I read some of your links, and, while I will admit that there are certainly problems with how software is built, I don't think you have a silver bullet.
If pictures are so good, how come a lot of hardware is written in Verilog/VHDL instead of schematics, when both are available?
If fine-grained parallelism is so error-free, how come chip companies spend a fortune verifying and validating both their schematics and their HDL before they tape out?
BTW, I make my living by writing software, writing Verilog, and drawing schematics.
FWIW, I prefer Python to Verilog, and Verilog to schematics. In fact, I check my schematic netlists with Python scripts to make sure there are no errors.
When it's important to get it right, I find text much preferable to schematics -- it can be easily diffed and version controlled.
Ignoring for a moment whether software patents are valuable (and, of course, most of them are crap):
1) If I have a patent, and don't do any GPLv3 contribution or distribution, and some GPLv3 code infringes my patent, in your opinion, would I be able to successfully sue people (or at least, large companies) who infringe my patent via use of the GPLv3 software? If not, please state why not and skip question 2. If so, go to question 2.
2) You imply that slipping this infringing code into a GPLv3 package and then tricking the patent holder into automatically distributing it "does not reduce their rights to sue, because they don't have them in the first place." I'm not sure what you mean. They would have had rights if they weren't helping out the community with a mirror. Do you believe distributing ANY GPLv3 code automatically removes your right to sue over infringement of ANY patent, or what?
So let's modify the scenario a bit and see how it fits:
Assume that party A (MIT in the original discussion) has a really valuable patent, and as a service to the community, performs automatic copying of open source projects (mirroring).
Renegade party B slips some infringing code into one of the projects, and party A automatically picks it up and distributes it. Party B gets it on the redistribution and starts disseminating it widely, claiming they have a license for the patent from party A.
So far, this is the original scenario. Now, according to your addition, party A sues party B for damages, for putting the infringing invention in the code in the first place. And I agree that there may be a good chance that party A could win a suit against party B for this.
But, according to a literal reading of the GPL, parties C, D, E, F... who also downloaded the code (written by party B) from party A, directly or indirectly, have now been granted licenses from party A for use of that patent in that code, and they don't suffer from the "unclean hands" of party B.
If party B does not have enough resources to make party A whole for the infringement (basically for letting the entire world use the patent), party A still loses, unless you can show that the last paragraph of section 10 is somehow undone by the suit filed by party A against party B.
So upon reflection, I agree that the unattractive prospect of being sued for the initial infringement makes the deliberate hack I described legally very risky, but if party C is profiting handsomely off the patented code, and party A cannot show any connection between party B and party C, it seems like this clause in the license might be able to provide income for trial lawyers for years.
And I've made it clear multiple times that I'm talking about section 10.
Now, maybe there is some reason why the last paragraph of section 10 doesn't apply to this scenario. If so, I'd love to hear it, and then I can go back to sleep.
But, absent somebody actually addressing my question about section 10, the fact that everybody keeps dancing around it and claiming that I'm talking about section 11 instead of 10 just makes it look like FUD from a different direction.
Section 11 says if you contribute code with your own patent, somebody cannot add code which implements another of your patents, redistribute it, and automatically grant someone else rights to your patent that you didn't. That's fair, but if you'll read my original quote carefully, that's not what I'm addressing.
Section 10 refers to people who "convey" the software, not "contributors", and seems to say that people who "convey" cannot sue downstream users for infringement of any patent implemented by the software. The definition of "convey" is quite broad, and it seems to me that it could cover the scenario I discussed.
IANAL either, and there could be some reason why this isn't an issue, but the issue in question has to do with section 10, not section 11.
In section 11, the patent section, what you say seems to be true. It talks about patents of "contributors".
But section 10 talks about the responsibilities of people who "convey" the software, who are not necessarily contributors, and states "For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License, and you may not initiate litigation (including a cross-claim or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling, offering for sale, or importing the Program or any portion of it."
Now, maybe that really only affects contributors, but then it would seem to be both in the wrong place (not under section 11) and slightly redundant to the statements in section 11.
Now, IANAL, and there's probably something really basic I'm missing that would prevent this hack from working, but I'll throw it out there as food for thought:
Some universities have a lot of patents and some of them offer free mirrors for things like kernel.org and sourceforge.net projects. It may be that the act of offering a mirror is protected under the DMCA safe harbor, but if copyright license law is as powerful as some GPLv3 proponents claim, it's not even clear that the DMCA safe harbor would override section 10 of the GPL. In any case, some mirrors work by pulling code rather than letting code be pushed, so that seems like an affirmative act of copying the software and then creating and giving copies to the general public, so an entity operating a mirror might be "conveying" under the GPL.
So, for example, if MIT has a patent I want to use, maybe all I have to do is get committer rights to some relevant project, code up something which infringes the patent, get the patch accepted (never mentioning the patent, of course), and it gets distributed to all the mirrors, including MIT's.
I download it from MIT, and voila! I have a license to use that patent inside that program (and apparently inside any GPLed derivatives I make of that program. Being the proprietary sort of guy I am, I wrap the GPL project's code with another completely proprietary program which controls it and lets the GPLed code do the patented dirty work.
I don't know whether this would work or not, but I'm starting to understand why companies are now marketing "open source" license scrubbers.
The FSF is certainly free to do this with the GPL. But while the consequences of just distributing source under v2 might have been intended to convey patents in this same way, a lot of people didn't realize that because the wording there is not as clear, and the remedies don't appear to be as onerous. V3 section 10 seems to make it very clear that if you convey code which implements a patented invention, you cannot sue anybody over using that invention in that code, and that "convey" would cover the act of proactively operating a mirror site.
This should give pause to a lot of people, not just Microsoft. Right now, "everybody knows" that GPL2 is a safe license, in the same category as BSD, well away from the category of any proprietary license, for being able to freely redistribute source code.
Those who assume the same about GPLv3 do so at their own peril, perhaps to their own detriment. It appears that, for an entity with a valuable patent, inadvertently distributing one copy of GPLv3 software could easily be much more costly than inadvertently distributing a few hundred copies of a Microsoft product.
The way universities work, it is unlikely that the legal counsel stays on top of things like kernel.org mirrors, but it seems that anybody with a patent portfolio who is running a free software mirror of any type ought to take a serious look at their policies and at the terms of GPLv3.
Perhaps one valid component of licensing strategy would be to repudiate GPL v3 (and any similar licenses which purport to appropriate your own patents), just like Microsoft has done. That would basically be a public announcement that, if anybody catches you distributing GPL v3 code, please let you know right away because it is not your intention to ever do so and you will stop distributing immediately, and if anybody thinks they're getting one of your patents out of the deal, you plan on fighting it every inch of the way.
I should have realized how many Nielsen/netrating fanboys there were on Slashdot, and been much more careful to choose my language so as not to inflame that adoring community.
They are practically shouting to the world that they don't have a clue how to pick the right site to advertise on (or that if they ever did have a clue, they just lost it).
If that's the real reason why these guys are blowing themselves up, it seems like the right kind of publicity could fix this whole problem.
We just need to explain why these women are virgins.
Even though down on earth, the male/female ratio was heavily skewed in the females' favor,
these chicks are so ugly, fat, and deformed they couldn't get laid on earth, and have to wait
in heaven for someone desperate enough to take them to show up.
The reason they can afford to give everybody who shows up 72 of them is that nobody ever
takes them up on the offer! Some of these virgins have been offered to over 30 suicide
bombers.
So, dear bomber, when you get to heaven, please remember that these poor women have
feelings too, and do your duty and give all of 'em all the sex they want and need.
Allah will thank you mightily for this, because some of them are starting to come
on to him on a regular basis, and it really creeps him out.
They have hired Chris Lattner, one of the lead developers of LLVM, which is an optimizing compiler. The bulk of LLVM is released under a BSD-like license, but for now it uses a front end borrowed from GCC.
But, that would be a highly application dependent, highly speed of transfer of VM dependent, highly CPU and disk speed dependent, completely useless statistic.
You're right, never mind, just remembered I'm on Slashdot.
I dunno, seems like a great way to frame somebody to me.
If you purchase music from iTunes, and somebody who doesn't like you knows a few of the tracks you bought, it seems they could create a fake chain of "provenance" which most judges in this country would agree proved that you violated copyrights.
No, your argument is as disingenuous as any old argument about "why should I care about if big brother is watching? I'm not doing anything wrong."
Actually, the Groklaw article implies that if party A publishes some code under the BSD, and then party B incorporates that into some proprietary code and republishes it, then any party C receiving code from party B is clear to redistribute it, because the original conditions from party A must apply to all the software distributed by party B. (As many here have pointed out, the article implies that the license is as viral as the GPL.)
So, for example, if Microsoft includes BSD code in windows (search for Greg Roelofs in that link), the article implies that it might be perfectly legal for you, me, and anybody else to start pirating WIndows.
Believe me when I say that this would cause a massive lawsuit (or even criminal charges), which would in no way involve the original author Greg Roelofs...
Slashdot Mirror
I got no mod-points...
I read some of your links, and, while I will admit that there are certainly problems with how software is built, I don't think you have a silver bullet.
If pictures are so good, how come a lot of hardware is written in Verilog/VHDL instead of schematics, when both are available?
If fine-grained parallelism is so error-free, how come chip companies spend a fortune verifying and validating both their schematics and their HDL before they tape out?
BTW, I make my living by writing software, writing Verilog, and drawing schematics.
FWIW, I prefer Python to Verilog, and Verilog to schematics. In fact, I check my schematic netlists with Python scripts to make sure there are no errors.
When it's important to get it right, I find text much preferable to schematics -- it can be easily diffed and version controlled.
If you'd RTFA (I know, I know, this is Slashdot) you would have found out that, all too often, Tasers FIRE lethal rounds...
"nanotubes on the lungs of rats produced more toxic response than quartz dust"
I dunno.
Electrical engineers build weapons systems.
Civil engineers build targets.
But, I guess if they take out your target, you get paid to build another, so maybe so...
Ignoring for a moment whether software patents are valuable (and, of course, most of them are crap):
1) If I have a patent, and don't do any GPLv3 contribution or distribution, and some GPLv3 code infringes my patent, in your opinion, would I be able to successfully sue people (or at least, large companies) who infringe my patent via use of the GPLv3 software? If not, please state why not and skip question 2. If so, go to question 2.
2) You imply that slipping this infringing code into a GPLv3 package and then tricking the patent holder into automatically distributing it "does not reduce their rights to sue, because they don't have them in the first place." I'm not sure what you mean. They would have had rights if they weren't helping out the community with a mirror. Do you believe distributing ANY GPLv3 code automatically removes your right to sue over infringement of ANY patent, or what?
Thanks,
pem
So let's modify the scenario a bit and see how it fits:
Assume that party A (MIT in the original discussion) has a really valuable patent, and as a service to the community, performs automatic copying of open source projects (mirroring).
Renegade party B slips some infringing code into one of the projects, and party A automatically picks it up and distributes it. Party B gets it on the redistribution and starts disseminating it widely, claiming they have a license for the patent from party A.
So far, this is the original scenario. Now, according to your addition, party A sues party B for damages, for putting the infringing invention in the code in the first place. And I agree that there may be a good chance that party A could win a suit against party B for this.
But, according to a literal reading of the GPL, parties C, D, E, F... who also downloaded the code (written by party B) from party A, directly or indirectly, have now been granted licenses from party A for use of that patent in that code, and they don't suffer from the "unclean hands" of party B.
If party B does not have enough resources to make party A whole for the infringement (basically for letting the entire world use the patent), party A still loses, unless you can show that the last paragraph of section 10 is somehow undone by the suit filed by party A against party B.
So upon reflection, I agree that the unattractive prospect of being sued for the initial infringement makes the deliberate hack I described legally very risky, but if party C is profiting handsomely off the patented code, and party A cannot show any connection between party B and party C, it seems like this clause in the license might be able to provide income for trial lawyers for years.
And everybody says I'm talking about section 11.
And I've made it clear multiple times that I'm talking about section 10.
Now, maybe there is some reason why the last paragraph of section 10 doesn't apply to this scenario. If so, I'd love to hear it, and then I can go back to sleep.
But, absent somebody actually addressing my question about section 10, the fact that everybody keeps dancing around it and claiming that I'm talking about section 11 instead of 10 just makes it look like FUD from a different direction.
Section 11 says if you contribute code with your own patent, somebody cannot add code which implements another of your patents, redistribute it, and automatically grant someone else rights to your patent that you didn't. That's fair, but if you'll read my original quote carefully, that's not what I'm addressing.
Section 10 refers to people who "convey" the software, not "contributors", and seems to say that people who "convey" cannot sue downstream users for infringement of any patent implemented by the software. The definition of "convey" is quite broad, and it seems to me that it could cover the scenario I discussed.
IANAL either, and there could be some reason why this isn't an issue, but the issue in question has to do with section 10, not section 11.
In section 11, the patent section, what you say seems to be true. It talks about patents of "contributors".
But section 10 talks about the responsibilities of people who "convey" the software, who are not necessarily contributors, and states "For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License, and you may not initiate litigation (including a cross-claim or counterclaim in a lawsuit) alleging that any patent claim is infringed by making, using, selling, offering for sale, or importing the Program or any portion of it."
Now, maybe that really only affects contributors, but then it would seem to be both in the wrong place (not under section 11) and slightly redundant to the statements in section 11.
Now, IANAL, and there's probably something really basic I'm missing that would prevent this hack from working, but I'll throw it out there as food for thought:
Some universities have a lot of patents and some of them offer free mirrors for things like kernel.org and sourceforge.net projects. It may be that the act of offering a mirror is protected under the DMCA safe harbor, but if copyright license law is as powerful as some GPLv3 proponents claim, it's not even clear that the DMCA safe harbor would override section 10 of the GPL. In any case, some mirrors work by pulling code rather than letting code be pushed, so that seems like an affirmative act of copying the software and then creating and giving copies to the general public, so an entity operating a mirror might be "conveying" under the GPL.
So, for example, if MIT has a patent I want to use, maybe all I have to do is get committer rights to some relevant project, code up something which infringes the patent, get the patch accepted (never mentioning the patent, of course), and it gets distributed to all the mirrors, including MIT's.
I download it from MIT, and voila! I have a license to use that patent inside that program (and apparently inside any GPLed derivatives I make of that program. Being the proprietary sort of guy I am, I wrap the GPL project's code with another completely proprietary program which controls it and lets the GPLed code do the patented dirty work.
I don't know whether this would work or not, but I'm starting to understand why companies are now marketing "open source" license scrubbers.
The FSF is certainly free to do this with the GPL. But while the consequences of just distributing source under v2 might have been intended to convey patents in this same way, a lot of people didn't realize that because the wording there is not as clear, and the remedies don't appear to be as onerous. V3 section 10 seems to make it very clear that if you convey code which implements a patented invention, you cannot sue anybody over using that invention in that code, and that "convey" would cover the act of proactively operating a mirror site.
This should give pause to a lot of people, not just Microsoft. Right now, "everybody knows" that GPL2 is a safe license, in the same category as BSD, well away from the category of any proprietary license, for being able to freely redistribute source code.
Those who assume the same about GPLv3 do so at their own peril, perhaps to their own detriment. It appears that, for an entity with a valuable patent, inadvertently distributing one copy of GPLv3 software could easily be much more costly than inadvertently distributing a few hundred copies of a Microsoft product.
The way universities work, it is unlikely that the legal counsel stays on top of things like kernel.org mirrors, but it seems that anybody with a patent portfolio who is running a free software mirror of any type ought to take a serious look at their policies and at the terms of GPLv3.
Perhaps one valid component of licensing strategy would be to repudiate GPL v3 (and any similar licenses which purport to appropriate your own patents), just like Microsoft has done. That would basically be a public announcement that, if anybody catches you distributing GPL v3 code, please let you know right away because it is not your intention to ever do so and you will stop distributing immediately, and if anybody thinks they're getting one of your patents out of the deal, you plan on fighting it every inch of the way.
I should have realized how many Nielsen/netrating fanboys there were on Slashdot, and been much more careful to choose my language so as not to inflame that adoring community.
They are practically shouting to the world that they don't have a clue how to pick the right site to advertise on (or that if they ever did have a clue, they just lost it).
Try to take over the world!
We just need to explain why these women are virgins.
Even though down on earth, the male/female ratio was heavily skewed in the females' favor, these chicks are so ugly, fat, and deformed they couldn't get laid on earth, and have to wait in heaven for someone desperate enough to take them to show up.
The reason they can afford to give everybody who shows up 72 of them is that nobody ever takes them up on the offer! Some of these virgins have been offered to over 30 suicide bombers.
So, dear bomber, when you get to heaven, please remember that these poor women have feelings too, and do your duty and give all of 'em all the sex they want and need.
Allah will thank you mightily for this, because some of them are starting to come on to him on a regular basis, and it really creeps him out.
They have hired Chris Lattner, one of the lead developers of LLVM, which is an optimizing compiler. The bulk of LLVM is released under a BSD-like license, but for now it uses a front end borrowed from GCC.
But Apple is busy fixing that, too.
But, that would be a highly application dependent, highly speed of transfer of VM dependent, highly CPU and disk speed dependent, completely useless statistic.
You're right, never mind, just remembered I'm on Slashdot.
Umm, assuming zero overhead, even with 10Gb ethernet, you could only move around 7.1 MB in 6 milliseconds.
With Gigabit Ethernet, you can divide that down to well under a megabyte.
While I'm sure your virtual machines are nice toys, they ARE just toys.
I dunno, seems like a great way to frame somebody to me.
If you purchase music from iTunes, and somebody who doesn't like you knows a few of the tracks you bought, it seems they could create a fake chain of "provenance" which most judges in this country would agree proved that you violated copyrights.
No, your argument is as disingenuous as any old argument about "why should I care about if big brother is watching? I'm not doing anything wrong."
...can be an invaluable search term.
So, for example, if Microsoft includes BSD code in windows (search for Greg Roelofs in that link), the article implies that it might be perfectly legal for you, me, and anybody else to start pirating WIndows.
Believe me when I say that this would cause a massive lawsuit (or even criminal charges), which would in no way involve the original author Greg Roelofs...