NEVER let Freebsd become like Ubuntu

Status
Not open for further replies.
Well, we all keep saying that, and I'm definitely one of the people who complain about systemd all the time.
HOWEVER: It (systemd) does attempt to solve a real problem.

Agreed 100%. Systemd is not all bad. In fact it has some awesome features (try journalctl), and attempts to solve some long standing Linux/Unix problems. It’s just invasive, really invasive. Distros need to be all-in or all-out, instead of shipping these hybrid configurations. The Arch Linux systemd experience is actually not bad because it’s 100% customizable.
 
Part of the problem of systemd is that it has to be invasive. Matter-of-fact, to clean up the ugly mess of electrical tape and baling wire that holds a normal Unix system together, it needs to touch a huge number of components. That means that people who do any configuration/maintenance/management of those systems will have to relearn everything. Most people don't like change, me included. And because that wasn't handled correctly (from a sociological point of view), it has caused a huge allergic reactions. And unfortunately, the one person doing it is unsuitable for the task of "socializing" it, rather on the contrary, because he's such a prick.

My hope is that over the next few years, it will get cleaned up, documentation will improve, a sensible set of people (under some corporate umbrella with steady paychecks) will handle maintenance and improvement, and after a long while of cleanup, it might be a joy to use. We're not there yet. One of the first steps in that journey is that the "community" (meaning the CEOs of companies like RedHat and Suse) need to get rid of the narcissistic sociopaths such as Linus and Lennart.
 
Part of the problem of systemd is that it has to be invasive. Matter-of-fact, to clean up the ugly mess of electrical tape and baling wire that holds a normal Unix system together, it needs to touch a huge number of components. That means that people who do any configuration/maintenance/management of those systems will have to relearn everything. Most people don't like change, me included. And because that wasn't handled correctly (from a sociological point of view), it has caused a huge allergic reactions. And unfortunately, the one person doing it is unsuitable for the task of "socializing" it, rather on the contrary, because he's such a prick.

My hope is that over the next few years, it will get cleaned up, documentation will improve, a sensible set of people (under some corporate umbrella with steady paychecks) will handle maintenance and improvement, and after a long while of cleanup, it might be a joy to use. We're not there yet. One of the first steps in that journey is that the "community" (meaning the CEOs of companies like RedHat and Suse) need to get rid of the narcissistic sociopaths such as Linus and Lennart.
That's not happening. Waiting for people whose names are synonymous with certain software to leave is fruitless. If you want that, you'd have to fork it, or better yet, you're already on a good OS.

Systemd is also a bad starting point for anything. Why want to fix that?
 
Regarding systemd, I'm surprised no one has mentioned The Tragedy of systemd, which includes a bit about how the BSDs could learn from systemd and perhaps implement something similar but better?

Back to the OP, I'm a bit upset that I've had to go back to using Linux on my laptop just to have the things I need to work working. FreeBSD really seems to be the better-organized/developed operating system base, but it's hard to get my desktop programs set up correctly, user mounting through the file manager, power menu options, etc. There are workarounds, but they were a bit too painful for daily use. Still, FreeBSD is much simpler to configure, IMO.
 
hard to get my desktop programs set up correctly, user mounting through the file manager, power menu options, etc.

The original fault of this goes to the Gnome 3 denizens who broke the sudo / hald system which used to work pretty well on FreeBSD.

However, it isn't all the fault of the Gnome guys. I suppose part of the bad decision was on the FreeBSD porting guys who unfortunately decided it was better to have a very broken bleeding edge version rather than keeping with a slightly older version of Gnome that was fully working. I personally find the "latest is greatest" mindset a little bit childish and unworkable.

The stupid thing is that keeping with the older working version would have actually been much easier too :/. Oh well, getting used to a spartan Window Manager only environment is more flexible in the long run for computer enthusiasts. It also means you can rely less on others.
 
My hope is that over the next few years, it will get cleaned up, documentation will improve, a sensible set of people

Good points. Correct me if I’m wrong, Lennart, despite the controversies, was the real victim. He was trolled in extreme ways, and even recieved death threats, for which he had to go to the police. Nobody deserves that.

Also correct me if I’m wrong, in one interview Lennart hinted that Freebsd might be working on a new init daemon? That should probably be a different post!
 
Some say systemd is just a poor implementation of SMF from Solaris. Unfortunately, the Linux community doesn't have a developed frontal lobe - so they reinvent things poorly than adopting existing proven solutions. Solutions that solved these aforementioned "problems". Btrs, eBPF, systemd, epoll, ext4, LXC, etc, etc, are all re-invented nonsense. Almost no innovation came from Linux, but from FreeBSD or Solaris.
 
Good points. Correct me if I’m wrong, Lennart, despite the controversies, was the real victim. He was trolled in extreme ways, and even recieved death threats, for which he had to go to the police. Nobody deserves that.
That doesn't make him into a good guy. Perhaps into persecuted guy, but being persecuted is not a moral victory. Even less does it mean that the software he worked on is well designed. And systemd isn't the first strike; both his Avahi daemon and the sound system were similar octopus-like disasters which have tentacles in everything.

A large part of the problem with Lennart and his systems is that the Linux ecosystem has a long-standing culture of "dominating alpha-assholes". If you are technically astute, and really obnoxious, you can take over the communication mechanisms, and make all the decisions. Whether you have the overall system architecture and engineering skills to make big decisions or not. A related part of the problem is that super-coders (like the young Linus, or Lennart) get decision-making credentials, even though they don't have the big picture skills. At the same time, sensible people who know how to engineer large and sustainable systems (like Andy Tridgell) get pushed out, because they don't want to play the a**-h*** game.

I'm always hoping that as Linux gets commercialized (a large fraction of the important developers are employed by RedHat, Intel, Suse, IBM, Oracle, ...), this kind of stupidity gets made more sensible.
 
My impression is that Red Hat only cares about controlling the largest slice of Linux development. I don't think they care about code quality (at least with desktop projects) or a specific architectural direction. Oh, and Lennart deserves everything thrown at him, including death threats. (No, personally I wouldn't bother.)
 
Lennart deserves everything thrown at him, including death threats. (No, personally I wouldn't bother.)

That's a bit much, don't you think? If anyone deserved death threats, I'd imagine it would be people who have killed others. Not someone who thinks their crap software is good, even if the crap software is used more than it should be.
 
That's a bit much, don't you think? If anyone deserved death threats, I'd imagine it would be people who have killed others. Not someone who thinks their crap software is good, even if the crap software is used more than it should be.

I don't believe he ever was in any danger, it's just a misdirection tactic on Lennart's part. Although, as a Russian guy I wouldn't mind him being punched in the face a couple of times. Different people might have different cultural standards.
 
My impression is that Red Hat only cares about controlling the largest slice of Linux development. I don't think they care about code quality (at least with desktop projects) or a specific architectural direction. Oh, and Lennart deserves everything thrown at him, including death threats. (No, personally I wouldn't bother.)
No, RedHat and consorts only care about making money. Matter-of-fact, as publicly traded corporations they really have to care about making money (if they fail to do that, their shareholders have every right to sue them into the ground). Now, they can be stupid and make money in the short run and shoot themselves in the foot for the long run, or they can be intelligent, and make money in the long run. In my humble opinion, they are intelligent, and are trying to make money in the long run, by creating and selling good products.

Note that they don't make money by creating code: All the code in Linux and its various distributions is open source. The code and its quality are not in and of themselves their product. They survive (reasonably well) by selling support and services. Having good code is a prerequisite for that business. Note that the big ones (Suse and RedHat) make a lot of their revenue from products other than support and services, such as consulting and administrative tools (such as Ansible).

An interesting question is this: Many discussions on this forum and in the FOSS world in general are about desktop usage. If you read this forum, it is dominated by people who run X-windows based GUIs on their FreeBSD machines, and are mostly upset about problems with desktop tools and GUIs. But that's actually not where the FOSS market is. The vast majority of all servers in the world run some form of Linux (a small fraction run Windows, and other Unixes have a vanishingly small market share). The vast majority of all desktop users run Windows, with MacOS second, a small number of Linux machines, and the other desktop OSes vanishingly small. Just think about that comparison for a while, and you recognize that the important thing in the *nix world are servers; desktops are mostly for hobbyists. And companies such as RedHat and Suse and iXsystems can not survive on the revenue from hobbyists, because those typically don't pay for their OS.

This still doesn't answer the question: Why does RedHat allow systemd to be created, and why do they allow people like Lennart to mess things up? I don't know for sure. Part of the answer may be somewhat hidden in what I said above: The traditional way of managing daemons/servers in *nix is a horrible mess, and hopefully tools like systemd can eventually clean that up, and make these systems more reliable, easier to manage, and more profitable. I think RedHat understands that to get to the long-term good state they have to go through a "valley of tears", and today systemd is in that awkward teenager phase of its life cycle.
 
No, RedHat and consorts only care about making money. …

Obviously, I believe that controlling the ecosystem makes it easier to sell things built on top of it. The part where we disagree, is that it all must be comprised of good code. Mediocre and even awful projects are all very useful as long as they contribute to Red Hat's dominating position and don't make experience substantially worse for clients. Indeed, some of them might eventually mature into actually serviceable software, but it's a pretty good business as it is.
 
This still doesn't answer the question: Why does RedHat allow systemd to be created, and why do they allow people like Lennart to mess things up? I don't know for sure.
I have to work on Linux (Redhat, CentOS) as part of my job. There may be some innovation in Linux, but I don't know what it would be. They invariably knock off everything that has been done before. systemd is not any different than SRC in AIX or SMF in Solaris. In the case of AIX, that has existed since at least 1995 when I started on it with (I believe) 3.2.5, and in the case of Solaris, since 2005. Dynamic i-nodes were available in JFS2 in AIX 5 which was 2001. Solaris released ZFS in 2005.

AIX has multiple page size support.
AIX has had multibos or live update support for over a decade, and then Linux came out with k-splice (another knock off).
AIX Workload Manager.
AIX Advanced Memory Expansion.
AIX Active Memory Sharing.
AIX LPARs/Workload Partitions.
Solaris LDOMs/Containers.

Those are enterprise features.

It always humors me when Linux cheers the adoption of X when it has been available for at least a decade - and sometimes two - in AIX or Solaris.

But the real question isn't "don't let Y become like Z; or why is Y like Z?", because they are both the same thing, based on X. It is, "Where do we go from here?"

UNIX was released in 1971 which is nearing the 50 year mark. BSD and Linux are all based on UNIX.

Plan 9 was developed in the late 80s to fix the problems with Unix like networking and graphics. But still used Unix as the inspiration.

Then you have Windows which was released in 1985, which is nearing 35 years.

Minix based on Unix.

Akaros, NIX, Clive, Inferno are all based on Plan 9. These four are distributed operating systems and only Inferno was released commercially and never caught on. The other 3 are research operating systems.

Android based on Linux.

iOS part of the macOS family. macOS originated with NeXTSTEP which was based on Unix.

OS/2 was a hybrid of Windows and Unix.

BeOS was not internally Unix, but after Apple went with NeXTSTEP, it never gained commercially.

There is nothing on the horizon to replace either Unix or Windows. In the next 50 years are we still going to have the same thing? Au Gratin, french fried, mashed, cheesy baked, tater tots, roasted. However or whatever you call them, they're still potatoes.

Plan 9 was thought to be better than Unix, but Unix was just good enough. Just good enough is what we strive for in an operating system?
 
What we have now works, even windows, which has it's market share because of market flooding and certainly not because of quality, but it does a decent job for desktop users that don't know any better.

Not sure we will ever reach computing nirvana in terms of operating systems. Maybe not in my lifetime but who knows. I am pretty happy now using FreeBSD as a desktop. I am certainly not the majority though but I have always bucked the system and done what few do because I hate being conformist.

Despite Linux's past and direction, I do appreciate the fact it was born a "rebel" so to speak. Open source, to me, is really what is important because it is a community. I can't stand having a single OS shoved down my throat by a company that could care less about its users and whose only driving force is profit: "this is how you will use your computer, and you will like it!" Open source represents freedom to me, and gives us all choices, which we didn't have prior to its inception. Sure, there may be quality issues but I can say I have had fewer issues with open source projects than I have had with commercial ones, because if I don't like something there are 14 other apps to choose from.

Computing to me, is a very large part of my life. I spend a tremendous amount of time behind the monitor and have done so for 20+ years. That I have moved of to a small corner of the computing world in FreeBSD is fine with me. I have changed the way I do some things fundamentally so I could embrace FreeBSD because I like it and appreciate what it is. Thank you to all the folks that have made it possible.

That was a bit of a digression but my mind works that way :)
 
Agreed 100%. Systemd is not all bad.
Unfortunately, I do not agree with this statement. I'll look at it from my belfry and mention Debian. This situation prospects a lock in systemd dependencies which is de-facto threatening freedom of development and has serious consequences for Debian, its upstream and its downstream. This situation is also the result of a longer process leading to the take-over of Debian by the GNOME project agenda. Considering how far this has propagated today and the importance of Debian as a universal OS and base system in the distribution panorama, what is at stake is the future of
GNU/Linux in a scenario of complete homogeneization and lock-in of all base distributions. Init Freedom is about restoring a sane approach to PID1, one that respects diversity and freedom of choice.

P.S.

If any of these claims are offended or turned out to be trolling, please forgive me and not be disappointed.
 
Devuan aspires to be Debian without systemd and, if I didn't have FreeBSD, I would probably be in a bigger hurry to check out Devuan, but I do have FreeBSD, thankfully.

Was impressed with Ubuntu at first, but very disappointed when I discovered "unattended-upgrades" or whatever it's called, running in the background the moment X gets started. Attempting to shut down the host, one discovers that Ubuntu has the same sort of daddy-knows-best dialog as Windows Update, saying something like "Don't try to turn off your computer, because I won't let you shutdown until I've finished my automatic updates, which are far more important than whatever trivial task it was that you bought this computer to do." Bad juju for me. Linux Mint has a similar auto-upgrade feature, but at least there it's easy to deactivate. Ubuntu is prettier, but Linux Mint has better manners.
 
So they can make money off of customer support for something that is not reliable. That's their business model.
Yeahh, and to ensure that customer support contracts keep being sold, it is essential to break things systematically or else customers will eventually learn how to fix stuff themselves.
 
Ubuntu Linux is often touted as “Linux for human beings”. This is really interesting. The other day on an Ubuntu 18.04 machine we were struggling to setup some very basic networking configurations. Something as simple as setting a static ip on a machine turned into a science project.

A few years back, simply updating /etc/network/interfaces, along with /etc/hosts, would have done the trick. Not anymore. It was then discovered that, once again, there’s a new networking application on Ubuntu. Introducing netplan, a new utility which solved the problem.

Without delving into netplan, the point here is to highlight the chaos of an Ubuntu network configuration. We have:

netplan,
network manager,
systemd-networkd
,
systemd-resolved,
nmtui,
nmcli,
systemd service files,

/etc/hosts,
/etc/network/interfaces,
/etc/resolv.conf, /run/systemd/resolve/resolv.conf, /etc/stub-resolv.conf,
/etc/nsswitch.conf
,
/etc/netplan/netcfg.yaml,

all working together as one happy family. We can be certain there are more files but you get the point.

We pray that Freebsd NEVER becomes like Ubuntu.

It will. Law of Moore.
https://en.wikipedia.org/wiki/Moore's_law
More transistors, more energy, and looks like computers are "slower" as ever ;)

The longer it takes, the larger the number of source code lines.

If you want to keep it clean and like Unix, you need to develop a new kernel and stick to Unix Originals, like they do:
https://github.com/DoctorWkt/unix-jun72
 
More and more do I appreciate that 99% of things can be configured in one place on FreeBSD (rc.conf).

On Linux it is becoming more and more inconsistent and eclectic than ever. These days I highly recommend doing everything (albeit manually) in /etc/rc.local and if they have removed that, then just do it under the root @reboot cron entry haha.
This is probably bad advice but I just cannot be arsed with Linux any more. They may have systemd but I notice that every distro (and release) now does networking differently and... messy quite frankly.

The big one for me is that on Debian the "long interface names" feature (i.e wl023spd32 instead of wlan0) has long since broken wpa_supplicant making wifi useless. And yet they haven't fixed it for multiple releases (just a slightly out of date bug report). Leading me to believe that Linux desktop / laptop usage is actually miniscule. Again I blame Gnome 3+ and KDE 4+ for being unfit for general purpose.
 
More and more do I appreciate that 99% of things can be configured in one place on FreeBSD (rc.conf).

On Linux it is becoming more and more inconsistent and eclectic than ever. These days I highly recommend doing everything (albeit manually) in /etc/rc.local and if they have removed that, then just do it under the root @reboot cron entry haha.
This is probably bad advice but I just cannot be arsed with Linux any more. They may have systemd but I notice that every distro (and release) now does networking differently.

It is getting maybe not Unix, really, like it was. Or Unix is evolving. Depends on opinion.
 
Well I guess Linux *is* just a kernel so we cannot use it to say UNIX has changed.
I think GNU/Linux may be kinda trying to fork.

GNU/Linux/SysV-like
and
GNU/Linux/Systemd

But the sad truth is it isn't really forking. Due to the tightly bound nature of GNU software, it isn't really able to fork and most distros will have to be pulled away from SysV.

Though Spartrekus, this is *not* a bad thing. I think GNU/Linux has too long muddied up UNIX philosophy and if we can almost see Linux as just another free OS (Like Haiku) rather than a UNIX-like OS then that will allow a true UNIX-like OS to take center stage and do things correctly. Remember, if BSD got their licensing shite together back in the day, Linux and the GNU mentality* might not have bloated out almost every bit of software created.

After all, "GNU is not Unix". It is almost not fair that it is leading the way for typical open-source UNIX-like software. I think it could potentially be doing everyone a disservice because it has very different goals and outlook. For example, all these big online-dependent package managers requiring large infrastructure and "community" is not actually a very UNIX-like philosophy of keeping things small and simple or doing one thing and one thing well.

* For the record, I think the GNU license is extremely important and great. Whereas the sloppy dependency mindset of GNU developers is terrible. As it stands, typical GNU/Gnome style software is almost on par with closed-source software in terms of digital preservation!
 
Status
Not open for further replies.
Back
Top