Respectfully, Jakedp, you apparently have not read possibly even half my posts since I stated I do not like it, I don't want to see it in Slackware, and specifically because it does nothing I can perceive is useful to me and at considerable cost. So please don't attempt to paint me into some BiPolar corner. I do not have a habit of demeaning people's experiences since exactly why I posted a few times asking politely that people stop with the same ol', same 'ol that we have all been through for almost a decade now and I stated the guidelines for this thread when I created it which specifically asked members NOT go through the tired old rants with no experience to mention AND the post you referred to had no experiences, nothing new to add, just one more attempts to drag this thread into flames.

Now for the analogy which I find severely lacking. Heroin provides something which users feel has been missing and in the process also triggers electro-chemical changes in the brain that cause severe discomfort up to and including death in some. Systemd can hardly be considered euphoric or providing some need people felt was missing. It also has no "hooks" in people despite that it does have extensive hooks in the PC operating system. My experiences with systemd has been that I barely even noticed it was there, significantly less effect than if I used some heroin. I'm pretty sure I would notice a change had I used some heroin. Systemd has no effects that cause great discomfort when one stops using it. In fact I mentioned trying to remove it from Debian and being upset that it didn't work. I don't see any meaningful parallel at all. If you can refute this experience please do have at it but kindly be a gentleman about it and do your research first, instead of just attempting to plug square plugs in round holes, because it suits your POV.

The equivalent of heroin is Facebook!

Agreed! I am social media free.

If systemd is like heroin then heroin needs to up its game.

I think the assessment has been pretty much the same from the beginning. Everyone expresses it in a different way. It all comes down to what one does value more.

Purely pragmatic there is no issue with systemd at all, as long as it does the job, and doesn't bother users with all kinds of issues. The same goes for stuff like snap, from ubuntu. Snap offers a nice feature, since it enables users of a certain system to run software that uses different libraries or frameworks. In return, it pollutes your mount points, where even a simple calculator application can have it's own mount point. Is that a problem? not at all. The same goes for installing your libraries and software to wherever you want on your system, as long as you make the right links. But it's probably for a reason to have some structure, so one knows what is what. Snap also shifts the single point of trust in a distro's repo, to review every package/vendor separately. Because, after the initial review process, publishers can push updates themselves to the users. Also not a problem if one doesn't care about publisher/software, but only about as long as it does the job.

Certain principles or structures have been created over the years for a reason. That's why they have become principles. Throwing principles and/or structures out of the window because one wants to create something new, will always cause resistance. Probably for a reason. People always have to relearn lessons after a few generations, because old knowledge became forgotten.

SystemD brings stuff like starting services in parallel. For the user or sysadmin, that only uses or administers the system, it doesn't matter that it is a complex system of many tools, as long as it brings the advantages and is easy to use. The fact that they put several features into one system, and more and more relies on it, means that alternatives for every feature that systemD provides, will have it harder to gain ground. For those who want to understand every bit of their distro, or the maintainer, it adds to complexity to.

SystemD takes away from the KISS design principle. And it also makes it harder to replace the individual functionalities for maybe better ones. People that value that, will repeat those issues, people who don't, will wave it away.

If you just want to run a system, and don't care about the underlying system, systemD is no issue at all. You can run windows, or ubuntu with systemD, or maybe slackware with systemD, and it will probably serve you well.

If you are more of a tinkerer, or just want to know how a system ticks, it's probably better to stick to stuff that respects the KISS design principle, because it is more doable to grasp. The transparency that we have with a system like Slackware is because of the old values of structure and design principles. It's good to value that, without becoming stale.

Luckily there is choice, and there always will be.

I see you have never talked to a heroin user and copied and pasted a generic textbook response. You know nothing of heroin as it does not provide something people feel missing. As most drugs. It is a retarded proposition from people that have never dealt with it in the field or people who have not used it.


I never said you were for sysemD in Slackware. You do have a weird way of campaigning for it through deflecting those who do not state a carefully graded school approved expression of why they do not like it.


If you do not notice a change without systemD then you are using some hardware more powerful than the average user. A systemD distro does give a different user experience than a non-systemD distro.

The parallel start argument is old. First, there is many arguments on why that is a bad idea. Neutral in this but prefer a linear start so do not have spend time thinking on what some cryptic error means or what the init decided to run what in what order. Second, there is inits that can run in parallel. OpenRC for sure I know can.

There is not a single thing systemD does that another init out there somewhere cannot do. Many could do what it does, pragmatically, but choose not too because they have a focus and are not octopuses.

Jakedp I really don't want this to drag on but I also want to at least give you the respect of an honest response. You have zero knowledge of my life and my experience yet you jump to the above conclusion. I was married to a heroin addict for 20 years. I'm not going to say any more than that as it doesn't matter beyond revealing that I have intimate experience with addicts so your conclusion is dead wrong.

No but you did say before that I shot down any negatives (which I obviously didn't by example) and this time see my willingness to discuss both Pros and Cons as somehow "campaigning for systems" a completely mistaken conception. I did not ask for nor want just one sided posts. I really want to see both as long as they are experiences and not just guesswork, rants, or jumping to entirely hypothetical conclusions on philosophy. Now that we have some 10 years of usage it seems to me interesting to know what people either love or hate about it's actual usage. Just to be absolutely clear, I am in no way campaigning for it. It holds nothing good for me.

It's probably ridiculous to think RedHat would abandon it since they invented and pushed it. Debian seems to sort of ride the fence excepting their solution for SysV as an option doesn't even work so why is it that they have yet to make that viable unless they are reasonably happy with it's performance?, assuming it isn't a big deal or that whatever benefits they perceive outweigh the need to make an alternative actually work. What about Arch and SuSe? Why are they still employing it? Surely they don't harbor some death wish which if systemd caused more problems than it solved would surely be the results. We can't possibly answer these questions if we stay stuck in 10 year old ranting, and I'd like to know. I am beginning to get a clearer picture that it has almost no effect on SOHO Desktop systems but has some important benefits for Enterprise systems.

What? I truly don't understand this assertion. What has hardware got to do with it? Have you actually booted a systemd based distro more than a time or two? I have tried SuSe, Arch, Debian and CentOS, both as SysV and post systemd adoption and on my SOHO Desktop boxes I really don't feel a difference with or without it. That is exactly why I'm curious about what is the big deal. I don't want to push it. I don't want to kill it. I just want to understand it better as it has actually played out.

Lately I have read/heard about some perceived value in stopping and starting daemons, and since I do that to some degree in Slackware, what can systemd do that non-systemd systems cannot? I'd like to know more about that... again, not hypothetically but in actual use. Is it just that it can automate that process? I don't use it either enough or in the target environment (Enterprise) to grasp the value of that. That's why I am intellectually curious about it. I have no intention of ever adopting it. Understood?

Well, it's not so much as stopping and starting but more along the lines of starting daemons in parallel as much as possible and restarting those that died for God Knows Why reasons as soon as possible.

Starting stuff in parallel means that you can do a rolling restart of your servers as fast as possible, keeping your downtime as small as possible.

Restarting stuff that has died for some ungodly reason means that (maybe) your customers don't notice that you didn't provide service for a short period of time.

You can do those thing without systemd. But, RedHat and Ubuntu both use systemd and they both have cornered the Enterprise market.

I've noticed that people who strongly support systemd often run server farms, whereas people who hate it are invariably home users. To go off-topic for a moment, I've noticed almost the same split for and against the new network device names.

Don't blame IBM on that. systemd was funded by Red Hat long before IBM bought it :D

PS That may be a coincidence but yes probably Red Hat has very few home users among its customers.

Because they do not want/need to pay for support.
There will be more CentOS home users though, for the same stable O/S without subscriptions.

If you follow the money - freedesktop.org hosts, or used to host, most of the projects which are inextricably tied to systemd or have been superseded by it, such as systemd itself, D-Bus, polkit, upower, udisks2, X.org, mesa/drm, HAL, consolekit, etc...

Everything freedesktop.org hosted is related to Linux desktops and in particular - gnome. It has very little to do with "UNIX"... Red Hat are the biggest contributors in terms of funding and developers to both gnome and many of the freedesktop.org projects. For example - HAL, D-Bus, polkit were all developed by Red Hat people.

The reality is that, when it comes to Linux desktops, Red Hat have been developing and funding a lot of that for well over a decade and they are not interested or invested in developing a "UNIX".

When you have a large corporate entity, which develops many of the components, they are in that position where they are free to unify those fragments and develop the end result as they see fit - because they pay for it and employee the people who do the work.

Interesting that the above are mostly software that I try to avoid. You have to have mesa of course. But I do without dbus, polkit, consolekit, etc. because I can't see any need for them.

I find myself to still be in a neutral position (neither I favour it nor dislike it, I can live with it though) regarding systemd, however, that comment regarding Red Hat's commitment to those components is true and that systemd itself has been used a lot on server farms. It is probably a component found in distros that are ran on supercomputers too iirc.