The "Devuan" Debian fork

> improvements in FreeBSD's init system, effectively making it more similar to systemd

There are already plenty of dependency-based boot managers. Technically systemd is nothing new, just NIH. And monolithic non-portable code can be found all the way back to the software stone age.

Nor is Embrace, Extend, and Extinguish new. Microsoft started using EEE decades ago.

FreeBSD may well implement an init system with similar functionality to systemd, but with modularity and without EEE it will not be at all similar to systemd.

You're mischaracterising people with a very broad brush here. Merely being of the opinion that systemd is badly designed and implemented does not automatically mean one is a backward luddite who won't "live with the times". This unnecessary personalisation of the discussion is one of the factors which have reduced the debate to a low level and made this sorry mess in the first place. It's one of the reasons I unsubscribed from the main lists a year back and greatly reduced my involvement (RSI was the other); I'm a volunteer who chooses to dedicate a significant portion of my life to making Debian, and being subject to continued abuse over a multi-year period is just not acceptable. I've got other free software projects I can work on instead which also provide me pleasure and accomplishment without involving disrespectful and rude behaviour and the resulting high levels of stress and upset. Looking occasionally at the list archives, the tone of the discussion has only become worse.

As one of the sysvinit/initscripts maintainers, I'd say that we were well aware that while they had served us well for decades they did have certain shortcomings and that a replacement which solved these would have been well received. We were in no way opposed to a quality replacement so long as it was backward-compatible to allow upgrades without loss of function or configuration. Three years ago, I was hopeful that systemd had the potential to be that replacement, and spent some time investigating it. For reasons which have been tirelessly debated and which I will not reiterate here, I did not ultimately feel that systemd was suitable as a replacement, and in the three years since then the course of systemd development has not changed my opinion on the matter; if anything it has further reinforced it.

The need for this fork could have been avoided entirely if certain developers were willing to accommodate continued use of the old init. But willing cooperation is not part of some developers' agendas, and that is a great shame since it is that which allowed us to build such a high quality and comprehensive integrated software distribution in the first place.
[To put this into perspective, I spent many man days of effort in the wheezy release making sysvinit align better with systemd to aid migrations and interoperability both ways; that's what you do when the goal is a well integrated distribution. I spent several solid days just making sure that hwclock timezone settings worked in every possible scenario. I think it's fair to say in retrospect that this was ultimately a fairly one-sided effort.]

While I still use Debian daily at both work and home, most of my systems have been running FreeBSD since February. I've needed to evaluate what the post-jessie migration paths might need to be. My plan is to switch over to Devuan once the dak instance and archive is set up, and participate in development at some level, though I don't know what the extent of that will be at this point. If it is small and focussed upon correcting this single major problem with Debian, I think it has a good chance of success, and we can hope that maybe the differences between the two distributions can be resolved at some point in the future. I'll continue to maintain stuff in Debian alongside this work.

Regards,
Roger

If you'll excuse the cliché, there's room for everyone at the party. Nobody's views or wishes need be excluded.

There's no need for the VUAs to "live with their times", as you put it, and likewise there's no reason for systemd proponents to not adopt systemd. The Devuan fork is not going to harm Debian, and may even be beneficial to Debian in the long term, as someone pointed out in another comment. There has already been a long history of Debian derivatives.

The majority of distributions have already switched to systemd. Surely having a selection of quality distributions which offer an alternative is a good thing? Why is choice suddenly bad?

I think there is indeed a part of the opposition having some kind of trauma with regards to Microsoft and their previous computing experiences. Reference to journald being "microsoft-like" due to binary format are common, and we can see reference at svchost.exe , conspiracy about companies trying to create a monopoly, etc.

I guess that for some people, the Microsoft hatred is a fundational brick of their free software involvement, and suddenly discovering that things are more complex than the binary manichean division of the world they had in mind shattered their views of the world, who had to be readjusted to the new changes, and readjusted using schemas of the past. They maybe are trying to make sense on what they didn't like without knowing why, being fueled by the fears of their pasts, about frustrating experiences they didn't fully grasp.

So I am not sure what they need exactly, besides people to spend time to hear them and try to understand what they really fear. The kind of customer services jobs that the free software have always been good at avoiding, by not caring and chasing people who would be good for that, by showing that only technical contributions matters, forgetting all the others.