[License-review] [2nd Resubmission] ModelGo Attribution License, Version 2.0

[Carlo Piana]

> Da: "Moming Duan" <duanmoming at gmail.com>
> A: "License submissions for OSI review" <license-review at lists.opensource.org>,
> "Carlo Piana" <carlo at piana.eu>
> Inviato: Giovedì, 30 ottobre 2025 5:40:51
> Oggetto: Re: [License-review] [2nd Resubmission] ModelGo Attribution License,
> Version 2.0

> Hi Carlo,
> Thanks for your review.

>> while reviewing the modifications to decide on the approval (actually it should
>> be better to withdraw and resubmit at the end of the discussion) I noticed that
>> the modifications only applied to the MG-by and not to the Open Source version.

>> Comparing the two texts, moreover I see that here the liability disclaimer says
>> "to the maximum extent permissible under applicable law, the Licensed Materials
>> are provided on an “as is" and “as available” basis without any representation,
>> warranty, *condition* or term of any kind (whether express, implied, statutory
>> or otherwise)...

>> "condition, or term" does not appear in the other license and I think it does
>> not make sense adding it at all, since the license has terms and conditions.

>> At least, however, I think you should port the improvements of the Attribution
>> license to the Open Source one, I really don't see why, while sharing like 95%
>> of the provisions, this has not been done to the identical ones here: the
>> rationale applies equally.

> It seems you’re comparing the 2nd resubmission of MG-BY with the (1st)
> resubmission of MG-BY-OS (which was renamed to MG-BY-SA in the 2nd
> resubmission). For clarification, the 2nd resubmission of the MG license text
> is in the email with the subject:

> [2nd Resubmission] ModelGo Zero License, Version 2.0
> [2nd Resubmission] ModelGo Attribution-ShareAlike License, Version 2.0

It's very likely a mistake on our side, sorry. 

>> I also object using "Source Code Form" for "preferred form for making
>> modifications, since most evidently the definition includes way more than
>> source code and it is not good practice to define something with a name which
>> is misleading.

> This definition refers to GPL-3.0, which states:

> The "source code" for a work means the preferred form of the work for making
> modifications to it.

I know, this is the same definition we used. The objection is not on the definition, it's on calling this stuff "source code", because it's not code and probably not even source. 

> I am also not sure that MG-BY-SA can meet the approval requirements. And as Pam
> previously mentioned, there remains a question of whether copyleft can
> effectively function in the context of ML models. I look forward to OSI’s
> decision and hope to engage in a deeper discussion on this topic in the future.

That's a shared concern, yes. 

>> I am still dubious that you can include the Output in the Derivative definition
>> and that a license can purport to control it. This shall undergo more
>> discussion, here and elsewhere, as it is a central point upon which I tried to
>> make an impression every time I discussed porting Open Source concepts to the
>> world of AI.

>> For the "zero", I think the title is misleading, because it has no "zero"
>> conditions, and I have not made a final opinion yet.

> Please refer to the 2nd resubmission version of MG, where I have already removed
> all the conditions as suggested by McCoy Smith. Thanks.

Then we have a problem with tracking down the re-submission, because apparently I am not the only one confused by that. 

Could you please provide us with a full set of the final, revised licenses, as well as a word-diff comparing them to the originals? That would be a huge help and save us from further embarrassment of looking in the wrong place. Thank you. 

Carlo 

> Best,
> Moming
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensource.org/pipermail/license-review_lists.opensource.org/attachments/20251030/ff9f532e/attachment.htm>