license issues

[simonschmeisser]

Hi @Jmeyer1292

is there a specific reason why you choose GPLv3 instead of something a bit more relaxed? LGPL, BSD like MoveIt! does , MIT, whatever? I'm really not sure if we can use it currently since it might mean we would need to distribute our software under GPLv3 as well.

Thanks!

[gavanderhoorn]

Yes, this is rather unfortunate and could potentially mean that it cannot be used for packages like https://github.com/ros-industrial/fanuc/issues/245.

I second the comment by @simonschmeisser: it would be +100 if the licensing could be changed to one of the non-copyleft licenses. Or perhaps LGPL would be enough, but anything with "GPL" in it makes people look for alternatives is my experience.

[Jmeyer1292]

When i implemented this from the paper there were some ambiguities. I found a matlab implementation that i thought was gpl, so i went with that. Looking back, I may have been mistaken. I dont mind changing it as long as @JeroenDM is okay too.

[JeroenDM]

That's ok for me, thanks for asking.

It's a good reminder for me that I should pay more attention to the license. Does this also mean that if I write a ROS package that depends on opw_kinematics and I don't add a licence to the github repo that I'm violating GPLv3?

(My argument would be no, since I made the source available and the user downloads opw_kinematics themself, but I really do not know enough about licencing...)

[JeroenDM]

Appart from my ignorance about licencing, I still encourage @simonschmeisser to make the source code of his application available if possible within your business context :)

[gavanderhoorn]

It's a good reminder for me that I should pay more attention to the license. Does this also mean that if I write a ROS package that depends on opw_kinematics and I don't add a licence to the github repo that I'm violating GPLv3?

apart from the possible GPL violations: not specifying any license is even worse. Code without an explicit license is almost toxic and cannot be used ever (or at least: in a commercial context).

[simonschmeisser]

Cool, thanks!

I would love to opensource everything but I don't think that would work as we sell low quantities at a relatively high cost per installation. So we even need to use hardware license protection. Which sucks (not only theoretically but also from a maintenance side). We do however try to contribute whatever changes we make to components.

[JeroenDM]

@gavanderhoorn Thanks for the explanation! I will add LGPL-3.0 to my projects as my colleges do for now, when the dependencies allow it. If someone actually wants to use the code and does not agree with the license, they can create an issue as was done here.

Any other suggestions welcome, but I guess this is not the best place to discuss general license choice for ROS-packages.

@Jmeyer1292 Just to avoid confusion, I do not suggest to use (or not use) LGPL-3.0 for this package.

[gavanderhoorn]

@gavanderhoorn Thanks for the explanation! I will add LGPL-3.0 to my projects as my colleges do for now, when the dependencies allow it. If someone actually wants to use the code and does not agree with the license, they can create an issue as was done here.

Any other suggestions welcome, but I guess this is not the best place to discuss general license choice for ROS-packages.

Unless you have a specific reason to use a GPL license, using something like Apache 2.0 or BSD might be better. Copy-left licenses are typically not very well received, although in academic environments that is much less a problem.

But in the end it's up to you of course.

[gavanderhoorn]

@Jmeyer1292: what are your current thoughts about this? Would you accept a PR updating the license now that @JeroenDM has given the OK?

[Jmeyer1292]

@gavanderhoorn It's done.

[gavanderhoorn]

Awesome. Thanks for this @Jmeyer1292.