Sorry, you need to enable JavaScript to visit this website.

The Poetterisation of GNU/Linux

Homer's picture

Poettering LinuxI've found a disturbing trend in GNU/Linux, where largely unaccountable cliques of developers unilaterally decide to make fundamental changes to the way it works, based on highly subjective and arrogant assumptions, then forge ahead with little regard to those who actually use the software, much less the well-established principles upon which that OS was originally built. The long litany of examples includes Ubuntu Unity, Gnome Shell, KDE 4, the /usr partition, SELinux, PolicyKit, Systemd, udev and PulseAudio, to name a few.

I hereby dub this phenomenon the "Poetterisation of GNU/Linux".

The broken features, creeping bloat, and in particular the unhealthy tendency toward more monolithic, less modular code in certain Free Software projects, is a very serious problem, and I have a very serious opposition to it. I abandoned Windows to get away from that sort of nonsense, I didn't expect to have to deal with it in GNU/Linux.

Clearly this situation is untenable.

The motivation for these arbitrary changes mostly seems to be rooted in the misguided concept of "popularity", which makes no sense at all for something that's purely academic and non-commercial in nature. More users does not equal more developers. Indeed more developers does not even necessarily equal more or faster progress. What's needed is more of the right sort of developers, or at least more of the existing developers to adopt the right methods.

This is the problem with distros like Ubuntu, as the most archetypal example. Shuttleworth pushed hard to attract more users, with heavy marketing and by making Ubuntu easy at all costs, but in so doing all he did was amass a huge burden, in the form of a large influx of users who were, by and large, purely consumers, not contributors.

As a result, many of those now using GNU/Linux are really just typical Microsoft or Apple consumers, with all the baggage that entails. They're certainly not assets of any kind. They have expectations forged in a world of proprietary licensing and commercially-motivated, consumer-oriented, Hollywood-style indoctrination, not academia. This is clearly evidenced by their belligerently hostile attitudes toward the GPL, FSF, GNU and Stallman himself, along with their utter contempt for security and other well-established UNIX paradigms, and their unhealthy predilection for proprietary software, meaningless aesthetics and hype.

Reading the Ubuntu forums is an exercise in courting abject despair, as one witnesses an ignorant hoard demand GNU/Linux be mutated into the bastard son of Windows and Mac OS X. And Shuttleworth, it seems, is only too happy to oblige, eagerly assisted by his counterparts on other distros and upstream projects, such as Lennart Poettering and Richard Hughes, the former of whom has somehow convinced every distro to mutate the Linux startup process into a hideous monolithic blob, and the latter of whom successfully managed to undermine 40 years of UNIX security in a single stroke, by obliterating the principle that unprivileged users should not be allowed to install software system-wide.

GNU/Linux does not need such people, indeed it needs to get rid of them as a matter of extreme urgency. This is especially true when those people are former (or even current) Windows programmers, because they not only bring with them their indoctrinated expectations, misguided ideologies and flawed methods, but worse still they actually implement them, thus destroying GNU/Linux from within.

Perhaps the most startling example of this was the Mono and Moonlight projects, which not only burdened GNU/Linux with all sorts of "IP" baggage, but instigated a sort of invasion of Microsoft "evangelists" and programmers, like a Trojan horse, who subsequently set about stuffing GNU/Linux with as much bloated, patent encumbered garbage as they could muster.

I was part of a group who campaigned relentlessly for years to oust these vermin and undermine support for Mono and Moonlight, and we were largely successful. Some have even suggested that my diatribes, articles and debates (with Miguel de Icaza and others) were instrumental in securing this victory, so clearly my efforts were not in vain.

Amassing a large user-base is a highly misguided aspiration for a purely academic field like Free Software. It really only makes sense if you're a commercial enterprise trying to make as much money as possible. The concept of "market share" is meaningless for something that's free (in the commercial sense).

Of course Canonical is also a commercial enterprise, but it has yet to break even, and all its income is derived through support contracts and affiliate deals, none of which depends on having a large number of Ubuntu users (the Ubuntu One service is cross-platform, for example).

What GNU/Linux needs is a small number of competent developers producing software to a high technical standard, who respect the well-established UNIX principles of security, efficiency, code correctness, logical semantics, structured programming, modularity, flexibility and engineering simplicity (a.k.a. the KISS Principle), just as any scientist or engineer in the field of computer science and software engineering should.

What it doesn't need is people who shrug their shoulders and bleat "disks are cheap".

Comments

Anony Mouse's picture

If this is the trend, and they did succeed, what now?

As someone who wants to get into Linux, what should I do now?

If what you've said is true, then which linux distro should I use?
Are there any "best" options, or is that a meaninglessly subjective thing to ask?
With respect to the systemd thing and other issues, is there any distro which hasn't caved in?

And if I ever wish to become a Linux dev, should I start with a more "mainstream" distro and then switch to a different one once I gain some experience?

Sorry if these questions are idiotic or redundant. @_@
I just don't trust other people on the internet, especially if they aren't aware of the issues you've mentioned here.

Homer's picture

re: "distro which hasn't caved in"

Yes and no.

There are still distros that do not use systemd as the default init system (if systemd can even be described as such, given that Poettering seems to want it to be an entire OS), and even a few that do not use it at all, but the problem is that no distro has the power to stop the Poetterisation of GNU/Linux at its source ... the upstream projects that make the components of every distro.

A significant portion of the Free Software ecosystem has now become poisoned by systemd (as Poettering intended), either by direct or indirect dependency. This means that, increasingly, there is simply no way to avoid it. For example, it is now impossible to run the Gnome desktop without systemd [ed: apparently now you can], although this is probably not that surprising, given that Poettering and the entire systemd cabal are also the core team for Gnome (and incidentally Red Hat/Fedora employees).

Of all the distros listed on without-systemd.org, Devuan is the only one that was created specifically for the purpose of avoiding systemd, so it seems like a very good long-term bet, especially as it's a fork of Debian, perhaps the world's best distro overall.

Personally I'm going to stick with Gentoo, which sadly ships systemd but doesn't configure it as the default. This is a meta-distro only suitable for the most experienced GNU/Linux users.

Arch used to have the best and most Unix-like philosophy (simplicity is the ultimate sophistication), but has since betrayed its principles and moved to the dark side with systemd - a bloated, monolithic abomination that's barely distinguishable from Windows, from an engineering perspective.

Ubuntu, in addition to being tainted with systemd, is essentially just a proprietary OS. At one point Canonical deliberately infected it with spyware. Avoid at all costs.

There are a lot of very interesting GNU/Linux distros out there that qualify, and they're all free (in both senses), so just try some out and see what you like, but ultimately you should probably opt for Devuan, because its Debian foundation means that you're pretty much guaranteed compatibility with everything, including proprietary GNU/Linux software, should you want it. Steam is developed on Debian, for example, and indeed SteamOS is simply a fork of Debian shipped with Steam (but sadly also includes systemd). Steam itself does not actually need systemd, though, and neither does any other commercial GNU/Linux software that I'm aware of.

Good luck.

Anony Mouse's picture

My 0x02...

So sadly I come late to this discussion...and I'm sad to admit, late to understanding the whole systemd thing. I -am- ashamed to admit that I've only just now looked into what it means, since it just appeared in Ubuntu 15.04. I'm horrified by what it represents just from a security standpoint..systemd (and all it's many tentacles) looks almost like a ready-made rootkit just waiting to happen. Someone at the NSA is no doubt already on this.

But regardless; maybe I've gotten lazy. Been using Linux since 1992...yes that long ago, starting with the long-forgotten Softlanding System (SLS) distribution. That's a long time. I've seen a lot of changes. Seen a lot of people come and go. But it wasn't until one singular event and one small cabal of people that I noticed things were starting to change, and not for the better. I'm talking, of course, of the appearance of the GNOME project and Miguel D'Icaza in particular. That one person could spread such discontent and malaise was almost unheard of until that point. And then RedHat started showing it's teeth at the same time. It's not a surprise that GNOME and RedHat became such BFF's. I suspected back then that there was an agenda at foot to take control of the direction Linux was taking. Bravo to you for taking the fight to Miguel over MONO. I got very weary on places like Slashdot trying to be a contrarian voice to his bullshit in the GNOME wars. I still haven't forgiven Slashdot for their large role in the maligning of the KDE project and the very nice people who started it.

Back to my original point: now that I know what systemd is, I am absolutely horrified. It is wrong on almost every level. And just like GNOME in the early days, the PEOPLE involved with it totally rub me the wrong fucking way, especially this Linnart clown. But systemd is the same agenda...once again I see it's all about taking control of Linux in such a way that even Linus can't stop. It's a different set of assholes playing the same game for the same corporation: RedHat. And it's no surprise to me at all that GNOME helped cement systemd by "requiring" it and tying it to all kinds of unnecessary things.

And now for my tinfoil hat, which is well polished. I've been in a lot of OS-related flame wars. I was there for the invasion of the Microsoft Munchkins, back during the "Steve Barkto" OS/2 wars and later on Slashdot. I've know what an astroturfing campaign smells like. And when it comes to systemd, I'm starting to strongly suspect that these people did their homework and have had a team of astroturfers hard at work on every Linux-related forum out there. They're slick, they're arrogant and they all use the same catch phrase: "Well if you don't like it you can always switch to FreeBSD." Interesting that they don't say switch Linux distros...it's always "go to FreeBSD if you don't like it." And this betrays their badly hidden agenda..they want systemd on ALL linux distros. You don't get a say. (Why the hell would I want systemd on my Raspberry Pi?)

Systemd has been likened to an "invasion" more than once. Indeed, now that all the major distros have caved in, I'd say it's more like Invasion of the Body Snatchers. Have the pod people really won?

Anony Mouse's picture

We need to change our way to see things

Yes we do.
We also need to change our way to talk about them.

There's one distro becoming bigger and bigger, this distro is the Gnome-Systemd Linux (let's call it RHGSDL) and it has quite a lot of derivatives which will become more and more similar : RHEL, Fedora, Debian, Arch, Opensuse, Mandriva...

The core idea of the developpers of this distro is that the whole system must rely on a unique process which control every other thing and that package/software dev/maintainers will work for that distro only. They think that all maintainers will do that because it's 'easier' than to work for themselves and the whole community with other distros.
Maybe they are right to rely on lazyness and selfishness of those who make the strengh and diversity of GNU/Linux but, on the other hand, those who chose to give their time and skills in an open-source/free softwares unpaid work have maybe other higher motivations. Time will tell.

Time will tell also if the technical choices, and, among them, a single point of failure were good for this distro. Also their political choices. Some of the derivatives force it as default and do just not support alternatives, some are pretending to support alternatives while, by playing with recursive (sometimes 'forged') dependencies and break of compatibility, they make it almost mandatory once it's installed. And they are the first distro who openly wants to either eat or destroy any other distro not using their octopussy monster by orphaning/bashing them.

Still, RHGSDL is only a distro and we need to help and support the others who want to offer the choice to their users. I mean : LFS, Crux, GNU Guix, Gentoo, Slackware, PCLinuxOS, Refracta, Sorcerer, Source Mage, Void Linux, Plop Linux, Pisi Linux, Bedrock Linux, GoboLinux, Kali Linux, 0Linux, 4MLinux, Puppy Linux, Tiny Core and, maybe, LSD Linux, Trios and Devuan (if they ever become real forks), Maui project (still alpha but wayland based)...

Here are my 2 cents.

Lib

Anony Mouse's picture

lenartz pottering have to

lenartz pottering have to start use windows, stop broking linux

Anony Mouse's picture

lennart poettering is a dangerous vile dare I say evil whelp

lennart poettering is the santa claus politician of *nix , well just Linux because none of his garbage ports.

He promises stuff that does magic and its free, but the reality is hey can deliver, nor his sycophants and minions, anything that works.

He is poetter-izing linux, making it suck in the same ways as many commercial OSen.

I have to say systemd is the crappiest rip off of SMF I have ever seen.

Anony Mouse's picture

The Poetterisation of GNU/Linux --> The worse Operating System..

Homer:

good blog, good avocacy, thanks !
I do agree quite largely with what you write, however I would suggest that Open Source Including Linux is about freedom.
Like Democracy Linux may be "The worse [operating] System... at the exception of all others". (Thanks Wiston Churchill).

I am also a software guy, and have observed with dismay people writing overly bloated software, quite hard to fix, debug, and update.

I have also on occasions, been quite upset when someone modified my "Clean and lean" code, with "crapware", meaning quickly-hacked, bloated, and impossible to fix without complete re-writing, usually impractical, because management tend to be overly pragmatic. The result being often unmanageable code... that I had to manage !

The main reasons for bloat are in my view two-fold:

1-Lack of knowledge making it very tempting to use bloated building blocks, to solve (avoid coding !) simple, classic problems.
2-Vendors indulging and providing "bloatware" because it is one of the ingredients to create a monopoly.

"How to avoid bloatware" to get into cleanly written software ? A question well worth asking !
The answers will vary with opinions, and tools... and willigness !
However one of the key ingredients could be well documented software (Using Doxygen or equivalent, makes it easy!).
Like every developer, I do have little appetite for writing documentation, but it is important...
I would suggest that "students" willing to contribute reading existing code, could add Doxygen comments in the code.

On Linux:
Linus Torvald did a reasonably good job, keepin the system "clean an lean" as a benevolent dictator.
I wish he did the same with respect to system packages : too many incompatible versions ! (DEB, RPM, ... and more).
In my view, a common database (of dependencies) standard could and should (it may exist ?) be developed as part of LSB (Linux Standard Base).
This would help writing a Linux Application single package which works on every 2.x or 3.x version of Linux : it is not too late !

That was my 2c.
Andre G.

Anony Mouse's picture

torvalds = cool, lennart poettering = devil

torvalds built something awesome being a benign dictator.

lennart is a low life, a scum, and he tries to pervert *nix and linux to be unreliable, complex, non-working cruft.

Homer's picture

Sucks less

Yes, I agree with most of what you wrote. Perhaps I should have been clearer that I'm in no way suggesting GNU/Linux is worse than other operating systems, merely that I find certain trends within the community annoying (and potentially catastrophic in the long term).

I have to disagree about the supposed need to eradicate choice by consolidating all package managers into one, or all distros into one, or all of anything else into one. Any given choice only exists because at least one person somewhere wanted it, couldn't find it, and so created it, thereafter choosing to share it with others. Trying to eradicate that is like trying to eradicate individuality itself, which is not only impossible but actually rather sinister. Indeed my opposition to Poettering's attempted hostile takeover of the init system is exactly why I wrote this article in the first place.

Anony Mouse's picture

"I have to disagree about the supposed need to eradicate choice"

I am afraid to either have been unclear, or have been misunderstood here.
Never suggested a single package manager : quite the opposite !
I rather suggested a single format (API) to manage the database of packages, and dependencies, the same way that there is the LSB (Linux Standard Base) allowing a Linux application to be easilly ran on all Linux bad platform, wit a common GLIBC, and diretory tree, etc...

Current failure to do this results in tremendous waste of energy repackaging (and re-testing!) the same software for each Linux Distro or package format.

Actually the problem is somewhat different for binaries and source based packages.
Packages compatibilty requires a common API and ABI for packages, but this topic is not simple.

Failure to get there explains why several vendors (Nvidia, Broadcomm, etc...) prefer to release huge statically linked binaries, a pain to use (and update!).
Creating and testing binaries in multiple formats is costly, and difficult to maintain, especially for Kernel (not relying on glibc, etc...).
This unecessary burden on developpers, out a damp on Desktop GNU Linux wide acceptance, favors closed (or less open source) versions of Linux, like Chrome. Android. A shame.

Andre. G.

Homer's picture

The API is still a choice

If RPM, DEB, Portage etc. all used the same API then they would all essentially be the same package manager, since the API is the only real differentiator between that sort of software. Consolidating all those APIs into one would therefore have the effect of removing choice. I don't want RPM or DEB or anything other than Portage. That's one of the reasons I chose Gentoo in the first place. You would deny me that choice.

What you characterise as a "waste of energy" is actually other people's preference. They deliberately chose to do things differently, because they were dissatisfied with the existing paradigms. Exercising their freedom to do things the way they want to is never a "waste", it's simply their choice, and it isn't for anyone else to tell then not to. Moreover you have no way of stopping them anyway, so such discussions are quite futile.

In other words it's their energy to "waste", and nobody else is compelled to support their choices if they don't want to. That includes who proprietary software vendors choose to support, and conversely which proprietary software any Free Software maintainers choose to support.

If these proprietary software vendors find supporting different package managers so burdensome, then they're free to leave and take their proprietary garbage with them, I'm sure. Certainly I won't miss them. I will not have my choice of package manager dictated to by some proprietary software corporation. I came to GNU/Linux precisely to get away from that sort of manipulative nonsense.

Anony Mouse's picture

The basement boys ca cry all they want..

.. but if Linux is going to have any chance on the desktop the next decade, Poettering et al. are the way forward. I could not give a flying toss about anybodies affections towards legacy *nix implementations and non-sensical idealistic portability requirements.. Fuck that. That shit has put Linux desktop and Linux gaming (amongst others) on the back bench for more than a decade.

Anony Mouse's picture

Flying toss. Asshole.

You couldn't give a flying toss about what we care about? How about we find you and MAKE you care then you fucking asshole. Why do you think it relevant to us what YOU do or do not care about. The world revolves around you? I hope you are taught a painful lesson asshole.

Homer's picture

"chance on the desktop"

Unfortunately this "desktop" thing is rapidly dying, therefore the significance of exactly what will run on it over the next decade is moot. In a twist of irony, this also means that the thing which has actually put GNU/Linux "on the back bench for more than a decade", Microsoft's OEM racket, will also cease to exist within that same time frame, but sadly too late to be of any benefit to GNU/Linux, at least in terms of mass adoption.

But then I was never especially interested in such trivia as popularity. If you are, then I suggest you wear a tutu and join the cheerleading squad. Personally I'm more interested in the integrity of the software I use, and neither Poettering nor his dysfunctional, convoluted, proprietary abominations have any, therefore I won't be using them. Ever.

Anony Mouse's picture

glad to see...

I am glad to see that others find a problem with this trend.

I think however that Poettering and crew is worse than Ubuntu. At least Canonical are not pushing their projects onto Freedesktop and making it seem as some grand community project.

As best i can tell, Poetterning's projects as of late benefit only one entity within the Linux community. And that entity is Red Hat. And RH is focusing more and more on some kind of server clustering, that are quick to deploy and bring up.

And to obtain this they are going for tighter and tighter vertical integration, and tossing out anything that allow for options and alternate uses. this because they may cost a second or two extra during initial loading if enabled.

Anony Mouse's picture

"I've found a disturbing

"I've found a disturbing trend in GNU/Linux, where largely unaccountable cliques of developers unilaterally decide to make fundamental changes to the way it works,"

Isn't that what makes Linux and open source so great? The fact that developers can take a project and adapt it to their own vision.

Since you censored my other comment with your copyright hypocrisy, I wonder if you will do the same with this one.

I'd suggest that if you find projects you disagree with, you simply don't use them. There's plenty of choice out there and another reason why OSS is so great. In your world maybe you want software freedom as long as it agrees with your unique "ideas" - I wouldn't call that freedom at all and I'm pleased to see your type of hypocrisy is limited to a very small audience.

Homer's picture

The "freedom" to screw things up

Having the freedom to modify Free Software doesn't somehow negate my freedom to criticise it if it's crap, nor is it any excuse for making crap software in the first place. I have no power over these people, nor do I aspire to any, therefore I can't force them to stop making crap software, and I can't force them to stop destroying existing software that I depend on. However I can complain about it, especially when there are no viable alternatives.

If you'd actually read and understood the article, you'd realise that, for example, systemd threatens to supplant all existing alternative init systems (and software far beyond the init system too), by means of universal dependency, at which point there will not in fact be any "alternatives".

Before using big words like "hypocrisy", I suggest you learn what they mean first. There's no hypocrisy in criticising crap software, or a rapidly diminishing number of choices, or being subjected against my will by automatic copyright terms.

As for the other comment I supposedly "censored", I didn't, in fact I wasn't even aware there was a comment until now. Because I get an inordinate number of fuckwits like you spewing bile on this site, no comments are published automatically, they go into a moderation queue for review first, but unless I actually log in and look at that queue, I have no idea there are any comments waiting.

Not that I have any particular objection to censorship, in fact I strongly endorse it, just as I strongly endorse removing fuckwits like you from the gene pool.

Anony Mouse's picture

Indeed, We must see it for what it is: a fork of Gnu/Linux.

We have to see it for what it is. Lennart Pottering and his acolytes
who work within other projects are essentially forking Gnu/Linux and
are creating LennartOS.

What should be done is to follow their often spoken refrain: fork
every project that relies on systemd, xyzkit, LennartStuff:
create X.org-concrete (if they go the Lennart way (there's a patch now!))
KDE-concrete, Gnome2 Gnome3-concrete. Security fixes would be
applied, as-well as legitimate bug fixes (all these are often
one to three liners). Otherwise the software would mostly retain
its form. Something familiar, mature, stable.

Remember: Software is much more an art form than a science. The way
it interacts is a choice not a foregone conclusion.

Lennart and his followers use the fallacy that software is a
science to bully all of us into following their direction.

A fork of (all but) linux is need now, it's already happening
on Lennart's Side. We had better meet them in kind and protect our
software system or be subsumed.

Anony Mouse's picture

Is there any way out of this?

Is there any way out of this? Everything seems to be going systemd (apparently they all all-ready went dbus for some reason). There's now a patch for X.org for-which attention is being demanded. The pottering acolytes keep saying "oh well you don't have to use systemd, you can FORK all the projects that rely on it now!". Honestly when and why did the "let's all use some monolithic second kernel and waste as many resources as we can on constructing a new religion" idea take root? I haven't payed attention for a few years but suddenly linux isn't what it always was any more (or is on the verge). It's like a disease of the mind.

There needs to be a fork of the whole "stack" (what the fuck is that, I never heard of a talk of "stacks" years ago) into a pro-unix enclave, and whatever operating system pottering and followers are creating for their own glory. Just like XFree86 was forked, everything beneath (or above?, this stack terminology...) needs to be separated into branches for the pottering followers and the old unix (and old is good) guys.

Pottering is a fucking asshole. I watched him take over a man's presentation, and even hop on stage. The man should have beaten him bloody, it's clear he's never been on the receiving end.

He likes to scream "it's free, you don't HAVE to use it, and you get it for free". A bullet is free too on the receiving end. And one wouldn't have to receive it if one just stepped out of the way!

Pottering and his fans reminds me of the invasion of feminists a few years back. They came on with a bang, claimed they were doing great work, made demands, and they were obeyed to some degree. They didn't create much of anything. They got some open source projects of mine taken down however. They get men kicked out of various projects to this day for saying or believing the wrong thing.

Is there any hope (Please debian, don't force systemd, stay what you always were. And Ian you created Debian, don't let it be taken from you by pottering follwers)
--MikeeUSA--

Anony Mouse's picture

4 Freedoms are for USERS, not contributors.

>but in so doing all he did was amass a huge burden, in the form of a large influx of users who were,
>by and large, purely consumers, not contributors.

I agree with some of what you say and not with others but you just lost me at this quote.

Listen to Stallman speak about the GPL, GNU and freedom and NEVER does he make your statement.
If anything Ive heard him many times say the exact opposite: users DONT have to be developers and what he and others do.

your quote is a mix of condescencion and arrogance..

As a developer, i dont expect users to be contributors.

Homer's picture

4 Freedoms?

Where did I mention the four freedoms?

I'm talking about technical inadequacies driven by misguided popularisation, not licensing.

Also, I don't expect Stallman to make my statements for me, I'm quite capable of making my own, thanks.

And no, I don't expect the great unwashed hordes of ex-Windows users to suddenly become GNU/Linux developers, and I don't want them to, but then neither do I expect actual GNU/Linux developers to slavishly pander to their Windows-centric demands.