And the beat goes on…
By now, much of the news and commentary is already out there about a fork of Debian called Devuan — pronounced Dev-One (sharp, folks) — and what it means to the newly minted systemd/anti-systemd rift in the FOSS world. I can’t add anything to the news part, but leave it to me to add to the commentary.
Forking is commonplace in the FOSS world, a part of its natural process. Someone thinks they can do something better — or it may be a group of folks of like mind thinking they can do something better — and they do it for reasons ranging from rational improvement to unabashed ragequit.
So personally, I wish this project luck. They’re going to need it.
This fork — seemingly a reaction by a group of developers who continuously battered the democratic process in the Debian community to derail the systemd process, like Republicans in Congress trying consistently in vain (at least so far) to repeal Obamacare – “aims to be a base distribution whose mission is protect the freedom of its community of users and developers.” Protecting the freedom of its community against what is a little unclear in that statement. Anyway, they continue: “Its priority is to enable diversity, interoperability and backward compatibility for existing Debian users and downstream distributions willing to preserve Init freedom.”
Oh. Init freedom. The God-given freedom to Init, for which many have fought and died. Why didn’t I think of that?
I have no problem with systemd and, like it or not, I think it represents progress. Those who are more knowledgeable than me have concerns that, no doubt, will be addressed — concerns around security and the inherent problem that a new program has to tackle a wide array of bugs. Some even bemoan the whole upheaval between pro- and anti-systemd partisans, which places FOSS in a position for a few rocky years ahead. But I think in the long run, whatever systemd flaws are now on the surface — both theoretically and technically — will be ironed out in due time and eventually it will be widely embraced.
As for Devuan, I’m sorry that I don’t have a lot of faith in their endeavor being successful. Successful projects usually aren’t the ones that are driven by anger or attempts to unreasonably hold onto the past. At the risk of sounding Californian, there are not a lot of good vibes or warm fuzzies that would normally draw people to such a project, so it will be interesting to see how many developers get behind this effort.
Something that hasn’t been addressed too much is the slap in the face to Debian, the real victim here. Debian has always been a beacon of FOSS excellence and the perfect FOSS citizen. Arguably this distro is weakened – hopefully not too much — thanks to a group of developers who have, in effect, tossed their toys around the playroom in a tantrum because things didn’t go their way. This should be a sobering indication of what to expect from these developers in the new project they’ve decided to pursue.
Clearly, systemd can be improved – anything this new could stand improvement, and chances are it will get better – and clearly the systemd project is not above criticism, which it will get from the more thoughtful of us in the wider FOSS community. Conversely, Devuan developers are clearly within their rights to follow a basic FOSS tenet of making improvements, either real or imagined, where they deem necessary.
And the beat goes on…
Larry Cafiero, a.k.a. Larry the Free Software Guy, is a journalist and a Free/Open Source Software advocate. He is involved in several FOSS projects and serves as the publicity chair for the Southern California Linux Expo. Follow him on Twitter: @lcafiero
And here I was hoping to read a well reasoned article that expanded on the technical merits.
The author is very rude, and insulting. He also fails to grasp the most basic element of Unix philosophy, and why it’s important. If you have to resort to straw man arguments and character assassination, then either you don’t have a point, or you’re just unable to understand the point being made by others. In this case, he simply doesn’t understand: do one thing and do it well. Systemd doesn’t do that, and it’s now expanding It’s tentacles into other areas that it has no real business in, like implementing a resolver. We have bind. We don’t need systemd to retake over bind. We need it to init. The only advantage I see is parallel daemon startup. This could be done without taking a wrecking ball to things that Poettering and friends don’t fully understand. It seems, neither does Larry Cafiero.
Just like GM food and nuke power there are upsides and downsides to this systemd/Init question. I wish Debian well and I will try Devuan asap. Larry, hey, it is another distro and some one I know always is saying “Choice is good”.
I have been with Debian or Debian based distros since I started
with Linux. There has been a lot of drama in the Debian Community. Very smart people with big egos are always a problem.
I say we appoint the Free Software Guy to be our research team on this, he can sneek over the mountain and check out what Silly Valley is thinking or doing.
As Disappointed said, it looks like you, my friend, have no clue what you are talking about. Good luck with your blog. Bye.
It absolutely _isn’t_ a slap in the face of Debian. It shows how robust open source projects can be. So yes if you really believed in open source you would want it to succeed.
The direction that Debian is taking isn’t liked by a group, so they are forking. That shows one of the main strengths of FOSS, so rather then playing the pro-systemd political game, try thinking and supporting FOSS!
I spent five days on the Devuan mailing list and I’ve seen enough crap…
Like Larry says, they are going to need an awful lot of luck.
In these 5 days, I’ve received 337 mails, of which less than 10% was actually worthwhile to read.
There was this anti-government type of guy, who kept talking about keeping Devuan *pure* of systemd. There was this guy talking about a possible fraud with the donations, not accepting a non-profit foundation and pgp keys of the maintainer as an answer.
There was this misogynistic “videogame programmer” and GamerGate supporter, and with every message he posted it became more clear that he has something against systemd because Poettering is married to an adult woman and he wants to marry a young girl (read: child).
There were posts about logo design, posts about the Google Safe Browsing present in Firefox, posts about what 4chan’s /g/ says about Devuan, posts about whether or not users should have a choice to use systemd, more posts about logo design, posts about whether or not Devuan is a good name, supplying other names.
There were posts bashing other users for earlier posts on the list. I’ve seen personal attacks, etc etc etc.
And there is far more of this crap down the rabbit hole where it all came from.
A systems design based on each part being simple in its essence that it does one job well, will obviously be destroyed by adding an element that flies against the overall design. Are we to ignore the intellectual foundations of unix as a philosphical system of design, and upon a whim introduce elements which by there very nature are alien to the system itself. This is madness indeed.
It is like a man who professes some belief regarding the world and teaches his philosophy to others. He lives in harmony with the world of Man and Nature and is at peace. But others see this as a threat, and bit by bit they convince him to adopt a perspective that flies against his inner core, that makes him contradictory. Then they just sit back and watch him fall apart. Make no mistake they want to destroy him less his philosophical teachings spread.
As a Slackware user on both servers and desktops, here’s why I have a problem with systemd.
http://www.microlinux.fr/download/systemd.jpg
We may have a different definition of warm and fuzzy.
Gee Larry, ya wanna be a little more biased? I for one don’t like the sound of systemd at all, and I’ve only been a Linux user for 10 years. It’s my understanding that the GTK/GNOME developers twisted the arms of everyone else with Debian to get it implemented. They already screwed up GNOME 3, which was bad enough. GNOME 2.x was great — but given how the current group of GNOME developers have adopted a “my way or the highway” attitude and have botched things up as a result, I have NO use whatsoever for them or anything they produce — NONE — and this includes systemd. They’re just plain becoming too much like Microsoft or Apple for my tastes and not listening to the wants and wishes of us, the end users.
> As for Devuan, I’m sorry that I don’t have a lot of faith in their endeavor being successful.
I’m personally the most excited that I have been over an FOSS project in almost two years. To the point where I’m following Devuan VERY closely. If it’s exciting me, who else could it be exciting?
Something to ponder. You may see systemd as progression, to a degree so do I. That doesn’t mean this isn’t a really big deal, with the collective talent pool between the few projects that are refusing to embrace systemd there’s plenty of capacity to pick up and work on the projects that systemd left behind in its wake.
I find it interesting enough that I may even participate, or maybe I’ll dust off some old code and start contributing upstream..and downstream again.
Maybe its time to cut the passive aggressive ‘good’ people of California routine Larry? Your post is very angry, and very much throwing your toys out of the pram because perhaps the majority of the Linux community is not awe struck by the magnificence of RedHats latest attempt to rewire linux interfaces for there narrow business interest to the detriment of non RedHat projects. the small group of RedHat financed projects has taken steps to intentionally break compatibility with non systemd non RedHat code in an attempt to force adoption. its grotesque Larry. and is a wakeup call to FLOSS developers and projects who want an open ecosystem where the best ideas and technology win.
there has been nothing but vitrol spit at the open source community by a small group of social media pundits and bloggers who have wrapped themselves in the FLOSS flag but seem far more concerned with defending RedHats interests. even if that means attacking FLOSS principles and spreading FUD. systemd and its interfaces were misguided and short sighted. and will not be adopted by a large portion of FLOSS developers and distributions who see it as a fundamentally flawed tautology. Devuan is just one of such projects and we should all support them.
Systemd is not just new and buggy…it is the worst kind of design failure. Feature creep and revinventing the wheel permeate it beyond all reason. But the worst part is not that it suffers from these things through accident, but rather that it suffers from them BY DESIGN by the admission of Lennart Poettering, in an effort to supplant the rest of Linux userland and become the defacto platform for all of linux-dom. Monoculture is a bad thing and I respect the developers who are working on Devuan.
Developers have the right to work on whatever they wish, and if Debian isn’t going the direction they like they are free to pursue other options. Who are you Larry, to suggest that is somehow wrong?
I’ve already had two bad experiences with systemd. One of the pluses of systemd was supposed to be faster booting. Alright. I tried it. It doubled the boot time of one system, insisting that apache finish starting up before xdm could start… On another, I plugged a thin client into my network. I got it to boot PXE and run very well in a few minutes, with LTSP. There were some graphics glitches but I was using Jessie/testing and that’s to be expected. I downgraded to Wheezy and what was working, sound, quit. The problem? Not systemd, I pulled that out. pulseaudio, another creation of Poettering, is tied into X-authentication and that took a lot of work to fix. Now I have a thin client that works perfectly with no errors, managers working, volume-controls working, and no sound and no diagnostics. This is stuff that was trivial a decade ago with esound.
These are all examples of why systemd should have stayed in testing for another couple of years and never made it into the default for the next release. Debian GNU/Linux was a great distro. It’s due for a fall with this decision. I can stay with Wheezy, I guess, for a while until I see whether it all comes tumbling down or gets fixed. Beta software is supposed to be fixed before being in the stable release.
These two examples are the sort of thing the UNIX philosophy prevents, single points of failure. Having one init sytem in the works is OK but it damned well better work perfectly. All I’ve seen so far is inflexibility and unreliability.
On the “UNIX Philosophy.” This is 2014. No one is developing operating systems and applications on PDP-7 minicomputers with paper drives anymore. BSD is UNIX. Allow Linux to evolve to solve modern computing problems using today’s every more powerful hardware.
I don’t think the situation with the promulgation of systemd and the politics surrounding the Juggernaut are quite as you describe (indeed, I don’t believe systemd has been gaining ground or “market share” purely on, or even largely on, its supposed technical merits), but you did bring up an interesting argument nonetheless.
I’m very curious how you would compare this systemd fracas with the fork from xfree86 to Xorg, which I understand was similarly contentious.
The UNIX Philosophy is just as relevant today. I would argue the need for it today is even greater than ever. In the face of the ever increasing technical complexity of modern computing the need for simple well-defined interfaces increases even more. We need more modular systems, not more monolithic ones.
I am a developer and I know “do one thing and do it well” has a strong parallel in modern software design practices – called the Single Responsibility Principle. https://en.wikipedia.org/wiki/Single_responsibility_principle
It isn’t just greybeards resistant to change that dislike systemd despite what the systemd camp would have people believe. There are plenty of younger, forward thinking, smart people who think systemd is a big fat pile of crap. Just like Microsoft found out with Windows 8, not all change is progress.
I has been a Fedora user for most than a decade. Fedora was the first distribution using Systemd. Until now, Systemd has done an overall good job, but I miss the days where I had the control from my init system with SysV.
The good news: My system is one minute faster in start up :).
The not so good: I restart my system once in a month :(.
Anyway Systemd has some good points and maybe will be suitable for most users, but for sure, not for **all** users.
I am not too worried about Systemd breaking Unix philosophy, or about lost of control, at this moment I don’t care the good points it can have. What real worries to me, is that Systemd has not been developed as an option or choice as originally intended; has been developed for take all Linux ecosystem, and that is not acceptable for me even been a Fedora user and admiring what Red Hat overall did for FOSS.
Free software is much about choice and control. If you don’t have it, there will be not much difference if you use Linux, Windows or OS X. Is sad to see many Open Source supporters does not appear to see in that way, and preach something like Borgs saying :”Surrender, Resistance is Futile”
I don’t know if Devuan will have a success or not, but I know I am not alone, and there will be a market for people who like its freedom.
BTW: I am trying Slackware distro right now :).
I hope both forks prosper. Because there is going to be a home for the Windows refugees and since they have the numbers they get Debian and RedHat/Fedora. Us UNIX folk are now looking for a Linux home. If this fork doesn’t get off the ground more will try until one or more succeeds. Where possible the two camps will share code.. until the two are so far apart it will be more of a port effort, then less so.
And that is why I will be going to one of the UNIX like forks, Pottering is an avowed ‘UNIX Hater’ and intends to eliminate every trace of UNIX from his fork. I’m having none of that. In a more sane world he would have went and helped a project that was already closer to his preferred design, ReactOS, Haiku or whatever. But RedHat won’t pay for that and besides, none of those projects are big enough for his ego.
I think this article is wrong. The backlash against systemd is growing. It’s a poor idea from an engineering point of view as well from a philosophical one. It controls too much and it seems to want to control the entire OS. Mr Poettering may have lied – initially it was only to be an init replacement. Now it’s expanding and taking over more and more functionality.
Manjaro have just released an unofficial, yet functional, build of their xfce distro using openRC instead of systemd. This for me is great as it means all my machines are now systemd free. This is how Linux should be – able to say “no, we don’t like that bit, so here’s an alternative”.
Systemd is a power grab. It puts more or less the entire community into the hands of one Corporation. The rug has not been pulled from beneath our feet yet, but now almost everyone is standing on the same rug, so to speak, and as such are now vulnerable. A license change is all it would take. Or perhaps Red Hat might declare that systemd will be free for home use but must be paid for in commercial settings. Whether they intend this or not is beside the point. They are in a position to do so. This is a colossal weakness for every distro.
The response of the tech press in this is quite interesting. It’s behaviour closely resembles the way the main stream media responds to people who think that Corporations should pay tax, or that Governments should be accountable – anyone speaking in that way is ridiculed, likened to fools, old men afraid of change, simpletons who ‘just don’t get it’, hopeless idealists, etc. This is how dissent is silenced – by bullying and ridicule. This is how the Corporations operate – by assaulting common sense and then bullying anyone who tries to point it out by using a host of ‘journalists’ and celebrities to push the message home – resistance is futile, accept the change.
No Sir, I will not.
Linux was created by thousands of people, working together mostly. Now one person intends to take it over and you’re all OK with this?
From this user the answer is “no, never”. Now there’s Slackware, Crux, Pisi, Manjaro and Devuan. A couple of weeks ago there were only 3, now there’s 5.
Resistance to tyranny is never futile.
Read ‘Kernighan and Plauger, Software Tools’. Then you will understand why systemd is abhorrent to the UNIX cohort (1970’s on). I said this about X, and I was right then too. Monolith is always the wrong way to write software. By now we should also have ditched the Linux Kernel for a messaging micro-kernel or something like it.
Full of anti-systemd posters as usual, small ranting minority with loud voices with no real knowledge of systemd configuration, its getting boring.
@ Colonel Panik
“Very smart people with big egos are always a problem.” Ilike that very much.
In another two years, you won’t be able to install a Linux distro without it requiring a GUI. No more CLI-only installs.
Man. Sure took long enough for the community to look like Microsoft.
There was an earlier “battle” between Andrew Tannenbaum and Linus Torvalds — microkernal vs monolithic kernel. Linus stuck to his guns and Linux is what it is today. Systemd is in sync with the LINUX philosophy. Linux is not UNIX. Linux resembles UNIX, but the design objectives are different.
“There was this misogynistic “videogame programmer” and GamerGate supporter, and with every message he posted it became more clear that he has something against systemd because Poettering is married to an adult woman and he wants to marry a young girl (read: child).”
Allowed in the Old Testament (Deuteronomy 22 28-29 in hebrew).
Suppressed in feminist countries, non feminist countries are bombed.
Many young girls are pretty. The japanese had the right idea. The hebrews had the right idea. The muslims had the right idea. The medieval french had the right idea.
You have prisons and schools and the willingness to bomb any culture that doesn’t obey your pro-woman ways.
“There was an earlier “battle” between Andrew Tannenbaum and Linus Torvalds — microkernal vs monolithic kernel. Linus stuck to his guns and Linux is what it is today. Systemd is in sync with the LINUX philosophy. Linux is not UNIX. Linux resembles UNIX, but the design objectives are different.”
AKA no philosophy.
Thank you and yours for murdering the bosnian man btw, he /really/ deserved to be beaten with hammers because a man who shared a similar skin colour shot a black man.
#JusticeForZemir
” By now we should also have ditched the Linux Kernel for a messaging micro-kernel or something like it.”
There are two good things about the Linux kernel:
1) though a fault in a single driver can and does crash the whole system, can and does create root exploits, it has a lot of those drivers for alot of hardware.
2) PaX / GRSecurity patch.
If Tannenbaum’s OS had just number 2 it would be the way to go over Linus’s kernel.
“Man. Sure took long enough for the community to look like Microsoft.”
When the women come in, the men go out
Apparently, most above commentators are philosophically opposed to or actually despise systemd, versus those expressing their clear preference for the protest-distro Devuan based upon its own merits.
Although a few above commentators are indeed hopeful that Devuan can and will succeed, there remain several key unanswered questions concerning the progress of Devuan development beyond optimistic blind faith. For example:
1. Sure, there is all the handwaving going on in https://devuan.org/ about all these supposedly great things to look forward to in this great, systemd-free, Debian fork.
However, besides eliminating systemd as an init system, what other objectives and accomplishment do Devuan’s developers REALISTICALLY hope to achieve in removing ALL systemd dependencies from the Linux kernel and userspace (D-Bus, GNOME3, … etc.) ????
2. What specific timecourse and project planning is REALISTICALLY set out in the “planning phase(s)” of Devuan’s https://github.com/devuan site?? As of this writing, the Devuan project’s milestones might be considered fairly nebulous at best.
3. What high-quality team is becoming established to advance development of Devuan?? Related to this and perhaps similarly important, are Ian Jackson, Joey Hess, and other key Debian GNU/Linux members really joining up to help with Devuan’s fledgling efforts?? For a point of reference, Devuan’s Dng mailing-list is https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
4. According to Devuan’s donation site https://devuan.org/donate.html , the project’s infrastructure funding currently sits at less than 2k USD. What will happen should Devuan’s crowdfunding efforts fail to reach their fundraising goal(s)?? Will Devuan development cease for a defined amount of time or perhaps even indefinitely???
These and other hard questions must be answered to determine how viable Devuan REALLY will be.
Anyone who think the big-party Republicans want to repeal Obamacare is too ignorant to post on any subject, with the possible exception of who tired their shoes this morning.
If you don’t have at least 200/300 servers running, you have no idea about what you are talking about! All cool for the fanboys which run Debian and Fedora… But for the hardcore world with 3K servers (my reality), systemd isn’t good. Of course fanboys and Linux VM users can eat whatever they’re fed with… We don’t.
Daily reminder that systemd is FOSS software and that it is good software that caters to the needs of modern devops, sysadmins,regular users and application developers. Since work on systemd restarted again in 2010 and its initial inclusion in Fedora in January 2011 it has gathered over 500 contributors of code and has become default in all major distros since 2012 and has devs from each major distro with commit access who have helped shape systemd the past 4 years.
Unfortunately as usual for the past 15 years since Linux gained major traction and contributions the bitter and violent *BSD cancerous userbase continue to try and cripple GNU/Linux, they leach off of all the GPL work done by Linux developers and spit in the face of them, insult GPL, and threaten violence and rape of GNU/Linux users and developers.
GNU/Linux depending on systemd is worse than anything so far for anti-linux *BSD lunatics, if applications make use of the way systemd exposes Linux-specific kernel features then *BSD ports/repositories suddenly start to become very empty because all the work is done by GNU/Linux users and developers.
BSD ports (ports means ported, leached, stolen packages) starts to empty or become filled with even more insecure, outdated and broken packages. This would not be the case if *BSD losers were not lazy hypocrites, but these cancerous excuses for human life cannot code and instead try to poison and cripple GNU/Linux out of jealousy by means of delusional social and philosophical engineering (much like how scientology cons people out of their money).
There are also thought to be many of these people on the payroll of Microsoft to try and destroy strong powerful FOSS projects by negative campaigning and lie/FUD spreading. Fortunately as usual for Microsoft their FUD campaign and paid shills turn up 4 years too late and don’t have technical arguments, making it obvious what they are: paid trolls.
The author said:
I have no problem with systemd and, like it or not, I think it represents progress. Those who are more knowledgeable than me have concerns that, no doubt, will be addressed —
So like myself we do not have an inner understanding of systemd, but at least you have to know what are those extra functions of systemd that people are crying about.
And then:
As for Devuan, I’m sorry that I don’t have a lot of faith in their endeavor being successful. Successful projects usually aren’t the ones that are driven by anger or attempts to unreasonably hold onto the past.
The first statement above clearly strips the author author from the ability to determine the faith of a fork.
“Since work on systemd restarted again in 2010 and its initial inclusion in Fedora in January 2011 it has gathered over 500 contributors of code”
Latrines have lots of contributors too.
Those who invoke the Unix philosophy don’t really seem to get it. It isn’t just “do one thing and do it well”. It is “do one thing, do it well AND play nicely with other tools”. The reason why Unix tools like grep, sed and awk are so useful is because they all adhere to a consistent interface so that you can use them together and pipe the output of one into another. That’s what systemd aims to do: provide a coherent set of tools that play nicely with each other. It isn’t about reinventing the wheel. It is about having all tools speak the same language so that they can work together.
The Linux kernel is also a single point of failure, yet nobody complains about it. Know why? Because it works well. In the end systemd will stand or fall on its technical merit. If it does its job right, fine. If it doesn’t, eventually someone will build a better alternative, and no evil corporation or developers cabal will be able to stop it. That’s the strength of free software, the best code always wins.
“The Linux kernel is also a single point of failure, yet nobody complains about it. Know why? Because it works well.”
Bullshit. The constant root exploits in every minor version are complained about. The less and less stability is complained about.
You, Onan, are a Liar.
Yup. Because prior to systemd, nothing on a Linux system was able to work with anything else on a Linux system.
Oh, wait, the whole reason Linux has gained such popularity and strength is because it worked so well.
So why are you now supporting unnecessary changes?
It’s pure sales talk – rubbishing an old product in order to sell a new one. That the old one worked perfectly well is of no concern because all that matters is selling a new one.
dv said: “Apparently, most above commentators are philosophically opposed to or actually despise systemd, versus those expressing their clear preference for the protest-distro Devuan based upon its own merits.
Although a few above commentators are indeed hopeful that Devuan can and will succeed, there remain several key unanswered questions concerning the progress of Devuan development beyond optimistic blind faith.”
Devuan may succeed or fail…that isn’t really the point for people who don’t like systemd. We simply applaud the fact that someone is willing to do the work to give people the option not to use it. Currently in Debian even making SysVinit PID 1 does not get rid of systemd’s influence. Thankfully, there are multiple options available. Slackware and Gentoo are viable alternatives.
Onan the Barbarian said: “That’s what systemd aims to do: provide a coherent set of tools that play nicely with each other.”
No it doesn’t. It replaces many working tools with half-baked alternatives in an effort to control the direction of development. People think FOSS always ensures freedom, but the reality is if you have enough sustained effort (say from a large corporation) you can control development through sheer velocity of code changes…unless you piss off people enough for a viable fork to take off.