FreeBSD Foundation Flounders on 15 with Rust, pkgbase, and KDE

Oh, I love a good RTFM response! Let's bring those back because now I might even have a [slight] chance to be on the other end of those. ...So, the issue with those have always been a "the ability to understand what I don't know" catch-22.

But I'm a bit confused. To what level (I know that sounds like a dumb question--and I'll follow it up with a really bad example which I hope you do not read as me being snide--but to what point/line/place/etc do you consider an amateur?). Should we remove the handbook and let them just read the man pages? ...I started using BSD in 2000, and I still read the handbook and man pages constantly (I feel like I'm still an amateur).
if one doesn't understand something after reading the documentation and man pages, maybe he/she should look at the source code first, and then ask questions (I don't do that myself, I ask questions right away, and look at the source code last...)
 
Last edited:
I partly agree that adding KDE to the installer is a bad idea. Not for a technical reason, after all it is always possible to say "no" when the installer asks whether X/KDE should be installed. But for a sociological / philosophical reason: It shows that the FreeBSD community is wanting to put effort into making desktop use easier and smoother, and I think that use case is a waste of time for FreeBSD, as there are many better options available for desktops. It is also the camel's nose under the tent: I suspect that soon, X and a DE will be a requirement for installing and running FreeBSD, probably in the name of streamlining and simplification, as has happened (long ago) to OpenBSD, where one can't run without xbase.tgz. That is not just unnecessary bloat, but a maintenance and security headache.

I have no problem with GPLed software being added to FreeBSD, either in base or elsewhere. While I think that the GPL is a dumb license (and RMS a clueless and wrong-headed person), the world has moved so much towards GPLed software that being dogmatic about the GPL is simply no longer practical. And as long as one doesn't (a) modify GPLed software, and (b) distribute that modified software to others, the GPL has no practical effect.

I'm very much in favor of adding Rust to the base system, and starting to migrate code away from C and onto a modern language. C was a great thing in the 1970s, compared to the alternatives (PL-S, Assembly, Bliss and whatever else we had back then for systems development). But that was 50 years ago, and today C (and even more C++) is a bad programming language, which doesn't make coding efficient and safe.

But the real topic of discussion that is unique to this thread is pkgbase, and I just don't know enough about it yet. I have read the Wiki page for pkgbase, and it is not terribly clear, but it answers a few questions, most importantly: An existing system that's based on bsdinstall / freebsd-update and txz files can be upgraded by first converting it using pkgbasify, so moving from 14 to 15 will not require a reinstall. But how does one get pkgbasify? The link goes to a GitHub page, and I need it either installed for me as part of base, or as a pkg, or shipped with the upgrade system. I don't even have git installed on my FreeBSD machines, nor am I planning to install it.

What is completely missing from that wiki page is a discussion of the "why". For me, as a user of the OS, what does pkgbase do better? How does it help me? Until I know that, why should I use it?

Where is pkgbase in the handbook? Nowhere. That alone disqualifies it from use.

The thing that is seriously disturbing about that wiki page is the complexity and lack of user-friendliness. To quote an old colleague: pkgbase operation (in particular regular upgrades) seems to have glass shards on all surfaces. Today, to go to the next minor version, I do "freebsd-update -R 13.x-RELEASE fetch/install", a 4-step procedure (fetch, then 2x install with one reboot in between), using a simple command and just one extra parameter on the command line (the -R flag). With pkgbase, that becomes a 7-step procedure, and several steps require long commands with multiple options. For major version upgrades it gets outright scary, when the instructions tell you to guess your ABI (why? the computer should know it!) and then the instructions tell you to not allow a step that you are prompted to do. To me, it seems that pkgbase needs a lot of sandpaper and smoothing before it is ready for prime time.

The upgrade to 15 becomes mandatory somewhere around 2028. If the FreeBSD community wants to move to a mandatory pkgbase by that time, it has a lot of work to do first. I have no visibility into whether that work is being done.
 
I partly agree that adding KDE to the installer is a bad idea. Not for a technical reason, after all it is always possible to say "no" when the installer asks whether X/KDE should be installed. But for a sociological / philosophical reason: It shows that the FreeBSD community is wanting to put effort into making desktop use easier and smoother, and I think that use case is a waste of time for FreeBSD, as there are many better options available for desktops.
To hell with "sociological" reasons.

FreeBSD has everything to make for a good desktop. It can be done and should be done. Nothing is being forced unto you, and as long as everything is optional, it's ok.

This overreacting is demoralizing to developers.

It's baffling to me that KDE being optional is such a problem to you but migrating everything to Rust, which is an unreadable mess of a language, isn't.
 
Don't expect that introducing the ability to install a graphical environment will attract new users. Or that it will streamline workflows...
Linux has had graphical environments in its installer for 20 years, and some distributions have installers focused on installing a given graphical environment. It has plenty of bells and whistles and graphical configuration tools, yet it can't reach a certain user threshold. Today's desktop world is mainly Windows and Android (tablets/phones/Chromebooks). Even people who spent a long time on Linux servers and Linux desktops are now switching to Windows or Mac OS on desktops, while still running Linux on servers. I don't know why there's so much argument about introducing a desktop installer. It's a waste of time. In my opinion, this functionality shouldn't be implemented. And don't let anyone tell me that FreeBSD once had sysinstall and offered such an option. I remember those days perfectly, but even then I didn't use it. I installed everything manually. It used to be Gnome2 -> today it's Mate.
I've been a Linux user since 2001, Debian Linux since 2002/2003, and FreeBSD since 2006. Observing the world of open source and desktop solutions on open source systems, this world is standing still. FreeBSD on Desktop users are mostly advanced users, so why introduce the ability to install a graphical environment in the installer? Completely pointless. Everyone will install a different graphical environment than the one in the KDE installer anyway. And few people who use Linux want to mess around with FreeBSD. And there's still "Windows"... Few people who use Windows will fall in love with FreeBSD either. Don't waste your time implementing graphical installer options in the FreeBSD installer. That's not the way to go.
 
Even people who spent a long time on Linux servers and Linux desktops are now switching to Windows or Mac OS, while still running Linux on servers.

I don't know anyone migrating to Windows, it's quite the contrary. Gamers are migrating from Windows to Linux because their Windows games run faster and nobody likes the Copilot spyware.

The only person I know using Windows is my wife because she hears me cursing all day at Unix because I do Unix things.

Wanting to give up the fight to have a better desktop is a losing mentality.
 
That's why I use FreeBSD because it's not Linux. It doesn't have "system-d"(created by Pottering, formerly working for Red Hat/IBM and now working for Microsoft) , but is divided into the main system and external software/packages, and has its own wonderful features, mechanisms, and way of working. And Linux on the desktop hasn't exceeded 2% of the world's desktop systems (for 20 years now).
 
That's why I use FreeBSD because it's not Linux. It doesn't have "system-d" , but is divided into the main system and external software/packages, and has its own wonderful features, mechanisms, and way of working. And Linux on the desktop hasn't exceeded 2% of the world's desktop systems (for 20 years now).
That percentage has doubled and I dunno if it takes ChromeOS into account (because yeah, ChromeOS & Android are Linux inside).
 
PS. I know a lot of Gentoo Linux server administrators who used to have Gentoo on their desktop and now have MacOS. I know a lot of Debian Linux server administrators who now use Windows or MacOS on their desktop. No one use FreeBSD. Only I. I use 3 systems on my desktops: windows11, FreeBBSD and Debian linux
]:>
 
It's the many preconceived ideas not based on anything but thin air. Reminds me of some Teams chats I've witnessed. As a co-worker of mine says, it's a grab the popcorn moment. Sit back and enjoy.
Yep popcorn time.
I think we have 4 more pages before the forum automatically locks the thread, don't we? (kidding here)
What my complaint is that there is too much emphasis or desire to appease the amateurs.
Like anything else in life, sometimes amateurs play at the professional level. Some of the smartest/best people I've worked with have been fresh out of school, book learning but put in the effort to real world learn. At the same time some of the worst people I've worked with are the ones who keep reminding you about their MIT PhD and tell you "I don't have time to read your half page of instructions to set things up. I must have a 3 hour call with so you hold my hand".

My opinion:
I don't care is the installer gets an option to add some kind of graphical environment as part of the install.
I do care that the installer itself stays text based/ncurses. I do not want a graphical installer like a lot of Linux distros.
Pkgbase I have no feelings on it. If it works correctly out of the box and we don't have a long cycle to get it correct (I'm thinking back on the time it took to get 5.x stable/correct) it should be fine. It sounds like there is a chunk of documentation missing (wiki, handbook, etc), especially around an "operation conversion" map: I used to do this with freebsd-update, how do I do it with pkgbase.
 
This thread is indeed very depressing and it is somewhat making me regret to have opened one about FBSD15 and pkgbase. The RTFM attitude I see here is despicable IMHO, I believe that this community is better than that; if people like to behave that way then the Arch Linux forums is their place to be, not this forum.

I see no issue whatsoever in having an OPTIONAL (in caps, bold, italic and underline) installation of KDE in the DVD, it's so obvious that people who don't want it just don't have to select it that I'm 99% sure that who is complaining is just doing it for the sake of it. And BTW, as it was already pointed out before, KDE is the most logical choice in 2025, for localization, accessibility and so on.

I'm still not convinced about pkgbase but people I learned to respect in the past year or so that I've been using FreeBSD don't seem to be so reluctant in adopting it and that's a good sign for me.

I also don't like the idea of rust in base but I'm not a developer and thus I cannot (and I won't) speak about technical merits, I'm just basing my position on what happened in the Linux kernel and I really wouldn't want the same here.

I also had something about systemd but I decided to scratch it as there's no need to add gasoline on the fire.
 
I am concerned but not sure if I need to be. I just don't know if pkg-base bothers me but the KDE thing does.
In some ways, I am the opposite. KDE used to be an option in the old sysinstall so the project is really just treading water like it is the early-2000's again. I'm sure in another decade when we get a new installer, KDE will be removed again.

But PkgBase, if abused (and it will be), opens us up to a loss of our concept of "base" which in many ways defines BSD. It will just turn FreeBSD into a ricers DIY plaything. And I don't think there will be any "undoing" it either. In 2025+ there is zero chance that the open-source community can sit down and make reasonable decisions of what goes into a base install. Instead the only way is to (arbitrarily) just go with SUS and what we were gifted by commercial industries decades ago. We won't get that back.
 
and now look where is the person who built subsystems for linux, system-d, pulseaudio, etc.

"Poettering worked for Red Hat from 2008 to 2022.He then joined Microsoft"
Poettering is the worst example because he always hated the UNIX philosophy. Pulseaudio was replaced by Pipewire anyway.

And much better he's at Microsoft now because not even Fedora uses systemd-networkd lol.
 
I'm sorry to say this but you really don't know what you're talking about here.
I may be wrong, but those are commercial operating systems that pay a lot to developers.

Entitled users that just complain and contribute nothing are a burden in open-source.
 
and contribute nothing are a burden in open-source.
Ironically this whole thread is about "less is more".

People who contribute a bunch of broken ideas (whether it comes with code or not) are in my opinion way more damaging to open-source.

Specifically I believe the project would be better for server use-cases if *less* people contributed home desktop-centric ideas and features.

FreeBSD can only compete in the server (and maybe technical workstation) world. Playing catchup with Windows, macOS, even Linux will never happen in our lifetimes.
 
Windows is spyware. The ex-Linux folks that use MacOS are designers, hipsters and managers that do presentation stuff.

MacOS is actually not bad if you want nothing more than xterms and web browsers.

Also, I can confirm that some gamers go to Linux now. Their games run faster and they hate Windows or at least recent developments thereof. The trigger for this is that Valve pushed Wine into working for a critical mass of games. Without that there would be no chance.
 
Ironically this whole thread is about "less is more".

People who contribute a bunch of broken ideas (whether it comes with code or not) are in my opinion way more damaging to open-source.
Who are you to define what is broken and not?

What is broken about making KDE an option? Hell, I don't even use KDE and now I want to switch.
 
No one has been working on PulseAudio for a long time, so PipeWire has replaced it in Linux. I'm just waiting for SysytemD to be removed from Linux and replaced with another tool. I hate SysytemD and wouldn't want it ever ported and implemented on BSD systems. That would mean the end of the Unix and BSD systems. I didn't like PulseAudio, and I don't like SystemD. I like BSD, especially FreeBSD and its system solutions, and I don't want it to drift into the same Linux development path and repeat the same mistakes.
 
Who are you to define what is broken and not?
Reflecting upon Linux means I don't need to define them. I can just list them based on common reasons why people have migrated away from Linux.

What is broken about making KDE an option? Hell, I don't even use KDE and now I want to switch.
In some ways this very statement is one example why its a broken concept. KDE as an option is now funneling you towards a half-functional DE that won't reflect the quality of FreeBSD and will leave someone feeling that FreeBSD is not polished; when really KDE is not at all part of FreeBSD.
 
No one has been working on PulseAudio for a long time, so PipeWire has replaced it in Linux. I'm just waiting for SysytemD to be removed from Linux and replaced with another tool. I hate SysytemD and wouldn't want it ever ported and implemented on BSD systems. That would mean the end of the Unix and BSD systems. I didn't like PulseAudio, and I don't like SystemD. I like BSD, especially FreeBSD and its system solutions, and I don't want it to drift into the same Linux development path and repeat the same mistakes.
You don't need to wait for anybody. Don't put yourself in a passive position a take an active role and migrate to a distro that doesn't use Systemd because there are many.
 
Reflecting upon Linux means I don't need to define them. I can just list them based on common reasons why people have migrated away from Linux.


In some ways this very statement is one example why its a broken concept. KDE as an option is now funneling you towards a half-functional DE that won't reflect the quality of FreeBSD and will leave someone feeling that FreeBSD is not polished; when really KDE is not at all part of FreeBSD.
KDE is documented in the Handbook and it will be a basic KDE installation.

We had a discussion about containers that made me realize you don't know what you're talking about so I take everything from you with a grain of salt.
 
Back
Top