[SOLVED] Is 2019 Still Too Soon For Intelligent Assessment of SystemD?
SlackwareThis Forum is for the discussion of Slackware Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Well, permit me to offer some details about the "arrogance" of those project leaders, from what I know.
The XFree86 development model was a monolithic source tree, where everyone to contribute. The source code to be maintained only on that tree. If you want a better insight about their development politics, it was just similar with the one applied by Mr. Torvalds even today with the Linux kernel.
But, someday important companies like Intel and AMD wanted to develop their drivers in separate packages, with separate release cycles. There appeared the conflict, and in the end the moneys spoken, and so we get the X.org.
Why is important for us (and this thread) the story of XFree86 vs. X.org ? If initially the X.org was just a modular alternative to XFree86, after years nobody used the second one, and the drivers started to become antique. The XFree86 itself started to become obsolete.
I think that systemd is the "new X.org" and the refusal to use it will end a Linux distribution in the same position as trying to use today exclusively XFree86. Maybe it will work, but with really huge limitations. Just my humble opinion - like I said, I am not a Linux developer.
That's an interesting post, but I think you miss your own point. Systemd is actually XFree86 in the story.
Maybe I'm just blind and systemd really is the horrible thing that some people say it is. Oh, well, none of my computers have burst into flames because of it yet, lol. I hope the Linux folks will figure something out that'll make everyone happy. As if.
That's probably a pretty good assessment. If you live in a tyrant state and the tyrant state provide you water, food and heroin everyday, you will not care about reality, you will even probably embrace it. That doesn't however mean it is a good state or what you are given is any good. In fact, heroin is very bad.
I've said this multiple times before, but the thing that grinds my gears the most about systemd is not the fact that it exists and has been so widely embraced; but the fact that so much other software has been subsumed by it in the name of "integration."
Because of this "integration" it is no longer possible* to, say, pull only 'udev' from the systemd git tree and build it - necessitating the existence of eudev - resulting in greater duplication of effort.
I wouldn't care at all about systemd if it were just an init system, even if that init system happened to be used by 99% of the other distros - our scripts and sysvinit system work perfectly well, thank you very much. It's just... so against the UNIX philosophy, what systemd has become... so overwrought and monolithic.
At least we, as a Linux distro, have the option to assimilate if we would like to. We could stop resisting, embrace systemd, and keep quiet. But as per usual, the BSD's and other Unices are left out.
But I guess no one (except a small subset of people, many of whom are likely on this forum) cares about maintaining a diverse UNIX anymore. UNIX/POSIX as a concept has lost importance, and Linux has taken over.
I recall feeling the same way when HAL and DBUS were first introduced and only worked for Linux. Hmm, now I'm getting all historical, let me look up that software and see who wrote...
... oh wait.
So anyway, no, now is not the time to embrace systemd, nor will that time ever come in my view. There needs to be at least one Linux distro out there that continues to embrace the old-school way, and remember its roots as a UNIX-like system.
The XFree86 development model was a monolithic source tree, where everyone to contribute. The source code to be maintained only on that tree. If you want a better insight about their development politics, it was just similar with the one applied by Mr. Torvalds even today with the Linux kernel.
But, someday important companies like Intel and AMD wanted to develop their drivers in separate packages, with separate release cycles. There appeared the conflict, and in the end the moneys spoken, and so we get the X.org.[...]
[...]If initially the X.org was just a modular alternative to XFree86, after years nobody used the second one, and the drivers started to become antique.[...][etc]
Quote:
Originally Posted by zeebra
That's an interesting post[etc]
It is also largely a work of fiction... at the best it's a generalisation, at worst it's an opinion based on flawed information or hearsay.
X has a twisted development history, of code, politics and people, which one could probably write a novel about.
The Xfree86 project made same poor decisions - such as removing Keith Packard's commit rights and then expelling him and then claiming that he was trying to subvert the project, etc. The project, as lead by the "core team" was difficult for developers to deal with - AMD/ATI, Intel, etc video drivers didn't even come into it back then (early 00's).
It was after all these events that the XFree86 Licence became the last nail in the coffin.
Packard and other developers forked the version of XFree86 before the new licensing was applied and the result was X.org Foundation's / freedesktop.org's own X11 release (and without the licence problem).
"X.org" X11 gave a more open development model and gives more control and access for developers. It's "cathedral vs bazaar" if you like.
The modularisation came a little later and was aimed at improving accessibility to all contributors:
It was aimed to help Linux distributions and other OS. It certainly helps with components which get updated more often and need to be updated independently (such as video drivers).
While, it may seem that I do like to blame corporations for everything, in the case of XFree86, a certain amount of what could be called "arrogance" within the project and the apparent stagnation of a project which many free OS depended on, contributed heavily to its demise - so as to make any corporate entanglement/interference "blame game" largely irrelevant.
What we think of X.org today and whether corporations have too much influence, or whether it's all about video drivers and 3D acceleration, has little relation to the fate of XFree86.
This thread resembles the reports my wife writes for the government. Somehow she is able to write pages without saying anything. And they like it.
If Slackware ships systemd tomorrow the wannabe geeks wont even notice. They blindly run slackpkg anyway. The fanatics will instantly change their tune because BOB told them so and will prize "Slackware's systemd - the best in the world". And the few remaining Linux users will use the skills they already have.
That's probably a pretty good assessment. If you live in a tyrant state and the tyrant state provide you water, food and heroin everyday, you will not care about reality, you will even probably embrace it. That doesn't however mean it is a good state or what you are given is any good. In fact, heroin is very bad.
I know I marked this thread "solved" but that isn't license for "anything goes". I encourage people to still contribute experiences but also emphatically asked that mere "rooting and ranting" be excluded. Please cease unless you have some actual experience or specific knowledge to convey. With all due respect, the above is nothing but a rant with no substance and no value to anyone. Systemd is not the equivalent of heroin. People are free to choose it or not choose it with no pain, no blatant loss or gain.
So anyway, no, now is not the time to embrace systemd, nor will that time ever come in my view. There needs to be at least one Linux distro out there that continues to embrace the old-school way, and remember its roots as a UNIX-like system.
* To the best of my knowledge
This is the same old "not Unix-like" argument I mentioned in the original post that we don't need to rehash yet again. The integration of which you speak seems a common progression for systems that embrace parallellism, and most do. Personally and for just one example I despise the Windows Registry behemoth which seems a natural consequence when things go parallel. That is mentioned in the linked video in the passage about "DLL Hell".
It is exactly because so many systems have gone parallel despite my own personal experience of that Hell it seems to inevitably lead to, so I can't help but wonder what I'm missing? Why does parallel make sense to anyone? That is why I created this thread. I do mildly understand the concept of quicker massive deployment but why would so many choose it for other niches? I want to make sense out of that, not fan old embers into flames.
This is the same old "not Unix-like" argument I mentioned in the original post that we don't need to rehash yet again.
It's a valid opinion.
Quote:
Originally Posted by enorbet
The integration of which you speak seems a common progression for systems that embrace parallellism, and most do.
Is it necessary though? It's complication for complication's sake. There's no reason for it. That goes against the grain.
Quote:
Originally Posted by enorbet
Personally and for just one example I despise the Windows Registry behemoth which seems a natural consequence when things go parallel.
Why is it a natural consequence? As I see it, there is no valid reason to use binary logging. None. So why did they choose to go down that path?
Quote:
Originally Posted by enorbet
Why does parallel make sense to anyone? That is why I created this thread. I do mildly understand the concept of quicker massive deployment but why would so many choose it for other niches?
Are you talking about the parallel starting of services? If so, in the modern day it makes no difference. The speed of SSDs have abolished that as an issue.
Quote:
Originally Posted by enorbet
I want to make sense out of that, not fan old embers into flames.
Good luck with that!
I found a good article which de-constructs systemd that you might find interesting:
People are free to choose it or not choose it with no pain, no blatant loss or gain.
Exactly. Use whatever OS or init system that meets your needs. Live and let live.
It seems that this topic is akin to politics and religion. People really get fired up about this issue.
....
It seems that this topic is akin to politics and religion. People really get fired up about this issue.
True. When I was a kid my grandma told me the quickest was to start an argument with someone is to either discuss politics or religion. But at that time systemd hadn't been invented yet
In a nutshell, is about some people who understand by "freedom" as something like "forcing" a bunch of companies to support products which themselves declared as being obsolete long time ago.
Could someone explain WHY some software companies SHOULD continue to continue to develop obsolete products and royally pay for, just for the sake of your "freedom" to grab them gratis?
You want badly "free beer" ? Good luck to find that!
There's nothing about some 50 years old UNIX philosophy or whatever, there's just and only about business, people!
The crude truth is that systemd already won by default, just and only because you people are not willing to pay for alternatives.
Stop bitching and show your money for real. If there are some good moneys on standalone udev or logind or whateverd, they will make them for you. That's certainly. BUT, IF AND ONLY IF YOU PAY FOR.
Last edited by ZhaoLin1457; 08-24-2019 at 02:11 AM.
I'm done. I honestly tried to steer this thread to not be a carbon copy of 8-10 years ago, to learn how things have actually panned out. I read the article rkelsen linked and while it was an interesting read it made not one new point and some are questionable. One example of that is claiming an inherent, across the board, 20% failure rate in boot up. I test lots of distros and while I don't spend huge hunks of time in them I do boot them a lot partly because I prefer LILO and most of the systemd distros default to GRUB and it's a fly by the seat of your pants situation to get them to work decently with LILO. I prefer booting to Multi User CLI and for example Debian seems to hate that, and Arch isn't a whole lot better. The only systemd distro I've tried that even mentioned an option to boot to CLI was an older (but still systemd) SuSe... B U T .... I have experienced nowhere near a 20% boot failure rate.
If, as ZhaoLin1457 thinks, systemd has already won (which I think depends entirely on the description of The Contest) and we are all bound to sooner or later to get forced into systemd (I also doubt that) I surely would want to know why and what it can actually do for me and at what cost? I don't think I'm going to learn that here or in any forum. It seems most are all TLDR and don't weish to participate in an actual discussion, a cost/benefit analtysis. Their "systemd" button got pushed so it's time to drag out the soapbox as if we haven't heard/seen it a hundred times already. Example: I do care that "not Unix like" is a valid concern BUT WE ALL KNOW THAT ONE, OK? Thankfully there have been a handful of posts with actual real world experience noted but it all seems still Much Ado About Nothing.
QED...
EOF
I recall feeling the same way when HAL and DBUS were first introduced and only worked for Linux. Hmm, now I'm getting all historical, let me look up that software and see who wrote...
That's an interesting point indeed, but I would say that the BSD people are free to implement in their Kernel some of the things offered in the Linux Kernel, but the fact that they don't and support much less hardware and functions is not really the fault of Linux or those who make Linux specific userland functions for things implemented in the Kernel. Afterall, it is open source, and as far as I can understand, other Kernels (BSD) are free to implement the same functions in the Kernel and use the userland functions to implement it.
I think due to the advanced Kernel, it is inevitable that Linux as in GNU/Linux as an environment is MOSTLY posix compliant.
I think the core parts are and could easily be POSIX compliant.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.