I think the definition of a republic is a government where the citizens are represented by their delegates in the capital.

the roman empire was a republic.

What a well spoken mouthful! ++dunric

How so? Again, like service supervision, "American" is a term that you think everyone just knows and agrees on. In reality, it carries a freight train load of connotations and assumptions, for example, how much american TV I might watch, all different for different people. I simply try to avoid those assumptions by not answering yes or no to the question.

I fear that even affirming my current geographical location will cause you to attach many of those connotations to me.

I might better describe myself as a Free Range Human Being. If you must have some state-based reference point, I am the free and independent state of "Robert".

See what I said about connotations and assumptions!

But that is the clincher - the evil of a democracy is not an implementation detail - it is the core principle!

If you think that one person or group may be justly dispossed of life and/or liberty by the headcount among another group - then you need to very seriously re-examine your foundation principles.

Once you allow a headcount to determine the course of actions affecting others, you effectively define right and wrong, good and evil, truth and lie as subject to a vote! And usually a vote heavily manipulated or influenced by those with a keen interest in the outcome of that headcount! Utterly insane if you will just think about it!

If you think that is actually a good idea then you share the psychosis I mentioned and should seek professional help... (all in serious good humor, but please consider that point!)

It is often said that no man is an island, but that is a mantra repeated within the mob to protect against those living outside (aka, the evil ones).

I can tell you that the best state of an intelligent mind is in solitude and peace, waist deep in a gentle sea, beyond the noisome pestilence of the onshore mob. (Alternatively, the solitude of a few hours spent at speed on two wheels comes close... choose your metaphor).

As for the feral republic under discussion... an intelligent study of history is very revealing, but would take us further into the moderators waters... But to show this is not something new, I highly recommend that you find the old Henry Fonda western, The Ox-Bow Incident, and you will see that it is actually lost knowledge - and enjoy a great movie!

To demonstrate how this post is actually on topic, I will point out that the turmoil created by systemd is very democratic in nature and progression. Its origin is within a powerful self-interested group that has pushed it into the Linux sphere without regard for those peacefully using existing inits (or possibly for the purpose of bringing them into their own influnce). It has been very much advanced by effectively growing the size of the headcount according to democratic principles, but still without regard for those with pre-existing uses who choose to not participate. The mantra directed at the evil non-participants is "Doing nothing is not an option! You have to join-up!".

I choose to remain at peace with SysV init. If changes are introduced which make that difficult or impossible in the long term, then that will certainly affect me. But I will still recognize it for what it is - an unwelcome intrusion on my peace that does not respect my pre-existing independent use of my own computing machines.

3 members found this post helpful.

Exactly, I've seen little if any real benefit to using another init and service supervision system, including systemd. In fact I have more headaches with systemd than any other init and service supervision system out there. Don't get me wrong Runit is nice, and yes I worked on scripting for it, but it's not perfect.

My example of why I hate systemd on the test server I have running it... For me, networkd is a constant pain in the ass to deal with, and if you accidentally disable it, it kills your entire network and regardless if you enable dhclient, dhcpcd, networkmanager, etc. the network will never come back up, and you can't really use dhcpcd, dhclient or networkmanager with networkd. I had to re-install systemd just to get networkd back.

So ready to get rid of the systemd system once and for all. I built it to learn how to try and use this thing and I'll be damned if Windows Vista wasn't as big a pain in the ass. Sysvinit and bsdinit are just so much easier to work with.

4 members found this post helpful.

Thanks for the considered reply in post #61, skarnet.

I started a reply twice but decided to wait until I could offer a worthy response. Unfortunately that moment has not come and I spent my LQ time today on other responses, mostly in this thread. The weekend is upon us so please do not think I have ignored your thoughtful post, and I will offer a somewhat truncated and hurried reply as follows...

Glad to hear that. And I should add that I did not mean to imply any kind of technical excellence in systemd - it is a wasteland of bad ideas... IMO.

I cannot so easily follow the argument that "It was evil, but now it has some good points!".

The primary evil of systemd from my perspective is just that - it does things, many things, that should not be part of an init function. There may be some features there useful to others, although I see none of interest to myself and thoroughly resent the way that they have been forced into my field of view. But if others need them and want them, it is still evil to include them in an init that changes how everyone else must use their systems! That fundamental principle of Unix, doing one thing and doing it well, cannot be repeated often enough!

To implement them as part of some other init because systemd did so is still as wrong-headed! Implement them if you must, but as completely optional, completely separate projects and packages. Please do not further the systemd agenda by duplicating it, no matter how well you may do so.

Please do not participate in Lennart's "sins"!

The distros may be your target, but without end users there will be no distros! Lennart is after end users too, distros are the path to increasing his own user count simply because he knows that end users would NOT have flocked to systemd in droves otherwise. So by forcing it into the distros he forces the end users to sign on, like it or not.

Don't compete by duplicating his effort and methods, compete by offering a DIFFERENT CHOICE! In particular, offer the choice to keep what we have! And in the process be careful to respect the current and future uses and peace of those who have no compelling need to make a change.

In my own case, and apparently a good many others, we do indeed already have everything we need - in SysV init! Systemd is doing damage to the Unix/Linux ecosystem - please do not facilitate it by replicating its API or ABI!

If you have a new feature or set of features, then by all means develop them and make them available to us. But please, please, please allow us to select based on merit and do not jump on the train that is running over us and shout how your boxcar is preferable to theirs! When the wheels of the train slice off our existing appendages, we don't really care which boxcar it was that got us. Get off the train and help derail it before it plows through the innocent crowd of end users...

Your perspective is obviously that of someone with a "product" targeting this same "market". Your "we" and the "we" of many end users is fundamentally not the same. Please respect that.

Doing what you call "nothing" is exactly what I and a good many others intend to do. This particular "nothing" - maintaining continuity of our current uses, in particular with continued use of SysV init if possible - is apparently going to require a lot of effort and dedication! It has already disrupted my own computing uses and resulted in me spending quite a bit of effort planning and implementing methods to assure that "we" remain systemd and systemd-look-alike free. "Nothing" takes a lot of work that I would rather have put to other projects!

Every implementation that adopts or mimics systemd features and methods makes it just that much more difficult for us. But as I have said before, defeat is not inevitable.

So if you are really listening, I and others are asking you to please make your own implementation a non-monolithic (as opposed to merely less monolithic) set of independently usable features. In particuar please respect the current and future uses of all us "do-nothings", and facilitate our independence where possible, interfere with it never.

6 members found this post helpful.

Astrogeek sometimes the biggest and boldest things come from doing nothing. Doing nothing on an issue can mean the same as standing firm on an issue equally. I think doing nothing at this point is the best course of action, and doing nothing is best for Slackware. Should Patrick choose nothing and keep the simplified bsd-stylized sysvinit method, he's making a bold statement.

1 members found this post helpful.

I apologize for not making this clearer in the first place: this is exactly my intention.
I agree with everything you wrote in this post, and I will ask you to give me a little more credit, because we are definitely in the same camp and have the same objectives. I am not on the "everything in the init system" train, and will never be.
Do not believe me at face value: look at my software and how it is designed. I will never, ever write a monolithic thing that does not abide by the Unix principles. I am probably more of an extremist than you are on that point - the goddamn shell was too monolithic for me.

My fear, my big fear, is that by "actively doing nothing", i.e. fighting for the status quo, you are ultimately committed in a losing battle. systemd has huge manpower behind it, and it is evolving. Day after day, it's getting more shinies, and more and more distributions are succumbing to its sirens. I'm extremely afraid of kdbus - as long as poetterware is userland-only, I can always do without it, but once it has started creeping into the kernel, Linux as we know it is over; and it doesn't sound like Linus is going to raise a strong opposition to it.

I feel, very acutely, that in order to have the slightest hope of stopping the systemd epidemic, more resources need to be dedicated to the fight. I don't want to downplay your efforts, not at all, but what I am seeing is that systemd is still gaining ground today, and I think that means that more needs to be done.

Education and communication are the resources we lack the most: the systemd clique is excellent at communicating - as I like to say, if only Lennart was half as good an engineer as he is a communicator, the world would be a much better place - and the propaganda works beautifully. We need a team dedicated to dispelling myths about systemd (the biggest myth being that an integrated init system is needed at all); maybe boycottsystemd.org could fill that role - I haven't contacted them yet, but in time, I will. That team would need to make noise. A damn whole lot of educated, technically sound noise.

But that will not be enough. When you talk about systemd, systemd, systemd, even if you bash it (even with the best reasons in the world), that's still more publicity for it. And at the end of the day, people ask "okay, and what do you suggest instead ?" and you answer "do what we've always done before". That's not sexy. Nobody's going to follow you on that, at least not distributions; systemd does have a few interesting points (all of which predate systemd, but were not advertised outside of very small technical communities), and main distributions have shown that they prefer having the features with the horrible implementation to not having the features at all. And if distributions do not follow you, Joe User will still have systemd on his machine and his life is still going to suck, and when your generation of knowledgeable users moves on to something else, technical excellence will be lost.

That is why I want to come up with an alternative. By alternative, I don't mean a monolithic monster that would be conceptually just as bad even if technically better; I mean individual, separate, optional packages that can be used to achieve an equivalent level of functionality. It is a lot easier to convince people that A is bad if you can show them a B that can do more or less the same thing but respects the principles you claim that A is breaking. (And the best thing is, you don't even have to use B if you don't want to.)
Mainstream distributions will not go back to sysvinit, I'm afraid that ship has sailed; and it will become increasingly harder to defend a systemd-free bastion if they keep producing systemd-tainted software and enforcing systemd "standards". But an alternative with similar features, and that happens to be less proprietary and more open, that is something they can accept.

I'm willing to do whatever it takes to stop the systemd virus; and I believe that what it takes includes communication, and software to back up the communication. I'm not that good at communication, but I'm good at system software, and I know a thing or two about init - so I'm going to use that power for Good. And you can help, by being convinced that I'm not just a guy with an agenda to ride the init system hype train to advertise his product

4 members found this post helpful.

Found this interesting link

4 members found this post helpful.

^ Sums it up pretty well.

1 members found this post helpful.

You know, as the article points out Moisespedro, software having "merit" should take precedence.

1. I've never like work-in-progress software and progressive-ware. Once a project goes out-of-bounds, it's not worth it.

2. I hated PulseAudio and I hate networkd. I hate hijackware that kills a system service if you disable it trying to use an alternative. So yes, I do have an opinion of Lennart-ware. It's garbage. If you like, fine, enjoy your treasure as it was my trash.

3. Why do I need GNOME? I don't even like GNOME. I don't like KDE that much either. I like the packages like KATE and KDESVN, and other accessory-ware, but I'm an Xfce-addict. I use Xfce. So why is logind a concern? Xfce uses ConsoleKit just fine and the Xfce devs haven't made any fuse over removing ConsoleKit or hald support. If that becomes a problem, I have good feelings towards FluxBox.

4. As far as simplicity goes Shell scripts are the easiest executable files to read, as are plaintext logs. I mean seriously learn the King's English or don't bother. Nobody wants to read binary or C if they can help it. We're not a robot named Bender or 160 year old Professor Farnsworth. We're a bunch of John Q. Everymans.

5. Fadware makes operating systems look cheep, cheesy, and unproductive, and GNU/Linux doesn't need someone else trying to be Steve Jobs. Nobody... Liked... Steve... Jobs! Except those who liked fadware... Namely the Apple crowd. Yes, I use an iPad but at times, I'm not a fan. Bandwagoneering only makes distributions look even worse. Look at how many people left Debian over systemd? Need I say more?

1 members found this post helpful.

Reaper, I didn't say I like it (systemd). I was just posting a link.

Oh I know man. However, I'm using the context of the article for my own experience with systemd.

When I get back from my vacation the system I'm using that has systemd is going out and one with traditional sysvinit is going back in. I tried to learn it, and I learned enough after about 8 months of headaches. I learned I don't want another Windows Vista.

Please get this thread back on topic now.

1 members found this post helpful.

I think we have our answer really. The majority have chosen "Other" with a clear intention on sticking with the current simplified bsd style sysvinit solution. It's the best solution for Slackware really. We need to uphold the educational benefits and aspect of learning GNU/Linux the right way and being a true system administrator without bells and whistles. We don't need service supervision if we learn how to write shell scripts correctly and learn how to use tools to manage systems by hand. We don't need eveything automated beyond udev's automation. We need sensible software, not fadware and progressivism. Sysvinit works, works very well, and doesn't need to be replaced. If anything else is needed to suppliment it, we have Runit in the SBo repository and dozens of scripts to work from.

4 members found this post helpful.

I voted for runit. It is true that there are 2 main features that these replacements (not all of these fully replace init) offer, e.g. parallel startup and process supervision. runit seems to be the most customizable, much like init it supports scripts. Finit seems good too, but it seems a bit less customizable.

^^^ +1.

I cannot in all good conscience abide systemd, its creators, or the methods by which it was brought about. No point in reiterating the same reasons. I've been in the field since the '80s, not a coder of any kind, but been around long enough to know a thing or two. I have fierce historical/philosophical objections. Sad, but I suppose this kind of thing is inevitable sooner or later wherever us humanoids are involved... Suffice it to say that my feelings on the subject cannot be expressed in polite company.

What's frightening is not the BS itself, rather it's the sheer number of us humanoids who buy into the spiels of these court-jesters.

1 members found this post helpful.