Because it does not require the release of private certificates to comply with the anti-tivoization language.
If you said something about how the company would have less control over the hardware, so they are reluctant anyway, instead of offering that link as an incorrect answer to the question, then you would not receive downvotes.
I mean, only if control of that private certificate is required to run code on the device. I'm sad to live in a world where allowing people to run the software of their own choice on their own computers is considered a "huge no-no."
The problem is physics. You have a bunch of people in fairly close geographic area transmitting to the same point in space on the same frequency band. One bad actor can interfere with everybody else's tx.
> What if people die because you cannot modify that code?
I think it's far better to trust unmodified safety critical code that's been tested than to trust Joe Blow's precariously modified braking code that allows his Honda Civic to drift around corners better.
> How about the software in the micro controllers in my car? Should I be allowed to modify that code too (if I mess up people could die)?
You're allowed to modify any other part of the car. Why should software be different?
> How about the software in my satellite internet radio transceiver (one rogue transceiver could take down the entire beam/jam the satellite)?
There can be some flexibility for the part of the code actually doing the transceiving.
Everything on top should be replaceable.
And radio is really a special case, because it's a shared commons that requires a lot of expertise and where one bad transmitter can cause problems for miles or more.
Its not about "their own computer". Its about consumer "devices", including ones where control should not be given to the consumer for whatever reason. e.g. electric cars or medical devices.
Think of what would happen if anyone was allowed to run their own AutoPilot software on public roads.
Stallman is a radical and he doesn't pay attention to real business needs of large corporations, which is why they all avoid GPLv3 like the plague.
Requiring that a company provides access to the consumer doesn't preclude the existence of regulations ensuring safe usage.
If a user/team of developers were able to develop their own AI system which could operate to whatever safety constraints are set out by the law, on what grounds would you deny them from running their cars using the AI?
Similarly, we can all execute whatever code we want on our laptops, which also allows us the potential to cause harm by attacking other systems, but doing so would still be illegal.
>Its not about "their own computer". Its about consumer "devices", including ones where control should not be given to the consumer for whatever reason. e.g. electric cars or medical devices.
Why should the owners of these devices not be allowed to control them? What makes them special?
>Think of what would happen if anyone was allowed to run their own AutoPilot software on public roads.
Sure, there are plenty of ways to modify the software in a car in reckless ways that endanger those around you. But it's possible to do a lot of things that endanger those around you: you can drink and drive, drive while tired, install headlights that blind other drivers, etc. We have laws to discourage that, and they probably need updating for our modern world of software-driven cars, but at a certain point we rely on the social norm of not recklessly endangering yourself and others. (And if nothing else, the owner of the car is liable if the modifications make it unsafe.)
This is not to say I think people should be able to do whatever they want on public roads; they're not yours after all, there are plenty of restrictions you have to abide by to use them, and I wouldn't mind having a software review requirement. But the car is yours: you bought it, you should be able to do whatever the heck you want with it otherwise.
This is more a matter for regulation than for copyright, I think. Even if the car's ECU and AI code were under GPLv3, governements could prohibit using non-certified versions of the software on public roads and levy hefty penalties and jail time when drivers are caught using wildcat software...
That's their problem. Refusing to sell your product is probably not the best long-term business strategy. If they want revenue, the will eventually sell. The government already distorts the market in their favor by granting them an exclusive copyright. Those rights do not include being free from the market forces like supply & demand.
> anyone can record it in full HD with an app on their TV
Why would anybody do that, when its easier to just download a torrent? DRM doesn't stop piracy, because "stopping piracy" has always been the PR-excuse. DRM is an attempt to usurp property rights, with a manufacturer retaining control over your device after the first sale.
Still is (all that 'licensed not sold' bs aside) but the manufacturer isn't obliged to support your activities (outside of right-to-repair type legislation). Just like they're not obliged to give you engineering drawings or internal specs or to explain principles of operation.
I would be annoyed at the lack of uptake if I were RMS, the GPLv3 is designed to make sure the user has the four essential freedoms, while Apple and Android OEMs (except those permitting bootloader unlocking) are all too happy to trample freedom zero: https://www.gnu.org/philosophy/free-sw.html
Without the ability to run software as we see fit, we are constrained to what these vendors permit us to do on our own devices.
If you like this constraint, GPLv2 and lesser licenses are for you, otherwise a license like GPLv3 is advisable to ensure your users aren't trapped with an outdated, vulnerable version of your software that their OEM refuses to update.
Well, yeah; it depends on how closely the FSF's goals align with your own. Personally, mine are very close to the FSF's, so I'm comfortable releasing code under GPL v3 or later. Yours may not be.
It's worth noting however that it's much easier to go from "v3 or later" to "v3 only" than the other way around.
> It's worth noting however that it's much easier to go from "v3 or later" to "v3 only" than the other way around.
Why so? I always wondered why say 'or later' rather than just adopt v4 (et al.) as dual (...) licences as and when, so I can believe it, I just don't understand it.
If you don't like the new version of the license, you can start releasing your newer code under v3-only. This doesn't stop someone from forking the old version and releasing it under v4, but for the vast majority of projects this isn't an issue; most are small enough that it's uncommon for forks to overtake the original while the latter is still in development.
Hunting down each and every contributor and getting their consent to change licenses or pull their contributions is a lot of effort vs putting “or later” and deciding when later comes whether to shift to the new license.
You can't grandfather in code under a different license. The recommendation when GPLv2 was published was to include the "or any future version" to allow relicensing within the same series of licenses. A lot of people didnt do that, some for lack of understanding and some fully aware.
> Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation.
The license itself admits that it may cover multiple versions. Being clear, as the FSF is recommending on this page, seems like the best choice to avoid confusion for your end users.
(in particular, this page seems to be about deprecating the existing ambiguous SPDX license indicators and replacing them with ones that have clear 'only' or 'later' suffixes. This seems like a good thing.)
>> If the Program specifies a version number of this License which applies to it and "any later version", you have the option...
> The license itself admits that it may cover multiple versions.
... along with the conditions under which this expansive licensing applies. While it's good to be clear, it's not respectful of author's wishes to presume a later license when they didn't explicitly allow that. The article suggests that this case is ambiguous, when it is absolutely not and I find that to be irresponsible
44 comments
[ 545 ms ] story [ 1538 ms ] threadIt seems anyone opposed to GPLv3 is being shushed by the pro-Stallman mobs.
If you said something about how the company would have less control over the hardware, so they are reluctant anyway, instead of offering that link as an incorrect answer to the question, then you would not receive downvotes.
I think it's a little more nuanced than that. Should I be able to run whatever software I want in my printer? Absolutely.
How about the software in my satellite internet radio transceiver (one rogue transceiver could take down the entire beam/jam the satellite)?
How about the software in the micro controllers in my car? Should I be allowed to modify that code too (if I mess up people could die)?
This sounds like a failure on the satellite part. Can you do satellite comms with SDR?
>How about the software in the micro controllers in my car? Should I be allowed to modify that code too (if I mess up people could die)?
What if people die because you cannot modify that code?
> What if people die because you cannot modify that code?
I think it's far better to trust unmodified safety critical code that's been tested than to trust Joe Blow's precariously modified braking code that allows his Honda Civic to drift around corners better.
You're allowed to modify any other part of the car. Why should software be different?
> How about the software in my satellite internet radio transceiver (one rogue transceiver could take down the entire beam/jam the satellite)?
There can be some flexibility for the part of the code actually doing the transceiving.
Everything on top should be replaceable.
And radio is really a special case, because it's a shared commons that requires a lot of expertise and where one bad transmitter can cause problems for miles or more.
Its not about "their own computer". Its about consumer "devices", including ones where control should not be given to the consumer for whatever reason. e.g. electric cars or medical devices.
Think of what would happen if anyone was allowed to run their own AutoPilot software on public roads.
Stallman is a radical and he doesn't pay attention to real business needs of large corporations, which is why they all avoid GPLv3 like the plague.
Nothing would change. You're responsible for what the car does.
If a user/team of developers were able to develop their own AI system which could operate to whatever safety constraints are set out by the law, on what grounds would you deny them from running their cars using the AI?
Similarly, we can all execute whatever code we want on our laptops, which also allows us the potential to cause harm by attacking other systems, but doing so would still be illegal.
Why should the owners of these devices not be allowed to control them? What makes them special?
>Think of what would happen if anyone was allowed to run their own AutoPilot software on public roads.
Sure, there are plenty of ways to modify the software in a car in reckless ways that endanger those around you. But it's possible to do a lot of things that endanger those around you: you can drink and drive, drive while tired, install headlights that blind other drivers, etc. We have laws to discourage that, and they probably need updating for our modern world of software-driven cars, but at a certain point we rely on the social norm of not recklessly endangering yourself and others. (And if nothing else, the owner of the car is liable if the modifications make it unsafe.)
This is not to say I think people should be able to do whatever they want on public roads; they're not yours after all, there are plenty of restrictions you have to abide by to use them, and I wouldn't mind having a software review requirement. But the car is yours: you bought it, you should be able to do whatever the heck you want with it otherwise.
currently, they are? https://comma.ai/
This is more a matter for regulation than for copyright, I think. Even if the car's ECU and AI code were under GPLv3, governements could prohibit using non-certified versions of the software on public roads and levy hefty penalties and jail time when drivers are caught using wildcat software...
DRM is toxic.
That's their problem. Refusing to sell your product is probably not the best long-term business strategy. If they want revenue, the will eventually sell. The government already distorts the market in their favor by granting them an exclusive copyright. Those rights do not include being free from the market forces like supply & demand.
> anyone can record it in full HD with an app on their TV
Why would anybody do that, when its easier to just download a torrent? DRM doesn't stop piracy, because "stopping piracy" has always been the PR-excuse. DRM is an attempt to usurp property rights, with a manufacturer retaining control over your device after the first sale.
edit: grammar
If you do it from your TV, its much harder to catch.
Without the ability to run software as we see fit, we are constrained to what these vendors permit us to do on our own devices.
If you like this constraint, GPLv2 and lesser licenses are for you, otherwise a license like GPLv3 is advisable to ensure your users aren't trapped with an outdated, vulnerable version of your software that their OEM refuses to update.
Putting "or later" at the end licenses your software under unknown terms. It could become CC0.
It's worth noting however that it's much easier to go from "v3 or later" to "v3 only" than the other way around.
Why so? I always wondered why say 'or later' rather than just adopt v4 (et al.) as dual (...) licences as and when, so I can believe it, I just don't understand it.
How well his ideas will survive him remains to be seen.
The more Orwellian Big Tech gets, the happier I am to be an FSF member.
So:
1. GPL3 introduced an incompatible, breaking change, knowing there was important GPL2 only code.
2. GPL3 did not include any way to grandfather in existing GPL2 code.
Being idealistic is great, but sometimes it has ugly consequences.
"Licensed under GPL 2.0"
Is ambiguous about whether later versions are applicable. They're not. Only GPL 2.0 applies because that's the listed license.
> Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and "any later version", you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation.
The license itself admits that it may cover multiple versions. Being clear, as the FSF is recommending on this page, seems like the best choice to avoid confusion for your end users.
(in particular, this page seems to be about deprecating the existing ambiguous SPDX license indicators and replacing them with ones that have clear 'only' or 'later' suffixes. This seems like a good thing.)
> The license itself admits that it may cover multiple versions.
... along with the conditions under which this expansive licensing applies. While it's good to be clear, it's not respectful of author's wishes to presume a later license when they didn't explicitly allow that. The article suggests that this case is ambiguous, when it is absolutely not and I find that to be irresponsible