[License-discuss] Total Reciprocity Public License (TRPL v1.0)

Shuji Sado shujisado at gmail.com
Sat Dec 13 06:05:15 UTC 2025


Hi Jay, Josh, and all,

https://discuss.opensource.org/t/distillation-and-oss-licensing/1333
I just noticed that Josh-san’s Discourse thread on distillation and
copyleft-style model licensing is closely related to TRPL’s Section 4.1
“Functional Output”, especially where it treats “ML trainer generating
Model Weights” as a derivative work and tries to impose reciprocity on that
output.

My main point is about enforceability and scope.

1) If the “functional output” (including weights) is not a derivative work
under copyright law in a given jurisdiction, then Section 4.1 cannot be
enforced as a copyright condition in the usual Open Source license sense.
At that point, the only realistic mechanism is contract (a promise by the
downloader).
2) But contract obligations bind only parties who actually received the
artifact and the license text (and can be shown to have assented). This is
a weak fit for distillation via API access or output scraping, where the
distiller may never have agreed to anything in the first place.
3) Also, relying heavily on contract increases jurisdictional uncertainty.
In Japan, licenses are generally understood as a form of contract, but they
are still analyzed primarily through copyright as the common baseline.
Moving the center of gravity to pure contract makes cross-border
predictability much worse.

So I think reciprocity on “things that look technically derivative but are
legally uncertain” may be possible in a narrow, symbolic sense, but it is
unlikely to achieve the practical deterrence people hope for (and it risks
drifting toward the kind of “tool controls the output” problem that Josh
raised under OSD9).

If the goal is an Open Source license, it may be safer to narrow 4.1 to
situations where identifiable portions of the Program are actually
incorporated into the output (for example runtime libraries), rather than
deeming all functional output to be a derivative work by definition.

Shuji


2025/12/10 0:09 Jay Patel <jaypatel.ani at gmail.com>:

>
> Modified original Licence text into this draft from Feedbacks:
>
> https://github.com/trplfoundation/trpl-license/blob/main/Draft
>
> On Wed, Dec 3, 2025, 10:17 PM Jay Patel <jaypatel.ani at gmail.com> wrote:
>
>> Thank You Gil for feedback.You identified a critical logical loop (the
>> 'Postman problem') that would have made compliance impossible.
>>
>> Maybe adding this address your poins:
>> 1. Adding a 'System Library/Standard Tool' exception: To ensure generic
>> tools (OS, Browsers, API clients) are not infected, targeting only specific
>> functional wrappers.
>>
>> 2. Tightening 'Intimate Communication': Removing 'not limited to' and
>> defining specific dependency types to reduce legal ambiguity.
>>
>> 3. Clarifying Deployment: Explicitly exempting personal/study use,
>> focusing 'Deployment' on organizational/commercial utility.
>>
>> My goal is 'Stronger than AGPL,' not 'Impossible to Run.' Your feedback
>> helps bridge that gap.
>>
>> Regards,
>> Jay
>>
>> On Wed, Dec 3, 2025, 9:52 PM Gil Yehuda <tenorgil at gmail.com> wrote:
>>
>>> Jay,
>>> If you intend to ask for critique, there’s quite a bit — from nitpicking
>>> details to fundamental flaws. I’ll list some of the apparent ones below as
>>> I read the license text.
>>> If you intend to suggest this as a new open source license that would
>>> meet the OSD, I don’t think this will do.
>>>
>>> As I read the license:
>>>
>>>    - Preamble: Licenses are documents that grant rights under
>>>    conditions. This text suggests that the license can guarantee freedom,
>>>    indeed “radical” freedom (I’m not sure of the difference) in software
>>>    architecture (not just code?). Licenses should articulate the rights they
>>>    grant and the conditions under which those rights are granted. Preambles
>>>    are great to convey intent which helps when trying to interpret ambiguity;
>>>    but they also reveal cases where the intent is to express a wish for how
>>>    things ought to be in the world. That’s better expressed in a manifesto,
>>>    not a legal document.
>>>    - “Intimate Communication” is one of my favorite terms found in
>>>    software licenses since it makes people think we’re also dabbling in
>>>    marriage counseling. My constructive comment here is that when licenses say
>>>    “This includes, but is not limited to” that automatically creates a speed
>>>    bump where a reader (and their lawyer) have to imagine if this includes
>>>    something surprisingly not intended. It creates a very broad scope — and
>>>    that’s going to warn me to stay away from using code under this license
>>>    because I might intent to comply only to learn that the scope was even
>>>    broader than assumed.
>>>    - “Content Output” is defined with two terms “human consumption or
>>>    data storage” — I understand the first to exclude non-human uses and the
>>>    second to exclude the use of data that is not stored. I note this because
>>>    of the next phrase...
>>>    - “Deployment” is defined with a curious inclusion of the term
>>>    “internally or externally” which I assume means in the context of a
>>>    corporation (not of “human consumption” in the above clause — right?!) If
>>>    so, then “internally” suggests that if I deploy my application onto my work
>>>    computer for use by my work colleagues, then the copyright license
>>>    considers this to be “deployment’ subject to copyright protection. I do not
>>>    believe that would hold up in the current interpretation of copyright laws.
>>>    - “Consequently” (line 40) is where this becomes quite challenging.
>>>    If I create a system with code licensed under TRPL 1.0 that shares data
>>>    with any proprietary software to achieve a unified functional goal — this
>>>    license declares that the proprietary software becomes part of a "Combined
>>>    Work” that I must release its source code under the terms of this license.
>>>    But what if that proprietary software is not mine to release? I might not
>>>    even have the source code? Let’s say I license the proprietary edition of
>>>    Postman and use it to make an API call to software under TRPL 1.0 —
>>>    internally (to my corporation, not in my body). I now have to acquire and
>>>    release Postman’s proprietary source code under the TRPL 1.0 license? How
>>>    would I go about doing that? Since there’s no definition of “release” here,
>>>    can I assume that if I deploy internally, then I can release internally
>>>    too? You see that would not help promote your intent. This section of the
>>>    license seems to convey how you wish software would work — but it does not
>>>    clarify how I, a potential user of software licensed under this license,
>>>    needs to do to make the world work that way.
>>>
>>>
>>> I’m concerned there is little practical use of this license since any
>>> software licensed this way, no matter how appealing that software may be,
>>> is automatically going to pose a threat to the rest of my software. Given
>>> that software is subject to copyright, and that as a user of software, I
>>> seek to honor other people’s copyrights, this license would make it nearly
>>> impossible to do so. I’d always have to limit my use of this software to
>>> ensure I don’t inadvertently infringe other people’s rights.
>>>
>>> Rather than a license, maybe we can collectively imagine what the past
>>> 40-50 years of technology would have been like had there been no copyright
>>> on source code. I imagine it would be different — better in some ways,
>>> worse in others. This license appears to invoke that imagination. But since
>>> source code is subject to copyright laws, I think the licenses should do
>>> their best to work within that context, granting rights that the grantor
>>> wishes to grant, and imposing conditions that the users of the software
>>> wish to, and can, comply with. This text falls short on the second part, at
>>> least for me.
>>>
>>> Gil
>>>
>>>
>>>
>>> On Dec 3, 2025, at 12:59 AM, Jay Patel <jaypatel.ani at gmail.com> wrote:
>>>
>>> I am reaching out to community to collect feedback on the proposed
>>> license.
>>>
>>> Here is text of License:
>>>
>>> https://github.com/trplfoundation/trpl-license/blob/main/LICENSE
>>>
>>> Thanks,
>>>
>>> Jay
>>>
>>> _______________________________________________
>>> The opinions expressed in this email are those of the sender and not
>>> necessarily those of the Open Source Initiative. Official statements by the
>>> Open Source Initiative will be sent from an opensource.org email
>>> address.
>>>
>>> License-discuss mailing list
>>> License-discuss at lists.opensource.org
>>>
>>> http://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>>>
>>>
>>> _______________________________________________
>>> The opinions expressed in this email are those of the sender and not
>>> necessarily those of the Open Source Initiative. Official statements by the
>>> Open Source Initiative will be sent from an opensource.org email
>>> address.
>>>
>>> License-discuss mailing list
>>> License-discuss at lists.opensource.org
>>>
>>> http://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>>>
>> _______________________________________________
> The opinions expressed in this email are those of the sender and not
> necessarily those of the Open Source Initiative. Official statements by the
> Open Source Initiative will be sent from an opensource.org email address.
>
> License-discuss mailing list
> License-discuss at lists.opensource.org
>
> http://lists.opensource.org/mailman/listinfo/license-discuss_lists.opensource.org
>


-- 
Shuji Sado
Chairman, Open Source Group Japan
https://opensource.jp/
English blog: https://shujisado.org/
Japanese blog: https://shujisado.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-discuss_lists.opensource.org/attachments/20251213/b7bc3b60/attachment.htm>


More information about the License-discuss mailing list