[Consortium] LA.org "official fork" software
Sat Feb 13 10:53:06 CET 2021
There was no consensus about the fork. Not by the original author, nor by core community members. Robin, you describe a picture where there was. This is just not true. Totally the opposite instead actually. Robin, was not involved in the discussion. I know, because I was.
I've been probably the most active advocate of NSM since it's existence. I'm even today responsible for advocating NSM support in applications like Seq66 and Radium and most feature request for NSM support are coming from me. I did a lot of testing for developers during their implementation of NSM. I recently reported NSM related issues for Ardour, Qtractor, AMS, Amsynth, Qmidiarp and more. Most of them initiated a bugfix. It will be hard to find someone else with the same passion for NSM and it's design. Does it matter? Not really, but it shows how I'm involved and that I know the situation very well.
There has been a active discussion about the technical aspects of NSM with the original author in the last half year or so, which I followed closely. The original author and designer of the NSM session API had very good technical arguments to not accept suggested changes and patches. Moreover the changes could be easily implemented in a other way, in line with how NSM is designed. See for his own take of this dispute here: https://non.tuxfamily.org/wiki/News
So these persons didn't want to listen to the original designer of the NSM API and decided to fork against the will of the original author and some of the core members and most active users of the NSM community. Be aware of the fact that session management on linuxaudio has a long history of smart people trying to solve this problem. From LASH to Ladish to jacksession to Non Session Manager (NSM). The designer/original developer of NSM was finally able to solve this issue for the community satisfactory after many years. I just want to point out that this is a major accomplishment. And one of the pillars of his success was keeping NSM limited and to prevent 'feature creep' in it's core. In other words, he has the authority to judge patches one could say. Which he did with good argumentation.
Now tell me which version linuxaudio.org should promote?
You see, that's the problem. It’s highly questionable that linuxaudio.org now promotes this version. And do you know why linuxaudio.org promotes the fork, while it contains rejected patches by designer of NSM? Because what you call 'high profile members', have a role there. So they use these roles to choose the forked version. You see the issue of conflict of interest? But in fact it’s the other way around of course. They fork and they choose linuxaudio.org to promote it. They can because they’ve a role there. It’s a misuse of the linuxaudio.org consortium and it’s resources.
You say, it's just about a technical document, but then it would make more sense to choose for the original project. The developers of the fork didn't design or invent NSM. The original author did and he claims (I think rightfully) that the developers of the fork don't understand his design fully. He didn't reject the proposed patches out of fun. It has been done with good technical argumentation. As a active NSM user, I’m also against many of the proposed changes. These are suggested changes by people who don’t use NSM themselves heavily, who are not particular a proponent of a modular approach or are just less skilled programmers in this realm.
So it's clear that the three developers (actually one developer backed by two other persons) of the fork use linuxaudio.org to promote the fork above the original version. While it is really questionable that linuxaudio.org should promote this version and not the original if it should promote any version at all. Or maybe take a step back and ask whether there should be two versions of the NSM API at all. Should there be a fork at all or should they just develop a different GUI for NSM (asking the question is answering it... ).
You say that hosting on linuxaudio.org is good for preventing fragmentation. Isn't that exactly what the fork has done, causing fragmentation? Before the fork there was one API, one NSM session manager. They change the API, so they've made the linuxaudio session environment already more complex. So your argumentation goes wrong here.
You say that they've to use linuxaudio.org because otherwise they have to start a new organization like LV2 or Faust. But these are just original projects, just like there is Non Session Manager (NSM) for more then 15 years. With it's own website and wiki here: http://non.tuxfamily.org/
There is already a organization and a second organization, leads to fragmentation. Nobody wants a new-Faust officially released by linuxaudio.org, without consent of the original developers, believe me.
Then there is also the social aspect and the aspect of Free Software and the role of the linuxaudio.org consortium. Being active in the LAD IRC channels, does give you the right to just fork a project from a fellow LAD developer and host it on linuxaudio.org? Is that how the linuxaudio consortium should threat developers of the community which are of tremendous value for the community? Why would the inventor of something comparable to NSM, release it’s software as GPL for the community in the future, knowing that some guys of the linuxaudio.org consortium can ‘hijack’ your project like this? The linuxaudio.org consortium should stay away from these practices.
To conclude. Robin describes a situation where there was consensus about a 'updated version’ of NSM. I think I've shown that there was no consensus at all. I think I've shown why it is highly questionable for linuxaudio.org consortium to promote this particular version of NSM and not the original one. That it’s questionable that developers who have a role at the linuxaudio.org consortium, forks a project of a fellow LAD developer, who solved a major problem for the community. I think I've shown that the fork hosted on linuxaudio.org isn't preventing de-fragmentation, but causing it. I think I've shown that linuxaudio.org is promoting this particular version and not the original, because the people behind the fork are moderators of linuxaudio.org and so have access to it's resources like the github page, the domain name etc.
I think it's pretty clear that the linuxaudio.org consortium isn't neutral here anymore. This is bad for linuxaudio and for it's developers. Moreover, there is a clear conflict of interest by the three people behind the fork who use their roles as moderators of linuxaudio.org to promote one version of NSM (not the original), which is 'accidentally' their own version.
‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Friday, February 12, 2021 9:12 PM, Robin Gareus <robin at gareus.org> wrote:
> Hi all,
> I think it is entirely appropriate when high profile members of the
> linuxaudio consortium announce a standard and use linuxaudio.org to
> promote it.
> This is particularly true for a session-manager that many other
> consortium members and developers should ideally adopt.
> Updating NSM was a linuxaudio community effort that took place on #lad
> IRC by various community members after discussions over a long period of
> time. While there was no rigorous peer-review, I think it is properly
> After all this is about a technical standard, and not a person.
> The alternative would have been to start a new organization (like LV2
> jackaudio, FAUST etc). Yet this would likely fragment the community
> further, which isn't in the interest of linuxaudio.org.
> I think it would be nice if more standards and best practice documents
> are released by linuxaudio.org, particularly ones that improve
> As for this specific release, I agree the name was not chosen wisely,
> and I hope it can still be changed. Also other mistakes in the release
> notes were unfortunate. But let's not throw the baby out with the bathwater.
> On 2/10/21 12:17 AM, Spencer Jackson wrote:
> > It was asked that this discussion was moved to this list since it is meta
> > to the LAD list itself. I will attempt to comply though this list has been
> > inactive for nearly 3 years.
> > The recent unfortunate situation highlighted by Jonathon Liles' lengthy
> > rant in [LAD] regarding his Non-Session-Manager software and the fork
> > called New-Session-Manager does raise some questions.
> > Johnathon has a history of being difficult to work with, caustic in his
> > writing, and I can understand the motives to fork, especially considering
> > the lengthy time since any previous NSM release. I don't think the fork is
> > the issue nor particularly germane to the discussion. I imagine the
> > situation would have been avoided with some more cooperative effort from
> > Johnathon, but I also won't criticize him too harshly for having a vision
> > for his project and rejecting requests that didn't fit in that vision.
> > While I do not subscribe to Johnathon's assignment of malice and subterfuge
> > it does seem that the list moderators releasing a fork under the
> > linuxaudio.org brand is probably overreaching the consortium's mission: "to
> > promote and enable the use of Linux kernel based systems for professional
> > audio use."
> > I believe that releasing the software under the names of the authors'
> > rather than under the would have helped make this situation look less like
> > an attempt of the organization to replace a developers project without
> > acknowledgement of his contributions. I'd like to politely request the
> > authors of the New Session Manager refrain in the future from releasing
> > software as official linuxaudio.org versions. It seems appropriate to me
> > that this be written into the policies of the consortium to avoid future
> > instances where developers may feel their software is being replaced by the
> > organization itself.
> > I am assuming the consortium would like to remain project agnostic rather
> > than picking the winners and losers of which projects get the LA.org
> > blessing.
> > _spencer
> > Consortium mailing list
> > Consortium at lists.linuxaudio.org
> > https://lists.linuxaudio.org/listinfo/consortium
> Consortium mailing list
> Consortium at lists.linuxaudio.org
More information about the Consortium