Contradictions in Licensing and IP Acknowledge information

For EPIC, Michigan University suggested to remove our license statement, but I’m nearly sure I can make them accept the new wording from @siljelb and the slightly modified text in the “IP acknowledgment” element. And in the order it is per now.

I’ll give it a try, and if I succeed, I can do this manually in the adl for the EPIC and EPIC-CP archetypes only.

But we need to work on this further, in the PROMs WG or wherever, I don’t mind as long as we continue and find a way to deal with this ahead.

I think we need a license for each purpose, separated, and support multiple purposes per archetype = multiple licenses per archetype.

That way we have the flexibility to keep the “access” purpose to have the openEHR license, but the “usage” could be something else.

I can think of another case where the content’s access is restricted but maybe the openEHR foundation agrees on a publication license so the “access” license is really licensed by openEHR from XYZ" (the copyright holder. So that’s a B2B agreement. I’m not sure if openEHR is doing that right now.

Good points. Let’s work further on this, in whatever work group we find suitable. Most likely the PROM WG, as that group is working on the same issue.

If that is possible in terms of policy, we’ll need to check how to define the licensing per purpose metadata inside the archetype, and on downstream elements (templates, OPTs, data samples, etc).

Because if I grab an OPT generated by a 3rd party, how do I know if I’m violating the license of the original upstream artifact without the license inside the artifact I’m using?

I’m guessing, but the OPT might need to have a license for each resolved SLOT that is equal to the license of the original archetype, again per “purpose”, though when you are dealing with OPTs that’s clearly “usage” not just “access”.

Maybe the generation of the downstream artifact is itself “usage”, I guess someone that knows more about the legal part of it should define those things.

So, University of Michigan has replied and are happy for now. This is only one IP holder, but we can assume there are/will be others with similar concerns about how their IP and licencing information is displayed and handled. @pablo’s points are also something we have to deal with.

For the record, I’m pasting the reply from Michigan here, as it contains a couple of useful suggestions on how we name and display things:

“Thanks, this will work for now and is a step in the right direction. I think it would be more instructive if the IP Acknowledgements label was changed to Copyright Notice: or similar and the license label to be “Archetype License:” for clarity. The term IP acknowledgement really does not address the key issue of ownership and licensing of the measure. While I appreciate that the license is about the archetype the reality is that users can easily conflate these two distinct pieces. We would of course like to see both headers bolded.”

Great news!

Also agree with @pablo , some int’l copyright legal advise may be required, although I fear that by the time somebody got their heads around this, openehr is running out of money :face_with_diagonal_mouth:

Regarding the concrete suggestions from Michigan:

  • Changing the licence label to “Archetype Licence” seems straightforward.
  • Boldening the headers is fine (although I remember experimenting with this for the various right hand sub-headers such as author name, Original namespace etc. and it did not look too nice.)
  • Regarding changing “IP Acknowledgements” to “Copyright Notice”:
    We have a “Copyright” field already which could be used and is displayed above the licence if it exists (as in many other archetypes). This is typically the copyright from openEHR but could be a combination. Having a “Copyright” and a “Copyright notice” (or ignoring the copyright field and then labelling the ip acknowledgement field as copyright) will just cause new confusion I fear…
    Maybe we could relabel “IP Acknowledgements” to “Third Party Copyright Notice” or “Third Party Copyright Notice and IP Acknowledgements”. While IP Acknowledgements is broader and in theory includes copyright and e.g. trademarks and licences for example, it is just too fluffy.

So something like this (I have not adjusted the text in the second):

Thanks Ian, in short licensing (or PROMs but applicable to other copyrighted digital asset) is in principle is not that hard. I won’t go into too much detail but the devil is indeed in the detail (pun intended!). But, I think we’re at another intersection compared to actual solution providers who put these assets to work and materially (or otherwise) benefit from it. To represent copyrighted digital asset in a different format for the purpose of making available is different that actual solution providers and at times private healthcare and related organisations should be a different concern. So, any license fees should be collected from the party that makes money from it (e.g. biopharma as is). openEHR will certainly not monetise it. Some copyright holders get it some don’t but things are moving in the right directions and in the case of PROMs, clinical use conditions are offered by decent authors (which happen to have decent assets too).
We discussed this in CPB and within PROMs WG - and converged on a model where we pitch to authors, which is real, that Archetypes are simply defining a substrate for data capture - full stop. So, it’s not too different than taking an asset published in PDF and converting to Word or CSV. This is exactly the model ICHOM is able to make entire standard sets freely and publicly available as Excel files, which include all PROMs data elements (including additional prose to be able to interpret such data - but not implementation specific (e.g. Please circle each option only once etc… but things like “in the past 2 weeks:” so data consumer would know the data points are not monthly or post intervention etc.). IMHO, if SNOMED made humanity’s most expensive terminology available from CKM with conditions on their term of course, PROMs or other digital assets shouldn’t be that difficult. And this is where it ends. I had these conversations with over 100 authors in the past 10 years and you come across oddballs which you can never convince. Luckily, there more tha 20,000 (yes that’s 4 zeros!) PROMs out there so finding an equivalent or at least good enough instrument is not that hard. And ICHOM makes an effort to recommend instruments, most copyrighted requiring license fees of course, from authors who get it. I recently made a LinkedIn post on this very topic - it’s not the purpose here but for those copyright holders hindering the use of validated PROMS or clinical scales and other digital assets which would definitely affect real lives, the issue is ethics and integrity. Just saying :wink:

Hope this is useful.

The ugly truth of PROMs deployment at scale: LICENSING.

I am confused on what distinction is being drawn with the dual licensing:

This is the case for other archetypes as well, and we need to be able to both state that the archetype itself is open source, but still point to the fact that to use the archetype is licensed. It is not super clear as it is now.

From by-sa 4.0:

You are free to:

  1. Share — copy and redistribute the material in any medium or format for any purpose, even commercially.

  2. Adapt — remix, transform, and build upon the material for any purpose, even commercially.

  3. The licensor cannot revoke these freedoms as long as you follow the license terms.

I am not a lawyer, but how can something comply with by-sa and not allow commercial use? If content is not compliant with by-sa, then it should not be listed as such. I cannot see how the definition/model/contract can comply and prohibit any form of use - what am I missing?

Commenting here by request…

Disclaimer: IANAL. But I’ve spent waaaay too many hours dealing with US and international copyright and related issues, and can therefore confidently assert that you need a lawyer that specialises in international copyright before nailing down anything here. Trying to apply common sense and logic to copyright issues will get you in trouble every time. The alternative is finding som pragmatic approach that completely sidesteps the throny legal stuff.

TLDR: one field for our own copyright, suitably defined, and a purely informational field to document (neutrally) what third parties assert but only to help reusers make their own determinaton.

So (long version)…

The one thing we control is the copyright on what we have ourselves produced (technically also other IP-related rights such as patents, trademarks, etc., but we can get away with ignoring that for the purposes of this discussion). So what ought to be specified and displayed in a prominent structured form is our copyright assertion and licensing terms (CC BY-SA). A copyright license presupposes that there exists a copyright to begin with, which is not a given for an archetype, but with a permissive license like CC BY-SA we can get away with kicking that can down the road a bit. But we should display this with phrasing to the effect of: “To the extent any independent copyright arises in the archetype, it is licensed under CC BY-SA”). We should also be clear on which version of CC BY-SA we mean (3.0? 4.0?) and whether we mean later versions to apply automatically (at the reuser’s discretion).

Other people’s copyright? Now you’re in the quagmire.

The most obvious issue is that under US copyright law the parts of a score (or whatever) that an archetype implements have a, at best, questionable claim to copyright protection to begin with. Mere information is not protected, only actual creative expression is eligible for copyright protection, so the archetype bits might not have a copyright to begin with (and thus no basis for enforcing a specific set of licensing terms). Even text copied directly and verbatim from the source may not be eligible and thus may not be subject to the licensing terms (unless you enter into an implied contract to abide by them, despite no copyright existing to begin with).

Another issue is that for a joint work, one copyright holder cannot speak for the other copyright holders (specifically when it comes to granting permissions to third parties like us), absent a specific contract between them (that we are not party to and cannot know whether exiosts). So whatever special permissions or deals we make will have to be negotiated, and documented, with each of the copyright holders individually.

And then we get to the issue of which copyright. (Almost) everybody is a signatory to the Berne copyright treaty, but that treaty is not self-executing: each country must implement terms compatible with it in their own legal framework. There are rules for determining what the country of origin is for any given creative work, and then the copyright is determined by that country’s laws (e.g. in the US the term of protection used to be 95 years after publication, where a lot of the rest of the world protected works for 70 years after the death of the author). Internationally, Berne requires each signatory to implement laws that respect that original copyright, but each country implements this differently and the terms interact in unpredictable ways with other parts of their copyright laws. For example, there is great variability in what the criteria are for being eligible for copyright to begin with, and whether you need to be eligible for copyright in country A for its copyright laws to grant protection to a copyright from country B. In the US you do not get protection for a foreign work that would have been ineligible for copyright protection if it was a US work. In the UK the situation is the opposite (or I think it was, but it’s been a while since I checked, I could misrecall).

There are three parties to consider (us, a third party copyright owner, and a reusers of the archetype), and thus potentially three sets of copyright laws to consider (in addition to the international treaty-based rules) for any given reuse of an archetype (which is what the license in the archetype is trying to regulate.

Confused yet?

Now we can start trying to handle other IP rights that may exist, such as patents and trademarks. These are completely different legal frameworks (there’s no such thing as “IP law”; the term is a gross oversimplification) with at least as many complicated interactions and concerns as copyright does. And then add to that some other communities refer to as “non-copyright restrictions” (by which they mean restrictions that are not from any of the usual “IP” issues). An easy example here is pictures of legal tender (bills, coins). These are variously protected by copyright in different jurisdictions, but are quite often illegal to reproduce irrespective of copyright in order to deter fraud and forgery. In some jurisdictions it is absolutely possible that an archetype, or the upstream thing the archetype tries to implement, would be covered by special rules that restrict or permit dissemination based on e.g. being medical information, necessary for healthcare purposes, etc.

And what if two or more parties both assert copyright in whatever upstream thing it is we’re trying to implement?

So… Dealing with all this complexity for an unbounded number of possible usage scenarios is entirely impossible. Which suggests we should… not do that.

What is achievable is to somewhere document “known third-party rights and restrictions”. And in that field to document, for example, that “University X, University Y, and Health Syztem Z assert copyright to the Flurble Score, and request that enqueries about licensing be sent to licensing@example.com prior to use in a concrete implementation / prior to commercial exploitation / etc.”.

And that phrasing should always be such that it attributes the assertion to the party actually making that assertion (so that it is not us that is making that assertion), and should always be described as an assertion and not as fact (because the copyright, or whatever, may not actually exist in the first place). By doing that we are only making information available so that reusers can do their own due dilligence, instead of making it seem like we have made a legal determination that we cannot and should not even attempt to do.

Apologies for adding to the confusion, but the point is: the confusion is inherently so great that the only sustainable approach is to completely sidestep the issue, without abdicating the responsibility we have to our reusers and without stepping on the those of the upstream providers whose cooperation we need.

Cheers,

Terje

Hi Terje, thank you for the detailed response!

I am also not a lawyer but have spent a bit too much time in their company (I married one and have permission to make that joke =).

I believe where I am trying to get to is that an archetype encumbered by the licensing in the example here is not compliant with CC BY-SA. If there is any listing or indication that the archetype is CC BY-SA, that is incorrect. Specifically, the quote I am questioning is:

This is the case for other archetypes as well, and we need to be able to both state that the archetype itself is open source, but still point to the fact that to use the archetype is licensed. It is not super clear as it is now.

I do not believe the above is possible in terms of CC BY-SA (though again, not a lawyer and this is not legal advice). The terms of CC BY-SA allow for commercial use, modification, etc.. Content can be dual-licensed, but each of the licenses need to be honored. The indicated content is not dual-licensed.

As an example from the HL7/FHIR side, X12 is a licensed product that we cannot publish any contents of. All official published guides (which are generally CC-0) have narrative descriptions of IP considerations (e.g., DaVinci PDex : Intellectual Property Considerations), which are generated based on the licensing information in artefacts themselves. A caveat in this scenario is that NO actual contents of X12 are made available - e.g., there are no examples that have X12 values, etc. since publishing the content would be a breach of their licensing terms. Definitional ‘things’ are just placeholders that direct people to the correct route to obtaining them (e.g., CodeSystem: X12 Claim Adjustment Reason Codes).

In this case, I assume the openEHR team has made arrangements regarding the publication of the EPIC archetype. I believe that content is in the “used with permission” space and the license can be clarified (as you suggest - though holders generally prefer avoiding the term ‘exploitation’ ;-). In my experience, that information needs to both be attached to the publication and the “publication version” of the artefact in question to minimize legal exposure.

So TLDR: I agree with your approach in principal but differ on the approach slightly. My opinion for sidestepping the issue is to be painfully clear when a different license applies to content. E.g., in any rendered view of an artefact, the license should be stated and references to generic terms like “open source” should be avoided.

Thanks!

That’s why I suggested to consult with a legal expert: can can interpret and use licenses without understanding all the requirements and scope, and we might be using them wrong. Though we still dedicate time to discuss them, while a single consultation might erase all doubt, though we are community and that’s in the scope of the Foundation/CIC, though the CPB can raise the issue to the right authorities.

Agree with you @pablo, this needs to be raised with CIC/Foundation board.

There are three relevant fields in archetypes: Copyright, licence and ip_acknowledgements - and definite legal guidance on their usage is needed in light of the comments of @GinoCanessa and @tbless. In addition, there are also the sections on “Intellectual Property” and “3rd Party IP” at Governance – openehr.org : In particular, the first mentions “Clinical models based on copyrighted ‘scales’ or ‘scores’ (eg Glasgow Coma, Braden etc)”, but does not reveal what the principle is.

Hi everyone,

Thank you all for this discussion. There is a lot of useful information here, and I appreciate everyone taking the time to discuss the topic. I have mostly been observing the conversation, as I do not yet feel informed enough to have a opinion on the matter.

I understand the challenges we are currently facing with these and other archetypes that have raised concerns from copyright holders over time.

I would like to compile a collection of all historical copyright claims regarding archetypes hosted on CKM. Many of those who have received these emails/communications are participating in this discussion. I would imagine it is typically the original author of the archetype who receives these claims. Could you please forward any related communications to comms@openehr.org ?

I will also ask the CPB to review this topic together with the PROMs WG. I am currently collecting your responses, as well as similar discussions posted on Discourse, mail archive, Confluence and CKM discussions, in order to prepare a document for presentation to the CIC Board.

As many of you have also suggested, I believe this is an issue for which we will likely need official legal guidance.

@sebastian.garde that’s another question for the “legal department”, are those fields enough for representing the level of complexity we need for differentiating licenses based on the requirements we have right now, and without contradictions?

Because, besides of which license applies for what, and checking potential incompatibilities between licenses, we’ll need the structure to store that complexity without ambiguity in archetypes, but also in al derived artifacts (you can’t miss the license in the OPTs!).