Comment by discardable_dan | Hacker News Reader

Comment by discardable_dan | Hacker News Reader

Parent Full thread

discardable_dan

Can you clarify what features did the new NWS add that old NWS was refusing to add?

That seems like the crux of the entire tiff, and without that knowledge it's very difficult to discern which side is more reasonable.

Edit: Honestly, some of the initial commits to the new repository are more in line with Liles' account: https://github.com/jackaudio/new-session-manager/commits/mas...

Starting in April 2020, changes include:

- Dumping the timeline, sequencer, and mixer: https://github.com/jackaudio/new-session-manager/commit/80e9...

- Removing any thing with logos or icons, e.g., "filing the serial number off": https://github.com/jackaudio/new-session-manager/commit/7992...

- Enable debug mode / verbose mode by default: https://github.com/jackaudio/new-session-manager/commit/3975...

- Change project build system over to meson: https://github.com/jackaudio/new-session-manager/commit/e431... (plus others)

- Update build system to exclude current GUI / make it optional: https://github.com/jackaudio/new-session-manager/commit/63e6...

- Set up symlinking by default in install: https://github.com/jackaudio/new-session-manager/commit/3681...

- Force the project onto GPL3, more serial numbers filing: https://github.com/jackaudio/new-session-manager/commit/2731...

After that is a number of fixes. Unfortuantely, from this list of "immediate" changes, it reads like the Argodejo developers decided they would be the preferred GUI, and decided to fork the project to abandon the current GUI. The README was even modified to this end: https://github.com/jackaudio/new-session-manager/commit/7f23... They also forced the project onto a different build system, and made new (and possibly incorrect) assumptions about installation. If the fixes came first (build fixes, etc.) and the other work came later, I would have a more-charitable view of this fork. Unfortunately, the immediate changelog sort of speaks for itself: they forked NSM, filed off all the names, changed over the licensing and build system, and ripped out the GUI. Then, a few months later, they actually started fixing bugs. The motivation here really seems to be subsuming ownership and default UI. This really just looks like a classic internet slap fight.

Of course, this is only conjecture from the git commit history, but... it's supported by the git commit history.

>- Force the project onto GPL3

Pretty sure this isn't allowed. The terms of GPLv2 allows the recipient of the software to modify and distribute the software under terms of GPLv2 or later. The software isn't public domain, someone owns the IP. You can distribute your own changes as you please. You cannot strip the IP owner's license and replace it wholesale.

They changed from GPL-2.0-or-later to GPL-3.0-or-later. The latter is a subset, so any content licensed under the former can be relicensed under the latter without consulting the original author.

The license sets the terms by which you can redistribute the software, it doesn't allow you to set the terms for the people you distribute it to. The only person/entity who can change the terms of the license is the copyright holder. People who receive this software will always have the right to redistribute under GPLv2.

As soon as you do any copyrightable change, you have the right to declare that change GPLv3 or later and that makes the overall project GPLv3 or later.

The previously existing code will still be GPLv2 or later, but you’ll have to trawl the commit history to find out what’s also available under GPLv2 or later.

So it is perfectly legal to relicense under GPLv3 or later, it’s just rude: https://www.draketo.de/english/politics/relicensing-gpl2plus...

If I have a program that consists of 5 files each of which is exactly 2000 lines lone and I distribute under GPLv2 or later. You like my program but realize that you have a novel new algorithm that will speed up the core functionality of my program by 300% and your algorithm only needs 100 new lines in one of the source files. You are totally free to add your hundred lines under terms that are compatible with my original license. You cannot proceed to remove or change my headers that specify my copyright and my license terms to the 10,000 lines of code of which I am the sole copyright holder.

As someone who's not too familiar with IP law/GPL versions: is GPL3 not "GPLv2 or later"?

They meant that you cannot once a piece of code is released by GPLv2 you cannot unrelease under GPLv2. So even if every new version and contribution is required to specify a GPLv3-or-later license all previous contributions and versions are still in GPLv2-or-later, which is more permissive.

I think the comment was addressing a misconception around how the project as a whole is licensed and its individual contributions or past versions are licensed and the fact that GPLv2-or-later works cannot be revoked as the user can always freely chose between the GPLv2 and the "later".

Rebranding a fork is absolutely normal. not "filing the serial number off". You realize that all git commit history is kept? They are not "scrubbing" Jonathan's work.

Do you even know what NSM is?

I believe the point was: “Their initial work was not fixes or changes they had tried to submit to NSM, but a rebranding effort.”

And no, I don’t know what NSM is, and I might be totally wrong on GPS intention.

From https://lists.linuxaudio.org/archives/linux-audio-dev/2021-F...

“We already had history with Jonathan; when we tried to do a few things that were useful to us but not within his design, sometimes even minor things like allowing to use 64x64 icons or having an option to export the whole session as a tar file, we were met with insults.”

The full pull request history for non is available on the Internet Archive. The only one I could find from Filipe is this one: https://web.archive.org/web/20201122235832/https://github.co...

I'm not seeing any insults in there, just a detailed explanation from Jonathan for why he didn't want to merge it. I don't understand the full technical details, but it's clear he gave it a lot of thought and there was a difference in project philosophy.

Yeah, although if I read the thread right his first comment was also 3 years after the pull request was opened.

"Every time I look at LV2, I am repulsed. LADSPA is KISS, one header file dependency and relatively straight-forward. LV2 is like the opposite of KISS. It's like FART: Forget About Representing Terseness. It reminds me of something from IBM. The dependency is like 6 different little libraries that no one will ever use for anything else. And all that this complexity really buys you is shitty in-process GUIs that crash your host program."

Lol. That's what made his software so great, he kept it non. But I can see the problem when others start wanting to add features via bloat. Two contrasting desires, does he stick with the thing that made his software great, or add any feature someone wants?

falkTX maintains a fork of Non-Mixer that adds LV2 support;

https://aur.archlinux.org/packages/non-mixer-lv2-git

Just FYI, I don't think any of the claimed heated threads were captured by the internet archive's sweeps. There were a couple linked to post-fork which were either archived too early, too late, or not at all.

disclaimer - I am one of the github users in the linked thread.

The nature of what different people consider insulting seems to vary greatly. I'd like to judge these supposed insults for myself. Have they been archived?

To the best of my knowledge, no.

Why would they remove the timeline, mixer, and sequencer?

The original upstream was more or less a monorepo of several subprojects. The fork only involved the session manager, so the other pieces were removed from the forked repository.

Reply on news.ycombinator.com