License discussion

[jchodera]

Did we ever decide on a recommended OpenMM-compatible license?

I think I had misunderstood that OpenMM was released under the LGPL license, but the main code seems to actually be the MIT license. Should we recommend the MIT license for omnia projects for maximum OpenMM license compatibility?

[mpharrigan]

I didn't realize OpenMM was primarily MIT except for CUDA and OpenCL.

The general "advice" has always been to pick a license you're comfortable with but try to avoid overly copy-left licenses like GPL. If you pick any reasonable license it will be "compatible" with OpenMM (for some definition of compatible)

Personally, I'm a fan of liberal (BSD/MIT) licenses. You obviously have to be careful about including code from projects with stronger copyleft provisions [example: mdtraj can't take any code from cpptraj or mdanalysis because they're gpl]

[franknoe]

We have definitely agreed on LGPL.

Am 05/10/16 um 19:15 schrieb John Chodera:

Did we ever decide on a recommended OpenMM-compatible license?

I think I had misunderstood that OpenMM was released under the LGPL
license, but the main code seems to actually be the MIT license.
Should we recommend the MIT license for omnia projects for maximum
OpenMM license compatibility?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#607, or mute the
thread
https://github.com/notifications/unsubscribe-auth/AGMeQq7pEXsHtLhWD-Ypmuf0cIbaaXGeks5qxD2mgaJpZM4KPa__.

---

Prof. Dr. Frank Noe
Head of Computational Molecular Biology group
Freie Universitaet Berlin

Phone: (+49) (0)30 838 75354
Web: research.franknoe.de

Mail: Arnimallee 6, 14195 Berlin, Germany

[franknoe]

Actually, PyEMMA used to be BSD initially and we did restrict it to LGPL
(with consent of all authors that time) because there was some concern
that BSD might be too permissive for Omnia products if they are used by
companies. If this is an issue, we should discuss it during the Omnia
meeting, but I'm pretty sure we agreed on LGPL.

Am 05/10/16 um 19:25 schrieb Matthew Harrigan:

I didn't realize OpenMM was primarily MIT except for CUDA and OpenCL.

The general "advice" has always been to pick a license you're
comfortable with but try to avoid overly copy-left licenses like GPL.
If you pick any reasonable license it will be "compatible" with OpenMM
(for some definition of compatible)

Personally, I'm a fan of liberal (BSD/MIT) licenses. You obviously
have to be careful about including code from projects with stronger
copyleft provisions [example: mdtraj can't take any code from cpptraj
or mdanalysis because they're gpl]

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#607 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AGMeQvUJ45DC2Zt1tHfsmLrEy9ixvE4Sks5qxD_sgaJpZM4KPa__.

---

Prof. Dr. Frank Noe
Head of Computational Molecular Biology group
Freie Universitaet Berlin

Phone: (+49) (0)30 838 75354
Web: research.franknoe.de

Mail: Arnimallee 6, 14195 Berlin, Germany

[jchodera]

I'm definitely not a license expert, but I've heard a number of concerns from others about LGPL, with the suggestion that a more permissive license like MIT is less troublesome. I also was surprised to learn that OpenMM is MIT, and had originally intended most of our code to be license-compatible with it.

@franknoe: Do you guys have strong reasons to prefer LGPL over MIT?

[jchodera]

Actually, PyEMMA used to be BSD initially and we did restrict it to LGPL (with consent of all authors that time) because there was some concern that BSD might be too permissive for Omnia products if they are used by companies. If this is an issue, we should discuss it during the Omnia meeting, but I'm pretty sure we agreed on LGPL.

Thanks for the clarification, @franknoe!

Totally agree we should discuss it during the next Omnia meeting. (Apologies for being away then---hope @Lnaden was able to successfully join in my absence!)

I'll do some reading about licenses before the next call to hopefully have a more informed viewpoint.

[rmcgibbo]

FWIW, if I was rewriting MDTraj from scratch now, I think I would have chose the MPL for the license.

It may be a good option to consider.

Sent from my iPhone

On Oct 5, 2016, at 5:40 PM, John Chodera [email protected] wrote:

Actually, PyEMMA used to be BSD initially and we did restrict it to LGPL (with consent of all authors that time) because there was some concern that BSD might be too permissive for Omnia products if they are used by companies. If this is an issue, we should discuss it during the Omnia meeting, but I'm pretty sure we agreed on LGPL.

Thanks for the clarification, @franknoe!

Totally agree we should discuss it during the next Omnia meeting. (Apologies for being away then---hope @Lnaden was able to successfully join in my absence!)

I'll do some reading about licenses before the next call to hopefully have a more informed viewpoint.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.

[peastman]

Personally, I'm fine with both MIT/BSD and LGPL. I generally lean toward more permissive over less, but LGPL is still "permissive enough" not to cause problems for most people and most applications.

I don't think it's necessary that all Omnia software have the same license, though minimizing the number of different licenses we use would be a good thing.