Let's discuss the future of our licenses

legal
feedback-wanted

(Michael Downey) #1

Hi all,

Let’s discuss the future of licenses for stuff we create.

In the past, OpenMRS used MPL 2.0 + HD. We had only a few cases of concerns with that license, but perhaps there is rationale for a difference choice going forward. See also the OpenMRS Contribution policy at https://wiki.openmrs.org/display/RES/OpenMRS+Contribution+Policy that @lrosen put together.

I know Judy had some concerns about MPL vs a stronger copyleft license like GPL, and I hope she’ll share those here, but maybe others have strong feelings one way or another. Let’s get all of those concerns & ideas out in the air here.

I hope we can use our decisions about licenses to respond to our actual concerns that we have and things we’ve seen, rather than a specific license for a license’s sake. :slight_smile: :scales:


(Lawrence Rosen) #2

Please don’t change. I LOVE your MPL+HD license. The open source community in general loves your license.

/Larry


(Michael Downey) #3

I don’t dislike it myself … but can we probe a bit more? Why is MPL2-HD desirable? Generally my observations have been that people see it as a “happy medium” between permissive & strong copyleft like GPL.

But perhaps some are concerned about people not contributing back, or somehow losing the value of their own contributions through non-open derivative works. Is such a concern valid? And if so is it warranted?

As I mentioned earlier there aren’t too many examples of downstream derivatives of OpenMRS. Judy mentioned to me a project from Mayo Clinic but it’s not clear that they distributed their work, so in that case maybe the license doesn’t really matter? ThoughtWorks created Bahmni under AGPL but that’s really on the other end of the spectrum.

Are there other issues to consider?


(Lawrence Rosen) #4

It isn’t “liking” or “disliking” actually. I’m sorry I turned the license discussion into one about love.

What I mean is that the MPL+HD license is professionally open source. It is reciprocal (i.e., copyleft) but not outrageously so. Without the “+HD” part it is widely accepted. It encourages the combination of MPL+HD software with other licenses, even the GPL. It is a damn good license.

Also, comparisons with AGPL are kind of unfortunate. The license spectrum is not like electromagnetic wavelength. AGPL is almost like GPL which is almost like MPL+HD, except for differences. :slight_smile: The main important difference for now is that AGPL also covers the distribution of licensed software over networks as well as on floppy disks. This is an appropriate way to deal with software nowadays. But it has little to do specifically with health software applications – except perhaps their user interfaces. As OpenMRS decided previously for a user interface contribution, even AGPL software is compatible with MPL+HD when the software is relevant.


(Saptarshi Purkayastha) #5

I don’t see that MPL (+HD) as a way to ensure that modifications are brought back to the community. The intent of the license is that, but the way in which the platform/software is designed, one could write a module and completely change every behavior of the platform. I have seen this happen in the past and in fact consulted a large ISV to use this architecture because they were building a proprietary product and couldn’t tell the shareholders/customers that they internally use open-source software. To me, the “happy medium” is only an intent, but can never be a force to get contributions back. In fact, GPL/LGPL or others wouldn’t either.

That is one reason, why I have been of the opinion that we should use a permissive license like APL and talk to the conscience/goodwill of people who modify it (which in any case is all that can be expected). On the other hand, I see more and more deployments need a web service, because there is lack of local medical informatics skills or we would provide analytics-as-a-service based on data from multiple implementations. Here, AGPL might be useful, because we can ensure that anyone who takes our service and modifies it, needs to give back those changes. Unless off course, they may implement a module that uses (overrides without modifying code) the API internally and again provide added services. There again is a way to circumvent AGPL.

So what does MPL+HD really achieve, given that we have a modular architecture that allows changing everything from the system, without actually having to modify the source at all?


(Michael Downey) #6

My understanding is that GPL (and I assumed also AGPL) required modules of GPL systems to also be released under GPL or equivalent terms. That has been the presumption with Discourse plugins. Is it true?


(Lawrence Rosen) #7

And also MPL-HD, for modules that have been changed or modified.

Many of us believe the same is true of the GPL licenses, and that the distinctions between the MPL and GPL licenses are illusory.

Permissive licenses don’t solve that problem either.


(Michael Downey) #8

But MPL 2.0 allows “secondary licenses”, so couldn’t I take an MPL2-HD licensed add-on module, add & change some things, and relicense it as GPL? (And the same would be true with the core platform product, too, yes?) I realize this is a slightly different tack than the earlier discussion but I’m just trying to make sure I understand all the possibilities.

And, more directly related to Judy’s concerns, couldn’t one take a MPL2-HD licensed core, as well as one or more MPL2-HD licensed add-on modules, and then create (and importantly, distribute!) a new derivative downstream product with a properitary license, as long as the original files were still made available under the original license?

I haven’t had enough caffeine yet this morning so forgive me if I’m missing something obvious. :slight_smile: :coffee:


(Lawrence Rosen) #9

You don’t need caffeine. You are correct on all points. Yes, the licenses are different. But not enough different to worry about it now.


(Lawrence Rosen) #10

There is no such guarantee in any open source license. Cooperation with the original community is generally a “practice” implemented for practical and cultural reasons rather than a legal requirement. (That is the very definition of “forking”!) Not that copyleft is meaningless: Copyleft obligations require the disclosure of modifications or changes of the software to the licensee. That licensee is then free to distribute them to anyone, including the original open source community.

That may sound like a difference without much of a practical distinction, but I think it is important to let community contributions (other than “modifications” or “changes” to specific open source modules that must be published) be voluntary.

Some of the GPL license interpretations are stricter in “linking” interpretation, but I don’t agree with them as a copyright law matter. Your mileage may vary.


(Michael Downey) #11

On a related question, what is the status of re-using the Healthcare Disclaimer? Is it, as a license, copyrighted? The disclaimer itself is not an OSI or FSF approved license.

If I look at http://openmrs.org/license/ it would suggest that the text of the disclaimer is released under CC BY 4.0. Is that accurate?


(Lawrence Rosen) #12

Re-use freely.

[The rest of this message is nonsense because Discourse requires at least 20 characters before it will accept “REPLY”!]


Ultimate Discourse "minimum length" complaint thread :mega:
(Michael Downey) #13

2 posts were split to a new topic: Ultimate Discourse “minimum length” complaint thread


(Saptarshi Purkayastha) #14

@lrosen what do we need to do with the LICENSE file and the license header in all the OpenMRS source files. e.g. APIAuthenticationException.java e.g. the LICENSE file starts with a Mozilla Public License, version 2.0 … but then ends with

Copyright © OpenMRS Inc. OpenMRS is a registered trademark and the OpenMRS graphic logo is a trademark of OpenMRS Inc.

which is likely for the healthcare disclaimer. You did mention that was reusable. The source files contain the following:

/**

  • This Source Code Form is subject to the terms of the Mozilla Public License,
  • v. 2.0. If a copy of the MPL was not distributed with this file, You can
  • obtain one at http://mozilla.org/MPL/2.0/. OpenMRS is also distributed under
  • the terms of the Healthcare Disclaimer located at http://openmrs.org/license.
  • Copyright © OpenMRS Inc. OpenMRS is a registered trademark and the OpenMRS
  • graphic logo is a trademark of OpenMRS Inc. */

Does this mean the source is copyrighted to OpenMRS Inc.?


(Michael Downey) #15

I’ve always thought this claim dubious since there was never any copyright assignment from individuals to OpenMRS Inc. (or its predecessors) other than a handful of forms filled out by some people way back in 2006 or so. And despite corporate personhood, I always thought that a corporation can’t create works on its own without some type of work-for-hire agreement or copyright assignment.

But very interested in what @lrosen or others think…


(Lawrence Rosen) #16

Has this question been here since June 29? Sorry. Can you add a nudge button to Discourse?

LibreHealth claims copyright only on the aggregate project applications that we distribute, not on the individual components within it. The copyrights in the components remain individually owned by their contributors, or the employers of those contributors, and licensed to us under an open source license.


How to "invite" people to a topic
(Saptarshi Purkayastha) #17

Documentation license: This is relevant to the documentation team and the education team. But important for all projects and some advice from @lrosen would be great. Since we will be releasing a lot of education material and documentation, I would like for us to choose a license. I selected the BY-NC-SA 4.0 here - https://gitlab.com/vchilkuri/lh-toolkit-dev-guide/blob/master/LICENSE . Many of us work in university settings and would like the material to be used for non-commercial purposes. Since we have discussed that we want to bring the changes that one makes to documentation back with the same principles with which we originally licensed material, ShareAlike seems like a appropriate one to choose.

Does anyone else have a different opinion? More permissive or less than this?


(Robby O'Connor) #18

I can support this. :+1:


(Saptarshi Purkayastha) #19

We have lh-toolkit-dev guide here as CC-BY-NC-SA 4.0 - https://www.gitbook.com/book/librehealth/lh-toolkit-dev-guide/details @rallam92 is starting to move a lot of content from OpenMRS wiki and rewriting some of those as gitbook chapters. I wonder whats the best way to attribute all those people who worked on the content. I am thinking we should have an intro where we acknowledge all those who had edits on the pages. Is that acceptable? Or do we need to point to the actual history somewhere? What do others and particularly, @lrosen think about re-licensing the material from the OpenMRS wiki as CC-BY-NC-SA?


(Robby O'Connor) #20

Isn’t it already that?