Ecce Lennart

[Sarge-in-charge]

Linux is the last frontier where the user has a real say in determining those relationships.

Linux users have less say in the direction of Linux, than Windows users have in the direction of Windows.

It is Linux developers (be it kernel, or userland) who have ALL the power. They answer to none but themselves.

At least Microsoft with Windows has to care for the users, in order to stop losing market share and thererore, to stop losing revenue.

[Roel63]

Linux users have less say in the direction of Linux, than Windows users have in the direction of Windows.

It is Linux developers (be it kernel, or userland) who have ALL the power. They answer to none but themselves.

At least Microsoft with Windows has to care for the users, in order to stop losing market share and thererore, to stop losing revenue.

Interesting. When Gnome 3 was bashed throughout previously faithful Gnome user land, Mate and Cinnamon were designed and processed.

So, where is the power?

Furthermore, a change in the Linux system is noticed and visible. We don't know how MS is screwing us within the black Windows box.

[Randicus]

I feel the potential for freedom is still there with this:
https://github.com/helixarch/debtap
Though I hope to never need to use it.
I am working slowly on a way to hack systemd-shim into a fully functional system.
Inch by inch.

You do realise Arch was the first distribution to adopt systemd?

[bdtc1]

the conclusion that all sides of the debate, have their valid reasons to believe what they are insisting is true. [developers' freedoms] [users' freedoms][corporate conspiracy][development kit availability]

Another major concern: Serious design flaws, bugs, and incomplete code. I.e. shouldn't be in Debian Stable.

[Sarge-in-charge]

the conclusion that all sides of the debate, have their valid reasons to believe what they are insisting is true. [developers' freedoms] [users' freedoms][corporate conspiracy][development kit availability]

Another major concern: Serious design flaws, bugs, and incomplete code. I.e. shouldn't be in Debian Stable.

No only that, but also that by adopting the heavily redhat-based SystemD, Debian looses its identity as a separate Linux offering.

What is the point in using a hacked systemd in Debian, when you can run the real glorious thing on Fedora/RedHat/CentOS?

[mor]

Now we're getting somewhere (and I would like to ask alan for a comment on this particular post by millpond).

I am not questioning the *right* of gnome to do what gnome wants.
What I am questioning is the judgement of the Linux distros in permitting packes from gnome that do not also explicitly give users the right to opt out of a corporate sponsored system that may not be in their best interests.

This is an entirely different matter from arguing that systemd does not respect the Free Software definition which, I want to stress out loud and clear, is what I have challenged you for.

In fact, as you say later:

Certainly not against GPL, but certainly an infringement of *user freedoms*. In spirit if not in letter.

What's a violation "in spirit" is quite a different thing than a violation "in letter".

But we were not arguing about the "spirit" behind the definition of free software and behind the movement, we were talking about its true meaning.

What you are saying now, is about the decision by distros to go along with systemd which, considering that this is a Debian forum, means you should have called on violations of the Contract and Guidelines instead, not on the definition of Free Software.

But even if you call out for Contract and Guidelines' violations, I can only argue that "in spirit" it was a debatable move, but in practice it was perfectly within the boundaries.

Now, if we want to talk about the "spirit" of decisions, that's fine and fair because even many non anti-systemd users are not happy with the way things have gone.
The problems start when what's felt "in spirit" is presented as fact, when a perfectly legitimate developmental decision is presented as a violation of a supposed user's freedom to have what he wants.

There's no such thing, neither in the definition of free software, nor in the contract and guidelines, that guarantees a user's right to have what he wants, where "what he wants" is intended as something that others would have to do against their very and explicit right to do as they like (freedom one).

That is not to say that end-user's input and contribution is meaningless, of course users are taken into account by developers and, like in the examples from alan, of course users can hire developers to hack everything to their taste, but ultimately the decision taken by those who do things (remember that Debian is do-ocracy, not a democracy) have to be accepted even when they are bad ones, otherwise, access, study and modify the code, or hire someone to do it (this is a user's freedom).

End users can in fact contribute any way they want, they can voice their opinion, even protest and complain, can support and finance, but ultimately there is always gonna be a choice to make and some end-users are always gonna be on the bad side of the decision.
Be it the color of a wallpaper or the entire freaking kernel, if development takes a road that brings to a aut … aut, it is unfortunate but there's no calling out for violations of "users" freedom when the decision is gonna remove a choice for some.

Bye

[Danielsan]

We are all happy the GPL has not broken, but insisting on the 4 freedom is hiding behind a finger.
The only lesson I received from this history is the possibility to do lobbying even using the GPL, however, responding to Edbarx, it is really enlightening these few paragraphs from the Devuan website:

Have you tried to opt-out of the systemd change in Debian⁽™⁾ and stay with sysvinit, or whatever other init you prefer? You will quickly notice that is not a matter of choosing packages and in fact Debian offers no choice.

We want freedom of choice, we want Init Freedom!

We are working towards a stable, production ready fork of Debian Jessie, free from the entangling web of dependencies imposed by systemd.

Actually this is topic of the debate: why can't we install another init system by default?
Fine, I can accept because there's no others volunteers to work on a different init, so is a mere fact to have less people to put working for two different init.

However why the majors DE decide to use systemd as deep dependency excluding in advance the opportunity to work with other init is the other second mystery.

The third and last mystery is the stunning and unbelievable, if not miraculous, act of converge of all the distro, all the majors DE in adopting strongly systemd. It is so extraordinary that we can split this in the age before systemd and the age after systemd.

How I am not hiding behind a finger I must sadly admit that it's just an act of lobbying from a big company, actually nothing new below the sun.

[Sarge-in-charge]

How I am not hiding behind a finger I must sadly admit that it's just an act of lobbying from a big company, actually nothing new below the sun.

The root of all this evil lays in D-Bus. Udev needs D-Bus, Systemd ate Udev and also needs D-Bus. Desktop apps want to use D-Bus, and with it comes Systemd in the background.

The distributed binary for Firefox ESR 24 was the last one compiled without linking the D-Bus libraries. The distributed binary for Firefox ESR 31 won't work on a system without the D-Bus libraries. Yes, you can grab the sources and compile Firefox ESR without D-Bus, but pointing that fact would be missing what the general trend is in Linux Desktop Apps.

And now with kdbus the D-Bus subsystem is going to be integrated into the kernel itself. I foresee a future of instability and exploits in Linux land...

And D-Bus, is a scheme from RedHat "to bind them all to their own very will". Make a overly complicated thing, which offers some nice features but which only your full-time paid developers can grok, and the next thing you know is all the bases are belong to you.

[Danielsan]

Everything is possible, for example the market of antivirus is not start yet on Linux server, now with systemd you may have need of an expensive antivirus to keep safe your enterprise data.

[Sarge-in-charge]

Everything is possible, for example the market of antivirus is not start yet on Linux server, now with systemd you may have need of an expensive antivirus to keep safe your enterprise data.

With D-Bus integrated into the kernel thanks to kdbus, I see a Linux antivirus software in Linux servers as a logical conclusion.

Very nice.

And the Linux antivirus software is likely to be a paid-for plugin for Systemd.

Even nicer.

[edbarx]

Everything is possible, for example the market of antivirus is not start yet on Linux server, now with systemd you may have need of an expensive antivirus to keep safe your enterprise data.

That can easily be verified statistically by asking those running systemd, now for months, whether their installations are any more vulnerable than they were before the advent of systemd. I should think, if that were the truth, members of these fora and anyone frequenting Linux fora, would have heard such complaints coming in, but till now, it is only a fear of the unknown, which is understandable.

There was a Devuan discussion about the prospect of kdbus being integrated into the kernel. Such an integration is likely to take the shape of a couple or so of critical functions that make such intergration possible with the least of kernel disruption. Kernel developers don't want to disrupt the well tried and tested code of the kernel, so most of the code should still remain outside the kernel. According to Devuan, this integration is not something to worry about. Moreover, in the discussion, which was conducted between Devuan developers, never mentioned that servers will need to run an antivirus.

[Sarge-in-charge]

Moreover, in the discussion, which was conducted between Devuan developers, never mentioned that servers will need to run an antivirus.

Is the Devuan mailing list like gospel, or what?

[edbarx]

Moreover, in the discussion, which was conducted between Devuan developers, never mentioned that servers will need to run an antivirus.

Is the Devuan mailing list like gospel, or what?

Gospel, I presume.

Joking apart, as I wrote, the incidence of viral infections can be statistically analysed which would provide the probability an infection occurring. Linux fora would have discussed if such infections were taking place. However, you are right to think for yourself, as I am in no position to understand what you have at stake on your machines. Prudence is never excessive.

[millpond]

Linux users have less say in the direction of Linux, than Windows users have in the direction of Windows.

It is Linux developers (be it kernel, or userland) who have ALL the power. They answer to none but themselves.

At least Microsoft with Windows has to care for the users, in order to stop losing market share and thererore, to stop losing revenue.

Interesting. When Gnome 3 was bashed throughout previously faithful Gnome user land, Mate and Cinnamon were designed and processed.

So, where is the power?

Furthermore, a change in the Linux system is noticed and visible. We don't know how MS is screwing us within the black Windows box.

And somebody took cinnamon off my desktop when I wasnt looking.

Now thats software *piracy*.

They better put it back. else its theft.

This is another aspect that needs to have drums banged and teeth gnashing.

Imagine if you went to downoad a nudie app and it took out your desktop.
If I worte that app I would be in Jail.
If Lennart writes that app, he gets a promotion.

[millpond]

Now we're getting somewhere (and I would like to ask alan for a comment on this particular post by millpond).

I am not questioning the *right* of gnome to do what gnome wants.
What I am questioning is the judgement of the Linux distros in permitting packes from gnome that do not also explicitly give users the right to opt out of a corporate sponsored system that may not be in their best interests.

This is an entirely different matter from arguing that systemd does not respect the Free Software definition which, I want to stress out loud and clear, is what I have challenged you for.

In fact, as you say later:

Certainly not against GPL, but certainly an infringement of *user freedoms*. In spirit if not in letter.

What's a violation "in spirit" is quite a different thing than a violation "in letter".

But we were not arguing about the "spirit" behind the definition of free software and behind the movement, we were talking about its true meaning.

What you are saying now, is about the decision by distros to go along with systemd which, considering that this is a Debian forum, means you should have called on violations of the Contract and Guidelines instead, not on the definition of Free Software.

But even if you call out for Contract and Guidelines' violations, I can only argue that "in spirit" it was a debatable move, but in practice it was perfectly within the boundaries.

Now, if we want to talk about the "spirit" of decisions, that's fine and fair because even many non anti-systemd users are not happy with the way things have gone.
The problems start when what's felt "in spirit" is presented as fact, when a perfectly legitimate developmental decision is presented as a violation of a supposed user's freedom to have what he wants.

There's no such thing, neither in the definition of free software, nor in the contract and guidelines, that guarantees a user's right to have what he wants, where "what he wants" is intended as something that others would have to do against their very and explicit right to do as they like (freedom one).

That is not to say that end-user's input and contribution is meaningless, of course users are taken into account by developers and, like in the examples from alan, of course users can hire developers to hack everything to their taste, but ultimately the decision taken by those who do things (remember that Debian is do-ocracy, not a democracy) have to be accepted even when they are bad ones, otherwise, access, study and modify the code, or hire someone to do it (this is a user's freedom).

End users can in fact contribute any way they want, they can voice their opinion, even protest and complain, can support and finance, but ultimately there is always gonna be a choice to make and some end-users are always gonna be on the bad side of the decision.
Be it the color of a wallpaper or the entire freaking kernel, if development takes a road that brings to a aut … aut, it is unfortunate but there's no calling out for violations of "users" freedom when the decision is gonna remove a choice for some.

Bye

What we have here is a very proper talmudic debate, with RMS/Moses as the origin our laws and now for us rebbes to *interpret* them for either social utility, or to the letters of the tablets. The genius of their system was to determine whether the rules served the public interest, and would rule against Moses and even higher if they represented actions that thretened to community. Common sense trumped divine law.

[Sarge-in-charge]

What we have here is a very proper talmudic debate, with RMS/Moses as the origin our laws and now for us rebbes to *interpret* them for either social utility, or to the letters of the tablets. The genius of their system was to determine whether the rules served the public interest, and would rule against Moses and even higher if they represented actions that thretened to community. Common sense trumped divine law.

Common sense tells us to not have a hot-plug based PID1, I don't want the state of my servers to dynamically change because of hot-plug generated, or d-bus received, events.

I want my servers to either stay at the runlevel/state I have defined, or crash. I don't want them to be dancing to unknown and undetermined tunes, you know.

Systemd can be kosher according to Free Software Guidelines, and still be utter rubbish.

[edbarx]

I want my servers to either stay at the runlevel/state I have defined, or crash. I don't want them to be dancing to unknown and undetermined tunes, you know.

-1 for systemd

[millpond]

What we have here is a very proper talmudic debate, with RMS/Moses as the origin our laws and now for us rebbes to *interpret* them for either social utility, or to the letters of the tablets. The genius of their system was to determine whether the rules served the public interest, and would rule against Moses and even higher if they represented actions that thretened to community. Common sense trumped divine law.

Common sense tells us to not have a hot-plug based PID1, I don't want the state of my servers to dynamically change because of hot-plug generated, or d-bus received, events.

I want my servers to either stay at the runlevel/state I have defined, or crash. I don't want them to be dancing to unknown and undetermined tunes, you know.

Systemd can be kosher according to Free Software Guidelines, and still be utter rubbish.

This is the crazy part about systemd.
Its primary risk is to servers, and most seem to be Redhat based (CentOS apparently a favorite).
And in fact there seemed to have been a mass of complaints from sysadmins.

Lennart thinks because his baby boots faster it has to be *better*.
And Redhat was (in)famous its ancient but unbreakable server software.

Its similar to someone putting screen doors on submarines to improve ventilation efficiency.

As a lowly user, my objection to it was that it broke some scripts. Something I consider a mortal sin, enough for Billy Gates to be standing on LPs shoulders in the 9th circle.

But I can just imagine someone running a server farm. Fully optimized. And then some propellorhead comes along and decides to *update*.
Also, the real last straw here was binary logs. I dont think Lennart created the idea, but it is antithetical to everything that I define as Linux. The beauty of Linux is that I dont need to be a programmer to use it, and use it effectively. A few bash or shell scripts should be everything that I should ever need to customize the system. And while I am sure someone at CPAN will come along and make some perl modules to deal with the new crap, the line has been crossed.

There is however, one point that needs correction about Systemd. When we say its monolithic, what is meant is in scope and direction. Its cut from one wafer (monosilicaceous???) - there are of course many components to it, and ever growing. That is why many of us regard it as not a symbiont, but as a cancer.

[golinux]

millpond . . . I'm actually starting to like you. Never thought that would ahppen . . .

[somebodyelse]

Lennart thinks because his baby boots faster it has to be *better*.

No. In fact, in spite of all the insinuations of cloak and dagger, the arrogant Mr Poettering has at least been transparent in the systemd project's goals:
http://0pointer.net/blog/revisiting-how ... stems.html

He is not obliged to make software to make you happy. All these unwritten laws about how when person x sits down to write a piece of software they have to carefully balance all the thoughts and feelings and tastes of any non-paying user ever. It's like this indignation about Mir. Really, if you've ever got wound up about Ubuntu/Canonical developing software for itself that suits its own purpose, please kick yourself. And not just a soft playground kick. A really good hard kick to make you think twice about the morality you unreasonably and arrogantly impose upon other people who owe you nothing.

As for the distinction between user and developer: the kind of "user" who is non-technical and therefore cannot enjoy half of RMS' freedoms, has no problem with systemd. Systemd may make a classical sys admin's life more difficult, whether on account of change resistance or any technical defects it may or may not have. However, for a user, who uses his/her computer simply to access software and do non-programming/non-system-administration things, systemd is not a problem. It is in conjecture/insinuation land but we don't care about that place.

I am currently running Debian Jessie and it's wonderful. It's not wonderful because of systemd. But having systemd hasn't changed the quality of Debian. Don't allow your feelings about systemd to colour your judgment of Debian.

Systemd may be bad for server administration but it is odd then that it has been adopted by Red Hat and SUSE, whose business is (largely) linked to servers.
Systemd may be a way for creating back doors for government agencies. But to date, there is no evidence of ANY kind. Only insinuations in article titles. "It's so big, it must be hiding something. Nudge nudge."

As for MS and their Window thingy, we didn't care about svchost. No, really. Precisely zero shits were given about svchost. That was never the problem. We cared about their alleged business practises and their alleged treatment of the user, none of which is replicated by systemd. And no, it doesn't make anti-virus inevitable likely.

Note that Debian 8 does not require me or lead me to sign in to a remote account. Nor does it come with crapware (except for systemd, yuk yuk). Quality software is available from the repositories. These things matter.

Does systemd suit your needs? No? Well, provide code to Debian to make it easier to work without systemd or help Devuan.

But stop (wittingly or unwittingly) working for Microsoft with all these "you may as well use Windows" statements/arguments. And stop moving the goalposts on software freedom. In all the ways that matter and in all the ways they have ever claimed or committed to, Debian remains free (as in freedom) software. Again, no one is obliged to develop his/her software in a direction that pleases you. And they don't become better people for doing that.