Possible licence conflict

[DaisyFarns]

Since the original version of Jake Taylor's synth was written with the CC-BY-NC-SA 2.5 Licence, the fact that this port is under the MIT license is slightly troubling.

Do you think that rewriting it means it's not a "Derivative Work", as stated by the CC-BY-NC-SA, and may such be licenced in whatever way you wish? Did you get Taylor's permission to licence this with MIT?

I'm not accusing you of wrong doing or anything, but I'm curious to here your thoughts

[parasyte]

(Disclaimer: I am not a lawyer.)

Thanks for raising this concern. On the surface, I would agree with your assessment that this port is an adaptation of the original work.

The original copyright notice was added in 70c809a and then a few days later, the MIT license was added as part of #4. I claim I had a temporary lapse in mental judgement. Can't really explain how that happened.

Did you get Taylor's permission to licence this with MIT?

I did not. We spoke briefly at a meetup many years ago, before I started working on sonant-rs. What surprised me about this conversation is that I had no idea at the time that he wrote the original synth. I used the sonant-x port for a project that I demoed, and he recognized the synth by ear right away.

One of the licensing complications here is that sonant-rs is an adaptation similar to the JavaScript ports, neither of which use CC-BY-NC-SA:

• js-sonant is licensed under zlib and includes the copyright notice for Jake Taylor and Marcus Geelnard.
• sonant-x is also licensed under zlib and includes copyright notices for Jake Taylor, Marcus Geelnard, and Nicolas Vanhoren.

sonant-rs was written using both of these ports, as well as the original C code, as references.

This seems to put this project into a very precarious position. It's possibly in violation of two separate and incompatible licenses. I don't think it is possible to dual-license under CC-NC-BY-SA and zlib, due to license incompatibility ¹.

At this time, I can see three possible avenues:

1. The least disruptive thing to do might be relicensing under CC-NC-BY-SA. I would have to yank all published versions and then publish a new major version (just for a license change) to avoid license conflicts for upstream projects. This would still leave the project vulnerable to zlib license violations.

2. I can reach out to each of the three individuals and try to get their inputs on the licensing situation (and possibly permission to relicense under MIT).

3. Do nothing. Make the case that if js-sonant and sonant-x can both get away with their zlib relicensing, then sonant-rs can just as well relicense to MIT. It is the copyright holder's prerogative to enforce their copyright (or not). See the discussion on Pouet for an example where Taylor expressed praise for js-sonant without commenting on the relicense.

Footnotes

1. I could be wrong about this, but it is my understanding after reading through the CC FAQs. And specifically, the BY-NC-SA compatibility page says it is incompatible with all existing licenses. ↩

[DaisyFarns]

(I too am not a lawyer)

2. I can reach out to each of the three individuals and try to get their inputs on the licensing situation (and possibly permission to relicense under MIT).

3. Do nothing. Make the case that if js-sonant and sonant-x can both get away with their zlib relicensing, then sonant-rs can just as well relicense to MIT. It is the copyright holder's prerogative to enforce their copyright (or not). See the discussion on Pouet for an example where Taylor expressed praise for js-sonant without commenting on the relicense.

Very interesting to see that he didn't mention the zlib licence in that post, however he may not have noticed. That post I think is talking about Sonant Live (the JS port of the tracker) which was made by the same author as js-sonant, Marcus Geelnard. Now, I've had a look and I can't find exactly what the license used by Sonant Live specificity is using, and I'd argue it wouldn't have been to Taylor either.

Relicencsing under CC-NC-BY-SA would be possible, it also isn't recommended by the Creative Commons folks. But it also might be the only option, as that's what the original software was released under. Getting an explicit comment about the MIT licence from Jake Taylor would be very helpful. For the record, I've emailed him with a question about the licence for this project before I opened this issue.

I'm not very familiar with the zlib licence, but it seems very permissive and MIT-like. You would be able to sublicense a derivative work as MIT, just from my first read of it. You might be able to get away with a copyright notice, and including a link to the licence in the README.md, and that's probably enough to avoid zlib licence violations.

[DaisyFarns]

It's certainly a complex licencing situation. I don't envy you 😅

[nicolas-van]

Hello. I'm the maintainer of sonant-x and I was asked by @parasyte to give my opinion on this subject, so why not.

I too am not a lawyer but I had some basic training on copyright law applied to computer programming as part of my University Degree and have some knowledge on open source licenses in general that was acquired with time.

I have a quick question regarding this issue: could anyone share a link to Jake Taylor's original C Code ? The oldest source I know is this article from Marcus Geelnard, sharing his js port under the zlib license. Work of Marcus Geelnard is, by the way, inluded in the sonant-x git history and here is a link to the specific "oldest commit" I have in my possession: nicolas-van/sonant-x@97f4619#diff-33d1b0d954195ad989731183c4c41b82e8fc30829782ca3ee2d5b332a3c56c55

It should be noted that, in the absence of the original source code from Jake Taylor that I never had in hand, I see no reason to consider that original Jake Taylor's code was released under the CC-BY-NC-SA 2.5 license. As far as I know, the oldest sources available where under the zlib license and I have no reason to doubt about Marcus Geelnard altering the license applied by Jake Taylor when he created his JS port, even less without Jake Taylor's consent.

[DaisyFarns]

could anyone share a link to Jake Taylor's original C Code ? The oldest source I know is this article from Marcus Geelnard, sharing his js port under the zlib license

The original code is here, there's a download on the right hand side for the zip file. That contains the original C code in the Synth folder. There's a license.txt file in the top level which contains the CC-BY-NC-SA 2.5 license. I will say that the original C code file contains a copyright statement, but nothing about the license. But it seems clear that the original code is under CC-BY-NC-SA 2.5.

@nicolas-van Did you have access to the original C code when you were making sonant-x? If not, then I'm guessing that sonant-x wouldn't really be a "Derivative Work" of the original sonant by Jake Taylor, but would be a derivative work of js-sonant by Marcus Geelnard. Then, if the zlib license for js-sonant is good (e.g. Marcus Geelnard got Jake Taylor's consent to license it that way), then the zlib license for sonant-x would also be good. Although that wouldn't necessarily imply that the license for this project would be okay, it does help clarify the situation.

[nicolas-van]

Thanks for the original C code!

So, now I can summarize my thinking. There are two possible issues discussed here and those are very separate:

Relicensing from CC-BY-NC-SA 2.5 to zlib during port to Javascript by Marcus Geelnard

Technically this is not authorized by the CC-BY-NC-SA 2.5 and those two licenses also have a major difference: CC-BY-NC-SA 2.5 explicitly state that no commercial use should be made while the zlib authorizes it.

So, in order to alter the license during his port, Marcus Geelnard should have asked permission to Jake Taylor.

But this is greatly mitigated by the following points:

• There is no reason to think that Marcus Geelnard acted with anything other than good faith. He explicitly gave credit to Jake Taylor by acknowledging his work in his release announcement and kept Jake Taylor's copyright indication in his license while he could have had valid claim declaring the fact it was not derivative work (see bellow why I consider it could be a valid claim).
• It should be reminded that, due to the fact it's the copyright holder's prerogative to enforce their copyright, Jake Taylor is the one and only person in the world that could contest that re-licensing. Due to the fact he seems to be well aware of the js-sonant port (based on the conversation you had with him @parasyte ) and never reacted before I think it's a clear indication that he doesn't care. It can also be noted that Jake Taylor has a quite active Github account but never seemed to bother just posting the code of his old Sonant application even for archival purpose, which is another indication he doesn't care.

In the event where Jake Taylor would contest that re-licensing then it would be an issue for js-sonant and all derivative work coming from it (including my own work on sonant-x). But frankly that seems very very unlikely.

Ideally it would be best if Jake Taylor explicitly gave his authorization about that relicensing in js-sonant, but I wouldn't stress that much over that.

Non-reuse of original copyright notices and licenses in Jay Oster's work

In his Rust port, Jay Oster did applied a completely new license file that doesn't contain any of the previous copyright older's names and also has a different license.

If Jay Oster's work was a derivative of the other projects that wouldn't be authorized by any of the used licenses (neither CC-BY-NC-SA 2.5 nor zlib nor MIT by the way). As far as I can tell it constitutes a clear indication by Jay Oster that his work is not derivative work but instead original work inspired by previous work in Sonant, js-sonant and sonant-x, which is completely different.

IMHO, that's kind of OK for a simple reason: that' a completely rewrite to a completely different programming language.

Yes, I know that there is kind of a consideration in the Open Source community that ports to another programming language should be considered derivative work, but that's hard to prove anyway in the absence of copy-pasted code.

There are also many arguments that could be in favor of the "inspired by work":

• The algorithms used in the synthesis engine constitute a quite limited amount of code in all those projects and are rather generic anyway. (We're speaking about wave generators, some filters, that kind of things... nothing truly original even in Jake Oster's work.) Concretely their main difficulty is "getting them right so it's not too slow nor buggy", which is completely language-dependent.
• Regarding the other re-used aspect, which is the compatibility with some sound file file format, it could be said that was done to make the applications retro-compatible which is authorized by many copyright laws related to software programming.

Conclusion

If we wanted to be in the "100% in the clear, nothing could ever go wrong" for all these projects, what we ideal would be that:

• Jake Taylor gave explicit consent regarding re-licensing of his original work to the zlib. (Would prevent issues in js-sonant and sonant-x)
• Jay Oster re-used original copyright holder notices and licenses from the project he consider sonant-rs to be a port from (either the copyright notices in sonant, js-sonant or sonant-x). In case of doubt just take the one you consider you inherited most code from (probably the C version).

But on the other hand, we could also be realistic and see the fact that there are only 3 persons in the world (Jake Taylor, Marcus Geelnard an myself) that could claim something on these subjects and that really, no one seems to care about that.

Please remind that we're speaking about the demo scene here, so relax 😄 . And as a personal note, I'm already quite happy to ear that some of my work was used for some demo and could be inspiring for other persons. 😁

[yupferris]

Hello! Original Sonant author here. I saw both @DaisyFarns and @parasyte have sent me emails asking about this, and it seems this is the best (and most open) place to respond, which is probably useful if this comes up again in the future. If you'd like me to copy/paste this into an email response for a "better" paper trail, I can do that as well, but I don't think it will be necessary.

First of all, thanks for being thorough about licensing @DaisyFarns ! If there were ever an issue, this is certainly the kind of correspondence I would have hoped for, and I'm happy to see it.

To be clear in writing, and with the caveat that I am most certainly not a lawyer:

Regarding the original Sonant/licensing/etc - don't worry about it. Use it how you want to, with or without attribution. This is common in the demoscene and what I originally intended (as far as I can recall, it was quite some time ago!). This my attempt to give explicit consent to consider it public domain.

In hindsight I don't think I would have chosen the CC license; if anything I probably would have tried to donate it to public domain and not require attribution at all. The original work is a collection of bits and bobs that I pulled together from various places, all of which are (and were at the time) well-known techniques, that arguably either should have required attribution themselves, or, more likely, could already have been considered public domain. It's a very basic synth (in fact knowing what I know now, it's a bit embarrassing, but it also makes it very accessible, as evidenced by the many forks, so ultimately I think that's quite a good thing and I'm quite pleased with that). I have no issues with the current port/augmentation/licensing/etc situation. I do appreciate a mention somewhere (which I think most folks have done, if not all) - but that's more for my enjoyment, and not hugely important.

That said, if you're pulling in one of the ports, I would recommend you respect the license that's distributed with it.

P.S. regarding this:

I did not. We spoke briefly at a meetup many years ago, before I started working on sonant-rs. What surprised me about this conversation is that I had no idea at the time that he wrote the original synth. I used the sonant-x port for a project that I demoed, and he recognized the synth by ear right away.

Haha, I remember this, that was cool/fun, and I remember following your parasyte accounts on github etc some time later, and didn't put it together until reading this that you were the same person! Wow, I feel a bit dim now 😅

P.P.S. regarding a document for the .snt format (requested in email):

Unfortunately there is none, besides the code. Please forgive teenage me. 😄

[DaisyFarns]

That certainly clears things up!

@yupferris Thank you for your comment! Very nice to hear directly from the original creator. I did assume that ultimately the folks involved here wouldn't have any issues with licensing, but it's good to have that confirmed.

Thanks to the work that everyone on this thread has done for Sonant. It is a lovely little bit of software, and it's nice to see it still getting attention all these years later.

I don't think there's any unresolved questions regarding licensing now.

P.P.S. regarding a document for the .snt format (requested in email):

Unfortunately there is none, besides the code. Please forgive teenage me. 😄

No worries!

[parasyte]

Thanks to everyone for your input! This issue has been closed, but it will remain available for additional discussion.

This is a satisfying and amicable conclusion, IMHO. I'll mention that I also attempted to contact Marcus, but have not heard from them yet. If they do have additional information or requests, I'll gladly receive them.

@yupferris,

Haha, I remember this, that was cool/fun, and I remember following your parasyte accounts on github etc some time later, and didn't put it together until reading this that you were the same person! Wow, I feel a bit dim now 😅

Yeah, it was at the GDC side-event in 2016. Good times! And of course the feeling is mutual! We share common interests, after all. I am glad you were able to commit some time to helping out. And I'm always looking forward to your next set of adventures. (You should resume steaming! So should I!) 😄

[mbitsnbites]

Thank you all for your efforts to bring clarity to the subject.

Sorry for being late to the party, and I also apologize for the trouble, as much of it is due to my actions and I realize that I really should have put some more effort into being clearer about the licensing back in 2011.

Especially thank you once again @yupferris for your generous and helpful ways, and for clearing things up.

For future reference, I would like to add my part of the story. Disclaimer: I am also not a lawyer.

Some may argue that js-sonant can be called a derivative work of the Sonant synth source code (Synth/sonant.c from the original release from pouet.org), although at the time I was implementing it I do not think that I thought of it as such, mainly because I implemented it from scratch in another programming language (nothing was copied, all was written by hand, as far as I recall).

Side note: I would love to learn where the line is drawn for derivative work in terms of open source software and source code. I have yet to find a clear explanation (especially when we're talking about re-implementations or new implementations that are inspired by other implementations).

That said, the js-sonant implementation was naturally inspired by the Sonant synth implementation, since that source code was effectively the specification of both the data format and the function. Since I had not written my own music editor at the time I wanted to use the Sonant music editor tool (sonanttool.exe) for creating the music data files, and one key goal of the js-sonant implementation was that it should be compatible with the original editor.

I had a brief e-mail conversation (2011-04-11 to 2011-04-13) with @yupferris when going public with a precursor to js-sonant (called "a web demo" in that conversation - released on pouet.net as 4k-synth on 2011-04-27), and I still have it in my mailbox. If necessary, and with @yupferris permission, I could provide that conversation, but for now I'll summarise it as a positive conversation between two enthusiastic software developers that were mostly happy to share their work with the rest of the world (and quite frankly not overly concerned about licenses). I specifically asked if it was OK to use a liberal licence (zlib/libpng) for a potential future release of the source code (that would be "js-sonant", although it didn't have a name back then IIRC), and I received what I perceived as a supportive response.

We also had a few later e-mail conversations:

• On 2011-05-26 and on 2011-06-06 relating to SonantLive (a Web-based music editor that's compatible with Sonant and using the js-sonant synth).
• On 2012-04-11 relating to SoundBox (a new synth but built in the same spirit as Sonant).

Both conversations were in similar spirit as the first one, where we kept each other up to date and I wanted to make sure to run my releases by @yupferris as to ensure that I didn't do anything out of order (especially w.r.t. licensing and attribution).

I would like to thank @yupferris for being very supportive of my work with js-sonant and other related works, and I hope that this account that I have given is accurate and that we all can feel confident about the licensing situation for all related products (sonant, sonant-rs, js-sonant, sonant-x, ...) going forward.

[yupferris]

@mbitsnbites cheers for weighing in and your summary of our conversations! I had considered going through those as well, but didn't feel the need, and that's especially true now - your summaries above definitely match what I remember, and yeah I don't recall ever having any issues with licensing/attribution, and still don't now. As you said, two enthusiastic software developers sharing stuff and not caring much about licenses 😄 !