ModelingToolkit V11 Library Split and Licensing, Community Feedback Requested

Hello!

First of all: thank you so much for everything. This community and the libraries these people have built have allowed me (and many others) to do amazing things. Even my Master’s thesis used MTK and Julia :slight_smile:

With that said, I want to give my opinion and an idea of how to proceed. My opinion, as a libertarian socialist (however you want to define it), is that MIT is a bad license. I like AGPL way more, for reasons I think should be obvious, given my political views.

With that said, I think Directus handles this nicely: they have everything under BSL, but require a paid license for any organisation that “deals with” (income, grants, etc) more than 5 million dollars annually. Maybe something like that would be good?

I am, in general, a big fan of everything AGPL, but I understand the downsides of that.

Also, maybe be a bit more explicit with donation requests, somehow? Maybe ask for sponsorships from companies that use the library? I don’t know if you have done it or not, just suggesting :slight_smile:

Thank you, once again, for everything!

1 Like

My understanding aligns with what the Google LLM blurb says:

The core purpose of the AGPL is to close the “SaaS loophole” in the standard GPL: if users interact with the software over a network, you must provide them with the source code

If it’s internal use software, then that should be fine?

1 Like

In theory yes! However, some people here interpret “derivative work” as everything including, for example, graphical results (plots) in a report document. Hence, in the current interpretation, everyone having a report with the results can ask for the source code due to the AGPL.

I know, this does not make sense. However, when people want to have access to things, those strange interpretations can be used. The amount of energy to fight judicially is huge. We are waiting for a global guideline here but it is unlikely to happen in a short-term.

2 Likes

Just to wrap up, my use case if very specific. Maybe it is unique given that I work in a research institute in the direct administration of the Brazilian government. Thus, I fully understand the decision to change the license model and I can only accept the consequences. I just want to point out my situation :slight_smile:

2 Likes

It might be worth reading through the GPL FAQs because it seems like you might be okay. My earlier post would suggest that if you are merely releasing reports/plots/figures then that is an exception to (A)GPL. Second, according to this FAQ .

"If I know someone has a copy of a GPL-covered program, can I demand they give me a copy? (#CanIDemandACopy)

No. The GPL gives a person permission to make and redistribute copies of the program if and when that person chooses to do so. That person also has the right not to choose to redistribute the program."

And this one

Does the GPL require that source code of modified versions be posted to the public? (#GPLRequireSourcePostedPublic)

The GPL does not require you to release your modified version, or any part of it. You are free to make modifications and use them privately, without ever releasing them. This applies to organizations (including companies), too; an organization can make a modified version and use it internally without ever releasing it outside the organization.

But if you release the modified version to the public in some way, the GPL requires you to make the modified source code available to the program’s users, under the GPL.

Have you brought this issue up with a government lawyer? This license issue would be worth bringing up with someone before making any final judgements.

2 Likes

Hi @mihalybaci !

Thanks for the info!

The history they told me was that a company received a report (that any citizen can ask from the public service) and this report contained some screenshot of a Qt software (internal software that was never released). The company said that given the software used GPL modules, the government institute must also sent the source code.

If I understood correctly, the case is under trial and the first instance said that those results in the report are under the definition of “distribution” and the institute should also send the source code. There are still many steps to go until we have here a conclusion.

I know that if everything is kept internally, I will have no problem. However, we must share some information in some circumstances. Then, until everything is decided, we were asked to do not use anything that can force use to disclose things we do not want to.

3 Likes

As far as I can tell, the problem here is that there isn’t a large body of case law making everything crystal clear with precedents on what you can and can’t do. So everyone can read AGPL licenses and conclude what they think is allowed, but at the end of the day that is just your opinion (even if you are a lawyer). Whether your opinion would hold up in court is another matter (not to mention how this could change from country to country).

4 Likes

That’s precisely the issue here! No one would have ever imagined that the court would comprehend a screenshot of software as a derivative work.

Therefore, I must continue, for now, with MIT (or similar) licenses, where the text is extremely simple, leading to no misunderstandings.

3 Likes

By the way, I want to mention again that this is very limited to my situation. Given that I work for the government, reports related to the work we are doing can be required. If we were a private company, there will be no problem at all, since we can just block any distribution of anything related to the AGPL software.

1 Like

@Ronis_BR Have you asked juliahub for a different license?

In principle, “software xyz is AGPL” is a category error. Instead, you can license software xyz from the copyright holder under any terms you negotiate. The reason why copyleft is so sticky is that “copyright holder” is often a diffuse mass of contributors, and tracking every one of them down to consent to any terms is infeasible (or strictly impossible: some people die, and the inheritance status may be in eternal limbo under crazy jurisdiction).

In this case, there is no such issue. Nobody would prevent your institute from getting a license to MTK from juliahub, for the princely sum of $1/year, under the “almost AGPL” that allows you to do everything that the AGPL allows you, including redistribution under AGPL; but also gives you “THE INSTITUTE” as licensee certain additional rights, like e.g. including output plots and screenshots in pdf reports, worded in a way that is unambiguous to even Brazilian courts (including snarky references to the problematic court decision).

Hi @foobar_lv2 !

Yes, that will certainly resolve the issue. However, since this is specific to my use case, I prefer not to bother JuliaHub until we have a guideline here regarding the licenses we can interact with during our paid hours.

1 Like

I think asking can’t hurt; “public sector in Brazil” is not that fringe, and this is a proactive way to shape things the way you want.

Your legal department gets thrown a bone “ok, what do we actually want?”, juliahub gets to publicly demonstrate good-will and how flexible they are in terms of licensing, and SatelliteToolbox maybe gets MTK integration.

I suspect that neither juliahub nor your legal guidance has appropriate experience / processes in-place; and this is close to the platonic ideal of “just say yes”.

Under juliahub’s interpretation of the AGPL, they shouldn’t need to sign anything at all, you already have a license under all terms you need!

And this would create an actual empirical data-point for your legal department / your courts to refer to, a la “in 31 out of 612 documented cases of (A)GPL trouble, an actual copyright holder could be found who has the legal standing to give out almost-GPL style licenses. In 31/31 of these cases, the copyright holder agreed that this is very silly, and in fact signed a license agreement. Maybe use that as a hint for how to deal with the remaining 581 cases of diffuse copyright holders?”

9 Likes