the same rights as a German worker. Under H1B, an Indian programmer does not have the same rights as a US programmer; he is basically an indentured servant, who must accept any conditions his employer imposes or face immediate deportation.
The argument for H1B is the claim that there is a shortage of skilled technology workers in the US. At present, there is not a shortage, except in very limited cases. However, many companies prefer H1B workers to citizen or permanent resident workers, because they can drive them harder and pay them less, holding the threat of being sent back to India or China in reserve.
Does this mean that Netscape (rather, AOL-Time-Warner) is withdrawing its support? Will they still be providing facilities, network connectivity, etc. or will the Mozilla Foundation have to raise all that on its own? Will Netscape be providing any money to the Mozilla Foundation?
If you think that PKD's writings are "strongly rationalist", you haven't read much of him.
The NY Times has serious flaws (as well as serious virtues). Conservatives usually scream about Jayson Blair; liberals on the other hand are appalled at Judith Miller, who seemed to be some kind of mole working for Ahmed Chalabi, the convicted embezzler who wants to run Iraq.
But as every other newspaper and television station in the US basically just reprints and rewrites what the the Times writes, you're pretty much stuck with them. Small newspapers and local TV stations don't have the staff to find non-local news on their own, so they just read the Times and the Washington Post and print what they print.
It's not true, as the article claims, that making one process look like two doesn't buy you much. The reason is that cache misses are getting more and more expensive: without hyperthreading, a cache miss might cause the processor to wait a hundred cycles. With hyperthreading, we simply switch to the other process, and pay a far smaller cost.
In particular, the Debian non-us site might have to moved to a country that is not in the European Union (so they can continue to distribute multimedia apps that infringe software patents).
They didn't just distribute in good faith something they got from elsewhere. Their engineers worked on it; they sold it; they explained the implications of the GPL to their customers. They knew what they signed up for. Worse, they used Linux money to buy Unix! That is, they used IPO cash they raised on the basis of being a Linux company, to buy the original Unix code.
In the current Debian vs FSF fights, the issue is that the Debian folks think that the FSF isn't being pure enough. They object to the Gnu Free Documentation License as not being sufficiently free, and are threatening to remove some of the FSF manuals from their distro.
Microsoft didn't develop their kernel from zero with nothing but GNU tools, in a way that no other tools will even work. The Linux kernel is totally dependent on GCC. The only existing C library for Linux is GNU. In the case of Microsoft shipping the GNU toolchain, it's an add-on; in the Linux case, it's a vital component, without which the machine is useless.
I don't understand your argument about the GNU tools not counting. When you write a program, it doesn't touch the kernel directly; it gets all its services from language support libraries (the C and C++ libraries). These are all GNU. Similarly, without the GNU tools you don't have a usable system.
But in answer to some of your other questions: the FSF provided key funding to get the Debian distribution launched. Eben Moglen, the FSF lawyer, has worked behind the scenes to pressure more than a hundred different GPL violators to comply with the license, and has helped many people who didn't assign their copyright to the FSF deal with legal issues (notably the mysql folks).
The Debian folks are working on a distribution that runs the GNU C library and Debian packages on top of a BSD kernel. You won't be able to tell, when using such a system, that you aren't using a Linux-based system, because all the apps will work just the same.
You are right that the FSF tools aren't designed specifically for Linux, they are key infrastructure for all the BSD systems as well, and also Apple's OSX.
There was a time when Red Hat/Cygnus folks were doing the vast majority of work on GCC, but for several years now the release manager and several of the most important contributors work for a company called CodeSourcery. SuSE also employs several vital contributors, and Apple has gotten very active lately (donating support for precompiled headers, which is planned for the 3.4 release). Of course Red Hat still contributes a lot, but they are nowhere near as dominant as they used to be.
It's not about the best time to show ads, it's about the ads. If you're a heterosexual man, aren't you going to watch the beer commercial with the wrestling babes? Over and over? :-)
The problem is that any such scanning will be highly visible, and it will be easy to arrange that there are 1000 honeypot ipv6 addresses for every real address. So anyone who tries to scan will only find fake systems.
If a tag costs 5 cents and is 1/3 millimeter across, removing it from the product and inserting it into another product is going to cost a lot more than 5 cents (if a $6/hour worker requires 30 seconds to remove the tag and put it in another product, that's a nickel right there).
You want to have vastly more addresses than can ever be used; this will kill scanning attacks by black hats and spammers who just try every network address looking for a victim. Anyone scanning thousands of bogus addresses for every real one will trigger all kinds of alarms.