I don't want my init system to replace cron. I don't want it to manage logging. I don't want it to have any debugging capabilities. All of these things can be done with other programs, arguably in a much more flexible and robust way.
An init system should do one thing (well): manage system services. Within that context, it should start services on boot, keep them running in the background, and allow the user to create, stop, and start services. That's it. And it could be argued that even those responsibilities are too broad for a single program.
So I understand that you and many others like that systemd provides solutions to all of these tasks in an integrated way. But you should also understand that this does in fact go against the Unix philosophy, of small, independent, but composable programs with a single responsibility. It "proposes" alternatives to many other tools for no particular reason, until users are effectively using GNU/systemd/Linux.
And, yes, I know that technically systemd is not a monolith and is composed of many programs, but that is a moot point. It is a single project, maintained by a strongly opinionated team, and given the importance of it, most distros go all-in, so users are strongly recommended, if not forced, to use all of its programs. In many cases it's not even possible to use one individual systemd program idependently from another. This is why systemd is seen as a kraken that takes over the entire system if you do decide to use one part of it.
The Unix philosophy is not an end goal neither. It's not even something well defined. Everyone seems to have their own view on what it is. I personally take the "everything is a file" and "do one thing well and be composable" rules as a guideline, an ideal to consider when designing stuff, but not as a strict thing to adhere. It might be something that's nice to have in some contexts and something that's useless or even counter productive in others.
What I mean is that I take "does not follow the Unix philosophy" as something to look into to find potential improvements or design issues, but not as blocker or a counterpoint in itself.
To me, the Unix philosophy discussion is quite moot. Those discussions are often very vague. I don't care much that systemd follows the Unix philosophy or not. I'm more interested in what actual problems this does cause in practice.
You do, however, point out something practical here:
> An init system should do one thing (well): manage system services
I suppose one could consider that to manage system services well, you have to manage "everything". I also suppose systemd's scope is way bigger than "managing services", they do want to "fix/figure it all". It seems reasonable to me not to agree with any of these things.
I do believe the uniformisation systemd causes is a good thing, but I absolutely understand that the big scope can be seen as an issue, and I almost fully agree with your last paragraph. I would object to the statement that systemd is not a monolith and is composed of many programs is a moot point: this modularity still means that you can replace individual systemd programs with your own implementations if needs be…
… as long as you provide the expected features / APIs, yes, you are totally forced into this indeed. systemd is a de facto API. It brings / forces standardization at the cost of diversity. It broadens the standardization that comes with UNIX/POSIX and XDG. I'm sure this can be criticized in a few ways: the API design, the scope, the featureset, the way the project is lead…
The alternative to systemd is non-existent standardization and each alternative designs stuff its own way on their side. For the better and for the worse. I can see how systemd can be criticized for when we are in "the better" cases. I personally easily see the worse side where several projects (for instance desktop environments) would each have to implement features that come with systemd. And programs on top of these environments now have to implement APIs of each desktop environment to be well integrated.
This is more work for everyone.
I guess this is a diversity vs efficiency balance to strike and we don't all see it at the same place.
I suppose another alternative would be to have different people working on different implementations that are then grouped in some common "system core" package or set of standards that everyone adopts. I'd probably be happy with that, if this is at all possible.
You're right. But what I take issue with is that systemd authors deliberately decide to go against it. We know because there are other init systems that do follow these design principles much more closely.
Of course, an init system is not trivial, and is a special program that must be given additional permissions over most user space programs. But the problem with systemd is that it's not just an init system. It is a collection of tools that also manages logging, networking, DNS resolution, containers, and a bunch of other system tasks, which, in my opinion, it has no business managing. When you add to this the fact that these programs are all interdependent in some way, and that I can't use e.g. `journald` without systemd itself, it really drives the point that this is an attempt to establish a cohesive and centralized system, rather than rely on a collection of independent but composable tools, most of which already exist. So I get the appeal why some people would prefer this, particularly if they're not already experienced with existing tools, but it's also not a surprise why experienced Linux geeks would scoff at this.
In my experience, systemd doesn't give me anything that I can't do well with other tools. And instead of having the choice to use a tool of my preference for each individual task, I'm forced to use a gargantuan system designed by a single group of people. Whether or not this ultimately makes my life easier, it goes against the primary reason why I choose to use Linux in the first place. If I wanted someone else to make decisions about how I use my computer, I have Windows and macOS for that.
Tangentially, this is also why I have a love-hate relationship with NixOS. As much as I appreciate reproducibility, atomic upgrades and rollbacks, and having a fully declarative system, its authors insist on managing every part of my system with Nix, which is completely insane to me. So, for example, it tries to replace every single package manager in existence, whereas I much prefer using something like `mise` to manage my development environments. Technically, I still can and do that, but it's certainly not the "Nix way".
Interoperability and composability are the core tenets of the Unix philosophy IMO. It's this that allows me to use programs written decades ago together with programs written today, without either tool being aware of each other. In contrast, tools that try to take over my machine forcing their own UIs on me—no matter how supposedly superior they might be—are hostile to my overall computing experience.
Which one(s) would you recommend/suggest?
I felt like systemd was an epiphany. Software doesn't need to a collection of simple tools that do one thing really well. You can have one tool that does everything shittily, the pdf reader of init if you will, and that's systemd. The author went on to do brilliant work with pulseaudio, you know, the whole /dev/dsp everything including sound is a file, oof. Let's make it a weird complex server process, oh, and let's make another sound system after that too.
I was very happy to see Lennart Poelering had joined microsoft to bring his brilliance to windows. I'm sure he's just cranking out masterpiece after masterpiece of design for them. I actually switched from unix to windows after being so inspired the tremendous quality and sensical design of both pulseaudio and systemd. Oh, and both very reliable, simple, and intuitive.
One does not need cron or systemd for scheduling tasks.
Have you written your own Cron in Ruby?
- You claim that the Unix philosophy only survives to the GNUtils. Well, that shows to me a lack of understanding what the philosophy is about. Everything is a file is similar to the OOP approach of everything is an object. I recommend watching Ken Thompson when he was young here: https://www.youtube.com/watch?v=tc4ROCJYbm0
It does not capture all the UNIX philosophy but it does extend on the reasons why that philosophy works well. The philosophy is bigger than that of course but it helps serve as a counter-argument.
- The example of "writing your own script" is no different to a non-systemd system. Why would a script work or not work based on systemd? You mention as example FreeBSD debugging a shell script. Well, others use proper languages such as ruby or python. Everything that can be done via systemd I can do without it too and, in fact, have been doing so. Ruby essentially runs my system as the primary layer on everything (granted, it runs Linux, and thus mostly C, and ruby is at the end of the day a syntactic wrapper over C). I never understood why systemd would matter. I read the advertisement of the systemd devs - none of this applied to my use cases, so I never "embraced" systemd, simply because I never needed it. I did point out the increased complexity of it as a negative trade-off and this has been true til this day.
- Former "hater" also implies that criticism is not based on rationale and logic. This is not the case either. It's funny to me how the pro-systemd camp isn't really able to come up with compelling arguments on their own.
I would say your comment either shows a lack of understanding, or that you completely missed the point.
> The example of "writing your own script" is no different to a non-systemd system. Why would a script work or not work based on systemd?
Of course, you can write the service itself in Python or Ruby or whatever regardless the service manager. The point is that with systemd, or upstart, or other service managers like this that make things more declarative (launchd?), you don't have to write a script to manage the service at all.
On systemd, you declare which services yours depend on, how to run it, which user should be used to run your service, and many things are handled for you, including many security mechanisms you don't need to think about and provide further config for this stuff that would be a mess to handle with the traditional way of writing a custom rc script per service.
The problem is not being able to write in languages like ruby or python. It's to have to write something at all.
systemd makes many things declarative that were historically procedural, potentially painful to debug, code.
This eases distro maintenance and I suppose is one of the top reasons most distros adopted it.
wrt the Unix philosophy, discussions about it related to systemd are often (always?) too abstract to be useful, I'd suggest talking about specific problematic points.
No.
cat something|filter step 1|filter step 2|filter step 3
filter step 1 something|filter step 2|filter step 3
filter step 1 < something
filter step 1 -i something
filter something step 1
cat something|filter step 1
< something filter step 1|filter step 2|filter step 3
I think it's a good example of when it's worth straying from the philosophy.
Yes, it 'prints' the file to stdout which is consumed by the pipe and turned into the input for the next command in the pipeline. It doesn't matter whether you're only 'printing' a single file or a bunch of them.
As for GNU utils and the examples you mention, those indeed align with the Unix philosophy, which you clearly misunderstand.
Convenience? `tar` integrates well with compression tools, but doesn't implement compression itself. This is the epitome of the Unix philosophy. You can just as well pipe its output to any compression tool of your choice, if you prefer not using its CLI.
> why find has a DSL?
Describing an advanced CLI as a DSL is a stretch. But to humour you: flexibility, and because files have many attributes which a good finding tool should expose to the user. Whether you like its CLI or not is a separate topic, but you're mistaking minimalism for simplicity as a requirement of the Unix philosophy. Some tasks are inherently complex, and forcing a tool to be "minimal" at the expense of flexibility would be counterproductive.
Besides, you're free to choose any other tool you like more to find files on your system. The fact GNU `find` is easily replaceable is precisely a sign that it follows the Unix philosophy well. I personally use `fd` and ripgrep more often than `find` these days.
Re: `cat -v`, I hardly know the history behind it, but it doesn't really matter. As a sibling comment mentions, there are no hard rules around this topic, and people will disagree about what it really means, and how a program should be designed. If I had an opinion on the topic of `cat -v`, I would probably argue with Rob Pike about it as well. None of this means that these design principles are not worth upholding, or that we won't make mistakes along the way. But going back on topic, it's a problem when a project like systemd explicitly chooses not to follow these principles.
I remember embracing systemd funny enough due to when I was running alpine as my server and I had to write my own r script and boy did that quickly make me remember the awful times on freebsd debugging sh.