The main concern is that Slackware won't run various "desktop stuff" in the future, if it doesn't adopt Systemd. But my expectation is that exactly that stuff will be abandoned by upstream in a few years anyway. The end result will be a bunch of LennarX distributions with Systemd, running XTerms inside Fluxbox and FVWM. Once Mozilla has abandoned the Linux desktop as tier-1 platform, if will be hard to even get a web browser running.

So trashing Slackware to just have some short-lived fancy bling-bling isn't worth it. Slackware already has an advantage by not requiring a GUI installer, so it doesn't even depend on a display server. We need to keep the stuff working, that makes Slackware a great, reliable and stable experience.

You can have a perfect Linux appliance for that: http://openelec.tv/ That's exactly how Linux is used by the end-user nowadays.

DirectFB is about the only alternative to X and Wayland for desktops but I have no idea how well it would work for a generalized desktop.

Someone can write a nice desktop environment using Qt, running directly on the top of KMS framebuffer...

You can even have (E)GL and even a nice web browser built in the top of (Qt)Webkit. ;)

But, the people will want more, and slowly, your super-nice and simple Qt Desktop Environment will gain Plasma and other Qute Things, and will become a second KDE. But without X... :P

I.e, I look at NetBSD and I see a very simple platform (, but with very poor hardware support, compared to Linux,) which can be emulated by stripping down your Slackware installation to cca. 100 packages.

Then, I see its very nice ports system, having around of 12000 packages (three frakking times of SBo!), with a nice dependency tracking system and all the bells...

Same we go with FreeBSD. But there we have around 40000 packages in the ports.

Linux desktop environments are a thing of the past. ;)

But the point of having a computer is not running a web browser. Everything with a LCD can browse the web nowadays.

Slackware has strengths in serving web applications, databases and files, routing and filtering traffic, being a great development workstation and so on. Don't let fancy bling-bling must-boot-faster get in the way of that.

Well no, even by stripping down Slackware you won't get the small footprint of NetBSD. Just compare the size of glibc with NetBSD's POSIX-conformant libc. And of course, once you have systemd in the system, you wont get anything that even remotely resembles a BSD UNIX...

Well, the Real Life demonstrate that you are epic wrong. Because for every lonely GNU/Linux user, there are 5000 frakking Android users. Whis is (esentialy) a Linux desktop environment.

Sorry, but I don't give a crap for a (personal) computer unable to run a web browser. Sure, I work with servers and I earn a meal from them, but them are NOT my frakking computers.

Sure, using a Linux desktop environment known also as Android...

Sure, you see Slackware's future just as a web server, nothing more. Or as a appliance? :D

BUT, again, I don't give a crap for a great development workstation unable to run a web browser. Because I need a web brouser for reading documentation, news, even to post there. Or you think that I post there using netcat and I read LQ using wget?

Yet again, I don't give a crap about small memory footprints on my computers. The lamest of them have "only" 8GB RAM. So, what's the point to care?

Also I suggest you to finally dump your techno-crap and buy a modern computer... ;)

Of course, you don't have. But there is no law which force Linux to resemble BSD UNIX or to die... :hattip:

Doesn't DirectFB have X-compatible extensions (XDirectFB?) to support X application and DEs like Xfce?

You can still be logged on a Linux development workstation while browsing the web and reading documentation on a different (most likely Linux-based) device. Remember the times, where these "man pages" came along as a printed book? Unix was created in those days. :)

That's the point of the discussion. Making reading LQ and watching YouTube the main purpose of Slackware Linux and sacrificing everything else for this is not the right way. We need reliable Unix for tasks, which can't be accomplished with anything else.

Well, whatever lets you live up to your LQ username but you could use links instead. Which comes with slackware.

I often wonder if the fact that GNU/Linux has no absolutes as far as standards go it's ultimate weakness to all this infiltration being made against UNIX and other UNIX-like standards that have such clear defined standards? It would seem to me, and I'm sorry if this next phrase incurs the wrath of every GNU/Linux user, maintainer, and distributor, all this pride of openness to freedom, liberalistic values and ideas, and willingness to accept chance without merit has done nothing good for GNU/Linux period in the long term, only the short , temporary, and fleeting? And I wonder if this lack of foundation will be the downfall and destruction of GNU/Linux?

I think that we just suffer of the lack of a global system engineering (for the whole OS), as we have for the kernel. It's good to have only one person who ultimately makes integration decisions, and that partly explains why Slackware has been able to survive that many years, unfortunately no one controls random evolution of available components and sub-systems. Every time one of these changes, appears or disappears, we can just cross fingers hoping that integration will stay possible without regression of features and/or of non functional requirements previously fulfilled or met.

Exactly - hence why Red Hat hired many of the developers of the key components. Without being able to control your own destiny, it makes creating and sticking to software road maps very difficult.

Yes, but Red Hat is only a distribution/distributor of GNU/Linux, they shouldn't be the ones making the decisions on the future of GNU/Linux, the GNU project and the Linux Foundation should be making those Standards and deciding the future of Linux. This is basically going back to my argument on someone adding the proverbial onion to the soup that has no business in the kitchen or near the food.

It's nice that some companies and groups want to contribute back to GNU/Linux, but when a contribution is aimed at destroying the UNIX ecosystem to which GNU and Linux both belong to, along with hundreds of other projects, there should be Standards in place to prevent this from going any further, and those Standards are either too loosely defined, non-existent, or unenforceable by whatever means.

I honestly think an Open UNIX Specification that would state that any key structural system software that works for the GNU/OS has to be openly available for any other UNIX or UNIX-like system, minus tools developed by the Linux Foundation specifically to manage the kernel only, and if not, it can not be included in a public GNU/Linux distribution, similar to the Copyright laws that prevent CDDL licensed software from being redistributed along with a GPL licensed project in source or binary form... and then actually enforce the rules.

Fashion rules your life - the dark side
 

I am your father -- Lennart

On the flip side, maybe this will help in Hurd development. :p

Brief Off Topic Intermission
 

Yes I remember.

My 1978 7th Edition UNIX books (mostly man pages) reprinted in 1983. I haven't used them much in the last few decades however. :)

(Note: I'm not a time traveler from the future. The camera date/time is still set to an Asian time zone.)

It would be nice if HURD ever became viable for production level systems, but I doubt it'll happen any time soon. At this point the only safety nets left are big brand UNICES and the BSDs laying down the law on projects and the willingness of other Linux distributions to not jump off the bridge like everyone else.

Maybe in 1995 that was the case but they are not and haven't been for quite some time since they became established in the the enterprise space in early 2000s.

They aren't, directly - and that was part of what I was getting at.

They're a business with share holders (I used to be one since I worked there a few years ago). Customers and investors need to know what's coming down the line: why should they stick with the company? The company needs to be able to execute against it plans, which means taking control of what ever it can. There is a lot of good that is done by RH which is contributed back to the community, and part of the price to pay is that some of the direction of the community is swayed by RH due to its own needs.

Vincent Batts works for Redhat, btw. I'm not sure if he's still considered a "Slackware developer", or if he ever was, but I do know that his contributions to Slackware have been substantial.

as far as I remember, he started at slackware and moved to red hat.
over the past years, he has been contributing still to slackware

Red Hat should not be in control of upstream GNU/Linux core system developments to the point that the only software being used is Red Hat's own developed stuff like systemd for example. Just because Red Hat proved years ago that a company that promoted GNU/Linux could be profitable shouldn't be an open invitation to Red Hat to do as they please, nor give them any exclusive rights to do so, but apparently nobody questions why Red Hat's tentacles are so far deep into open source systems without some level of a stop gap, and why they allowed so much freedom to do as they please unchecked.

Not to restart a debate or kick a dead horse but you do have to realize that besides the kernel control tools, kernel, shell, and other general utilities is one major component of the system that nearly controls everything called the init system. Now seriously think, who created this new octopus of an init system, and who does he work for, and who would like to see less to zero competition in a market? The answer isn't difficult to decipher.

You said it best. Red Hat is a business and what do businesses try to eliminate to make more money?

"Timeo Danaos et dona ferentes", is that what you mean?

If you actually imply that systemd is a Trojan horse send by RH to eliminate competitors, I won't follow you on imputing them these motives.

And in any case, nobody's forced to follow them and eat the meal they cook.

But be it true, good for them. We all compete to survive, either in person or through our descendants, that's life.

Live and let live.

PS On a broader scope, I don't think that we (mankind) are smart or wise eliminating other (plant and animal) species from Earth to grow, but that's another topic.

Lennart Poettering started with it in his free time. He initially wanted to contribute to Upstart instead, but is no fan of Canonical's CLA. After debating with Canonical's lawyers to put the CLA down, to no vail abviously, he decided to start his own init system. I can't see how his original plans to contribute to Upstart instead are somehow motivated by Red Hat to want to be in control, in fact, if Canonical wouldn't insist on their CLA there would be no systemd, but Upstart with much more adoption. Is Canonical insisting on their CLA part of Red Hat's plan to take over the GNU/Linux world?
If Red Hat, the market leader for enterprise Linux systems, wanted to see less to zero competition, why do they release their stuff (L)GPL, why do they openly invite their competitors to work with them together on their software? Why do they allow their largest competitors, Suse, Canonical (commercial competitors) and the likes as Debian (non-commercial competitors) to contribute, if their aim is not to help the competition, but to extinguish it? This does not make much sense, I would think.

A large corporation never has a single controlling mind, and it's silly to imagine one. It's entirely possible for a large corporation to be nobly philanthropic and shockingly evil at the same time.

It can't be denied that the systemd / Gnome 3 / gtk+3 affairs have generated a lot of bad sentiment in the community towards Red Hat. That bad sentiment has the wrong target -- "certain projects clustered around Fedora" might be a more useful substitute -- but it undeniably exists. So the relevant question is, does Red Hat senior management know or care about the bad feeling? The answer, apparently, is no. That's not surprising; they're focussed on enterprise customers, the bad stuff hasn't hit their enterprise customers yet, and who will ever listen to the 'techies' when it does?

EXCITING UPDATE!!!
This blog post from Fedora's own Matthew Garrett, about Red Hat's genuine commitment to the community... and how they could still get it so wrong when they tried to eject Openstack competitor Piston from the Red Hat Summit this weekend:

It might be instructive to know if the boys over at Debian feel manipulated and maneuvered into place or the recipients of a beneficent gift.

Well for years there was pretty much no choice: sysvinit or nothing, now as a result of upstart/systemd there are a few choices of init systems and there could be more, not to mention more distros based upon them - so in a way this is a positive outcome if competition is desired. One can't really stop distros making the choice to switch to systemd - we may not agree, but it's developers who do the work, put in the hours and make distros available - and unfortunately it's big corporations funding that. Even your *nix of choice FreeBSD/PC-BSD has a relationship with Apple and receives sizable donations from google and others.

FreeBSD is also developed as a complete OS, a base system, so you really don't get to chose the init system anyway...

So really I think the anti competitive thing or "Red Hat takes over the world", is just digressing from your main argument which seems to be more about adherence to POSIX compliance and KISS principles? I can agree on that count, but systemd is not the only software of this type to creep into Linux distros over the years.

Didier, please edit out the political commentary at http://www.linuxquestions.org/questi...ml#post5123573 that has nothing to do with RedHat, systemd, or Slackware.

There are forums to discuss that topic and this isn't one of them.

I'll just quote Russ:

Full post: https://lists.debian.org/debian-ctte.../msg00390.html

¡Arde Troya!
 

Some people react against closed code, a good side effect at least! WYSIWYG
interfaces instead win total acceptance. WYSIWIG interfaces are the Troyan Horse!

Mom is bad. Gramma gives me candies, Gramma loves me.

We were sold the personal computer like one more electrical household appliance. That was the origin of the lie.

Desktop? It was replaced by mobile devices. The origin of systemd.

Server? Today most companies out there will not lose a minute i.e. hacking Slackware rc files (more than the half of the "Slackers" here neither, they come here to cry because Xfce/KDE didn't do this or that automagically). Everybody wants to push the ON button and call the service when something is broken.

That education has been implanted long time ago by Microsoft and Apple. Server users, desktop users were educated in that way. What do you want RedHat do? To explain their clients Unix simplicity? It would break tomorrow doing that.

Now you blame Gramma. Gramma lied you.

Actually with FreeBSD this changed. There is a choice, but only because bsdinit is kept as the default. Runit is in ports collection as an alternative if you so wish to switch to it. Documentation and sample run scripts are all on the author's website. Gentoo's BSD creation uses OpenRC which has been looked into for ports but due to lack of someone willing to write up the port for it, it hasn't been ported in yet to the main FreeBSD project. ArchBSD has a package for it though with scripts and a build script.

It may seem a far fetched argument, and possibly conspiracy theory, but when too many facts start making sense, too much against the norm occurs, and certain parties of interest start doing thing publicly such as trying to trashing POSIX as part of a manifesto, attempt to disrupt BSD and UNIX's ongoing efforts to create a working desktop environment by denouncing ithem as obsolete, and then trying to play nice by saying they are all for GPL/LGPL while pushing out non-RH developed software for RH-developed and controlled, you get to a point where conspiracy theory isn't just a conspiracy theory, it's just a conspiracy.

But those are just my observations.

Generally I'm skeptical about conspiracies largely because they're so hard to "pull off". That and I learned the old adage that "It is usually wrong to assume as malice, that which can be more easily explained by incompetence". Someone pointed out the lack of a single mind in corporations and with todays multi-layered, multi-national corporations not only does the right hand not know what the left hand is doing, they arm wrestle.

At this point it seems fruitless to attempt to see if this was a RedHat conspiracy. Even if RHEL 7 is a huge success nobody is going to release a statement saying "We lost share to Debian but only for a minute because we knew our machinations were in place that would turn them into 'also-rans' and we would once more dominate. Next is Microsoft. Salut!"

Since it is unlikely we will ever get reliable and direct confirmation all we can do is watch to see the fallout. The Init Wars are effectively over. Systemd already won in what feels like Blitzkrieg.

All that remains now is to see if RHEL 7 falls flat on it's face, confirming the worst fears about systemd, or if they make it work. I, for one, hope to finally find out what gains were/are expected aside from boot times. I still find this very cloudy.

With no other gains my path is clear - stick with or as close to SysVinit as possible perhaps watching to see if BSD gets a boost from all this "tempest in a teapot".

Thankfully, my distro of choice, Slackware, has at the helm a man more concerned with function than fad. It seems the most compelling reason to go to systemd is that udev was swallowed by it. Surely with all the die hard Unix devs out there, that can't be that hard to replace, if push comes to shove...and it has.

I totally share this wait-and-see, sky's-not-falling attitude. No one is twisting anyone's arm just yet, and won't for another year at least. udev, like you say, can always be forked. systemd developers themselves may come to realize that things they denounced in the beginning will have to be implemented if it is to perform on servers, and then we may see things like plain text logs, rock-solid compatibility with traditional inits, and who knows, may be even excellent documentation. And if the only gain is faster boot time, then so be it. Can't say no to that.

Edit: post count for the win.

The way some of you guys incessantly spend so many hours talking about systemd, you'd think your lives depended on it. It's a 1980s init replacement. ReaperX7, is your life going to dramatically change if systemd becomes the standard? Do you think init is that bulletproof and flawless that it's the worst thing in the world if it goes away? If init is so important to you I'm sure you can help find a way to keep it around for the rest of your life. It's just kind of weird to me that you're so concerned about it.

If done properly systemd would have been more widely accepted, but it's not due to many factors we must again need to re-dig up.

The fact that it's a direct slap in the face to the entire legacy of nearly 50 years of contributions by the entire UNIX ecosystem on the whole is a major beef with me.

"Write programs that do one thing and do it well." - Doug Mcllroy on the UNIX Philosophy

This is not the philosophy of systemd nor it's creator.

Systemd is not a program that does one thing, and does it well. It's an amalgamation of up to 70+ subprograms that all try to do many things, yet doesn't do them well, it just does them, and if it screws up, there's no way really of knowing until the system is supposedly booted because systemd runs in a sealed environment like a locked hypervisor.

You also have to look at who designed systemd... Lennart Poeterring who stated that kFreeBSD was a toy OS, BSD on the whole doesn't matter, and any distribution that didn't adopt his monstrosity was lost, confused, and doomed to failure without it. This is also the man who said to take POSIX and dump it in the trash.

Sysvinit is an init program, all it is is an init program, and it's done it's job very well to the point that it's been only in maintenance mode. It's been supplemented by programs like perp, runit, and various other software to add extended functionality on top of sysvinit. It can even be reformulated with RC-init scripting (Slackware's init style which adds sysvinit native script support) to work like bsdinit to expand it's capabilities. Sysvinit rarely, if ever, has any issues at all so much so that since x86_64 architecture arrived very little patching has been required of sysvinit. yes it was inherited from System V UNIX, but it follows the UNIX Philosophy entirely. It's one program that does one thing, and does it exceptionally well.

The sound design and operating principles of Unix **ARE** the value proposition delivered by GNU/Linux.

Linux was not conjured out of thin air, it was designed from the ground up as a freely available Unix-like operating system.

From kernel.org:

Removal or replacement of elements of that Unix foundation removes value from Linux, or replaces it with something else. And inasmuch as Linux is currently a major mover among the available free Unices, removal of Unix value from Linux has far reaching detrimental effects to the free Unix ecosystem itself.

Systemd, so far, has seemed to be a wholesale and broad assault on those underlying Unix value elements... little wonder those who appreciate and depend on those valuable elements are so "incessant" in their defense of them.

Speaking only for myself, I use GNU/Linux because it is a freely available Unix operating system. I use Slackware because it is the most Unix-like of the Linux distributions. If the Unix value proposition is removed from Linux I will have to use something else... I hope that does not happen.

Exactly. Clone of UNIX, or not, all UNICES whether they are UNIX-based or UNIX-like all share some level of interoperability betwixt and between each other.

UNIX matters today now more so than it ever has. While AT&T's creation and Novell's patented UNIX haven't really done much, it's children have branched out considerably from the tree they all once called home. They've had children and their children have had children.

UNIX in some way runs phones, embedded systems, tablets, even MP3 players, and other systems of various origins.

Even Microsoft wouldn't be where it has gotten without UNIX. In fact Microsoft owes a huge debt to it's existence because of UNIX. If it wasn't for Xenix and the fact Microsoft couldn't compete with AT&T, Bill Gates would have never been able to buy the rights to that obscure 86-DOS (QDOS) from Seattle Computer Products he eventually redeveloped into Windows.

And without Microsoft the personal computer might have not made such as great an impact thanks to mass marketing by IBM.

Without the IBM personal computer, we wouldn't have had 386BSD developed as an alternative OS on the Intel platform, and without 386BSD being delayed due to a lawsuit, Linus Torvalds would have never developed Linux.

Even though Linux may never get 100% SUS or POSIX compliant, it doesn't mean that we have to ditch those efforts to move ahead. SUS and POSIX are long term goals of Linux, BSD, and all other UNIX-like systems. We might not be able to pay the money to get the UNIX rubber stamp of approval, but we can at least adhere to the standards put forth by the POSIX and SUS specifications as best we can.

Agreement with ReaperX7. It's not just systemd, it's this whole wall of bad changes of which systemd is the kingpin. We can either despair and give up, or mind our own back yards and keep them in order.

Started poking at eudev today. It seems to be actively developed. Haven't compared it to mainstream udev yet to see how they've diverged (besides systemd).

Also, I found that I'd been laboring under a misunderstanding. Gentoo developers forked and develop eudev, but it is not the main udev replacement used by Gentoo (though it is an option for Gentoo systems). Gentoo also forked a later version of udev to free it of the systemd dependency, and are calling it "udev". This is the udev primarily used by Gentoo.

I'll give Gentoo's udev a spin once eudev's been spun. I'd like to see if eudev jfw with Slackware 14.1, and if it exhibits any problems with my various hardware (I have six systems to test it on). Will let you know how it goes.

Gentoo's contributions to FOSS have been largely overlooked because Gentoo is non-profit. They've made a lot of projects that have really been snubbed at in favor of others by those from profitable sources.

If we look at the history of the USA, we see that when a group refuses to adopt the laws by which others have decided to live, the group will get to live in an Indian reservation.

When all major distributions have decided to adopt SystemD, Slackware boldness to not to adopt this new management platform will most likely result in reaching Slackware on a Indian-like reservation.

I, for one, I welcome the Big Bear overlord!

But Slackware will not be "a Linux" anymore...

Make no mistake, while I am not worried the sky is falling I am disturbed by systemd's strongarm methods and I do think peoples arms have been figuratively twisted. I also wonder how much of this is a play on RedHats' part to satisfy stockholders and recover from a market share loss to Debian/Ubuntu. Furthermore I have zero trust in LP. I don't like his manner nor his methods.

I am not freaking out because I am first and foremost a Slackware user but I am surprised the ripples at the Debian camp aren't more tsunami-like. If this is the time where Linux forks then so be it. I know which side I will be on and it isn't Lennart's unless I see some really compelling advantages and boot times is NOT one of them. So far even the more enthusiastic adopters have literally nothing to say as to why this is so compelling. Anyone have an answer for that?
.

Not directly an answer, and while I'd like Slackware to stay with the current init system, I think this might have won over some people: http://0pointer.de/blog/projects/why.html

Quite a funny statement, systemd was definitely started on the desktop and the most used Linux based mobile OS, Android, doesn't use it. You might want to elaborate on that.

Anyways, your whole argument comes down to: Don't invent an OS for the masses, they can't fix it when it breaks, so they don't deserve an OS that fits their needs.
The funny thing is, something like the changes in the Linux world nowadays happened already with other things. Like, for example, the car industry. In the early days of the automobile the driver also had to be the technician, while nowadays most people that drive a car do not know how to fix it. And that is a good thing, why should people have to learn that? There are a number of well paid technicians there to do that for them.

But I would like to know how this (or your rant that desktop environments dare to automate things for their users) is related to systemd, I see no link.

And you can't stop the progress... :hattip:

So, as I already said (you may have missed that), systemd exists only for the one reason that Lennart Poettering and Kay Sievers did not want to submit their code to Upstart under Canonical's CLA, something that I would fully agree with. Would you like to explain Canonical's part in your conspiracy theory about Red Hat wanting to control the Linux world?

If Slackware would become a BSD-like Unix combined with the good hardware support of the Linux kernel, I'm perfectly fine with that.

There are over a hundred distros not using systemd at this time. And the Slackware not being a "Linux anymore" is a strange comment, as long as Slackware uses the Linux kernel it is as much Linux as any other distro, and since its the oldest distro still going imho Slackware is "THE" Linux distro! :Pengy:

I've read that before and besides the fact that it is going on 5 years old, when systemd was barely a wet dream LP had, it is written by the man himself and still is only a contrived bullet point wishlist. Why can't any ADOPTER say what it is that is so compelling, what systemd does for him and his system that he cannot do now with less complexity?

Example - I just watched an interview with Jos of OpenSuse and all he would say is "Systemd is the future". He never says why or to what advantage, just that it is and I see this over and over and it stinks.

I'm not exact with the timeline but didn't systemd start out as a pet project and THEN become a paid for by RedHat project? Is it not at least possible that RedHat saw an opportunity much like Microsoft did when IBM assumed they owned CP/M? or Steve Jobs did upon visiting PARC? Isn't there considerable precedent for exactly this sort of opportunism?

Someone else pointed out 2 important things here

1) A corporation is not of one mind and can be both humanitarian and evil at the same time

2) Corporations need to show stockholders a solid, well-defined future path with little room for variables

and despite Software's unique ability to "have one's cake, sell it and still have it, too" isn't business still all about ownership? and therefore doesn't it behoove corporations to "lock it up"?

Why is this so far-fetched then?

I'm perfectly fine with your concept too, but it will not easily be achievable if the good hardware support starts to rely on udev, and if the desktop environments start to rely on logind. But it's not all bad news. The BSD people and the non-Linux Debian people have every incentive to create workrounds for logind.

Meanwhile, Slackware 14.1 is what it is -- the 2014 LQ distro poll-winner and runner-up. I have it. You have it. The mirrors have it. "They" can't take it away from us. It works now and (with security fixes) it'll still work next year. So there's at least a year, maybe two, maybe *twenty*, before Slackware-next needs a roadmap that includes the new Linux plumbing. The goblin horde is not yet at the drawbridge and Mr Volkerding is not yet a prisoner in Isengard. The future will bring both unforeseen problems and unforeseen solutions. People should chill. Let the slack times roll.

If being "labeled" is bad and staying true to UNIX principles turns Slackware into "Soaring Eagle" or "Wandering Bear", I'll stick with Slackware and it's old fogey ways while the elders tell stories of times many moons ago through the harsh winters of greater times when the great GNU walked the lands, and Great Elder Patrick brought peace to the tribe, the lands green with grass, the water flowed freely, and SUS and POSIX were considered sacred... Praise Bob.