Funtoo Linux is shutting down
forums.funtoo.org
forums.funtoo.org
https://www.funtoo.org/Wolf_Pack_Philosophy
My understanding is that it was created after the founder of Gentoo had some disagreements about the direction of the distro and forked. Many of the big distinctions (like git instead of rsync for updates; gentoo supports both now) have eroded over time.
I also think we might accomplish more if people could find themselves to being more compromising. Not that we don’t benefit from having a bunch of distros, but perhaps there’s a bit too much segmentation.
> Funtoo started as a philosophy to create a fun community of contributors building something great together. For me, it's no longer that so I need to move on to other things.
It's clear that his intention was to build a community around his vision and it's always sad to see when that doesn't work out.
That's been my experience. YMMV.
Gentoo lets you customize just about everything, but it seems like with Funtoo they take a bit of that control ability out of your hands so you can have a functional system without needing to think so much about the minutia of installation and the like as you do with Gentoo.
That said, Linux is inherently a community project, and compromise is required for any large community. I think the fragmentation is the biggest thing holding back desktop Linux. Not only are talented people solving some of the same problems 30 different ways, but in doing so, ignoring other problem that need solving. This can also be viewed through the lens of user choice and allowing the best ideas to win. The larger issue (imo) is it creates a huge learning curve for new users just to get started. Windows and macOS don't have this problem, because there is a singular vision and the company choose a solution to move forward with that all the employees get behind. All that fragmentation and testing is done internally, and the public doesn't have to deal with it, for the most part.
I'm not convinced.
The main thing holding back desktop Linux is buggy upstream software not working, which usuall affect all distros. LibreOffice Writer didn't have a rastered ruler for 10 years, no quality GPU driver for some AMD card, some KDE app segfaults, some Python app has a value unexpectedly being None, a kernel minor version update just completely broke the version of VirtualBox on Ubuntu 22.04 LTS -- those were and often still are daily problems that no amount of distro hopping or fragmentation consolidation will fix.
The "large" distributions justify their existence by differing in key aspects that cannot be consolidated:
* Corporate vs community, e.g. RedHat/SuSE/Ubuntu vs. Debian
* Fundamental concepts, e.g. NixOS/Guix vs. everything else
* Up-to-date-ness, e.g. Arch vs. Debian
There are some aspects of fragmentation that _could_ be consolidated by better software and more configurability, e.g.
* Source vs. binary, e.g. NixOS provides both
* Choice of libc, e.g. NixOS ships glibc by default but you can easily get much software with musl, and it's technically possible (but needs more community power) to provide a whole system running musl with a single line of config for the end user
* GUI vs CLI configuration
For the small distros, they are niche enough that them merging with another distro doesn't have a huge consolidation effect overall. For example, I do not see how Void Linux and Alpine Linux merging would significantly improve the life of the mass of Ubuntu users, or signficiantly advance "Linux on the desktop".
It's true that there's some amount of duplicate work occurring, but much of that is often easily re-used across distros. For example, if in NixOS something doesn't work with the newest version of musl, one would usually check if Alpine already solved that and if yes, swiftly copy the approach. The main work is for the _first_ one to figure it out, and that work is necessary even if there was only 1 distro.
Reducing distribution fragmentation isn't what makes "Linux on the desktop" happen.
Fixing broken software, making good desktop environments, creating Wine or Proton or Pipewire or cross-platform software (LibreOffice, all browsers, Electron), does.
Due to the fragmentation, I don't think most users make it far enough to run into the application or UX issues. It seems like reducing fragmentation would get more people into the door, and once they are in the door, having those developers work on making the core experience better, is what keeps people around. Once the platform has more users, that will bring in even more developers... the ones who are there as a job, not just enthusiasts, which should lead to an upward spiral for quality.
In terms of the large distributions not being able to be consolidated... I don't think the average user (read: typical Windows and macOS users) care about any of that, other than not having to go to the CLI to change a config. A user should not need to know what glibc is.
In a large open community full of volunteers with no central leader with any real authority, it's going to be a bit of a mess without compromise. Various uncompromising projects might be very good, but the larger sum of all the parts won't fit together as nicely as anyone might want.
I used to be into Gentoo, but the community got pretty toxic around the time that Robbins left. My sense was Robbins was trying to start over with less negativity. Hence the "fun" part of "funtoo".
I believe ChromeOS was based on Gentoo initially. So by that time the community was nice enough for a big tech company to make a financial bet on it.
> Keychain helps you to manage SSH and GPG keys in a convenient and secure manner. It acts as a frontend to ssh-agent and ssh-add, but allows you to easily have one long running ssh-agent process per system, rather than the norm of one ssh-agent per login session.
> This dramatically reduces the number of times you need to enter your passphrase. With keychain, you only need to enter a passphrase once every time your local machine is rebooted.
Out of the nice utilities that funtoo has, keychain is the only one I've seen become available across all distributions, and I use this on every single machine. It's less of a necessity now that systemd can handle gpg-agent and ssh-agent unlocking on login, but I've never been able to figure out how to set that up, keychain is very set-and-forget. It just works.
I disabled the ssh and gpg functionality of it, and use gnome keyring only for programs that need it to store login information.
It then worked as a very configurable, but stable distribution. Mostly current packages, just despite the rolling release model not necessarily bleeding edge, which is a good compromise. And Funtoo kept out systemd, which was great. Even if you like systemd you might respect the thought that monocultures are not a good thing, alternative efforts always need to exist and should be fostered. Like Funtoo.
Flaws for me were the slowness of the package manager when calculating dependencies and that compiling everything is in the end just an unreasonable effort, especially on aging hardware.
Without opinionating on the latter Gentoo dispute- which I didn't follow but superficially from what I saw was sympathetic to Robbins- my sense was that Funtoo was literally just a rebound project. Surprised it lasted this long...but maybe it took this long for all of the spirit that drove Gentoo to reach its natural conclusion. Just factless psychological speculation, but things like that happen when project breaks occur.
That being said, I did think funtoo would be around longer. I totally get the motivation; if a volunteer hobby isn't doing it for you, you ought to be free to let it go. I'm sure that the remaining dev/user base is perfectly free to fork off and rename to maintain the spirit of the distro too.
https://en.wikipedia.org/w/index.php?title=List_of_Gentoo_Li... says it's 16 years old; I think it did just fine.
Good point; it's less a matter of time per se and more of structure and governance; a project with a single person point of failure is likely to have weaker long-term prospects than a project with a higher bus factor even if the latter is younger.
I'd much rather see what a unique or otherwise outstanding distro has to teach me by relying on it thoroughly for a while, even if it's small, and even if it doesn't last a decade. The most established distros are pretty much all the same as each other anyway so it's not like you'll miss much.
Interestingly, both projects actually have a bit of shared history[0].
Funtoo joined Sabayon to become MocaccinoOS[1], before the alliance seemingly fell apart[2] and the two projects went their separate ways.
[0] https://news.ycombinator.com/item?id=33204970
[1] https://web.archive.org/web/20220619205105/https://www.sabay...
[2] https://www.mocaccino.org/blog/2021/12/06/updates-on-funtoo-...
Sabayon shipped with all kinds of over-the-top compositor effects back when those things were new to Linux (and Windows, for that matter). It was also highly styled out-of-the-box, in a somewhat extreme way, a bit like Garuda Linux today. That stuff was a lot of fun and very appealing for me as a young teenager!
At the same time it had a binary package manager whose command line and configuration interfaces were very similar to those of Portage, supporting the same notions of 'masking'. It had very pretty colorized terminal output, and it was extremely fast. It was also compatible with Portage, though in a clunky way because it kept its own package database in a different format— you had to issue a special command to sync the databases of the two package managers so they could both know what all was installed on your system in an accurate way. Later on in the distro's life, using Portage became 'unsupported' in that developers didn't want to help users troubleshoot related issues. But I never stopped using both package managers together, and that never stopped working.
Depending on how far you wanted to go, Sabayon could work nicely as basically a stable, graphical installer for Gentoo (just abandon the binary packages after install and start customizing everything), or you could stick mainly to the binary packages and use the Gentoo repositories (and overlays) much like Arch users use the AUR. (Having run both for years, I'd say using Sabayon this way was definitively better than using Arch and the AUR.)
At the time, the Portage-like masking features made Entropy the most flexible binary package manager I'd ever used when it came to pinning and selecting packages. The combination of binary packaging (which is fast) and source-based packaging (which makes software easy to customize and patch) in Sabayon was amazing despite the clumsiness of using two package managers and convincing them to interoperate. Later (later for me, at least), Nix's transparent binary caching system would blow all that out of the water, of course. But Sabayon was awesome at the time. And it was the gateway to my first stage3 Gentoo install, which was a really productive and memorable experience for me, itself!
Anyway: RIP, Sabayon. RIP, Funtoo. Niche distros like these can be innovative, stylish, and fun. And often their small communities are outstandingly expert, which is a massive, massive plus for newbies who can figure out how to be thoughtful and polite when asking for help.
Always appreciate getting a glimpse into the mentality of the users/maintainers of the more niche Linux distributions out there
Thanks for sharing :)
It's a great distro but suffered from a slow and difficult on-boarding process and the docs seemed often out of date with the actual distro, sometimes years so.
I'd used Gentoo for a long time but struggled with Funtoo to get everything working.
Then, some time more than 10 years ago, perhaps around 2010 or 2011, the Gentoo package system began to have some annoying problems and I have discovered Funtoo at the same time. At that time Funtoo was clearly better, so I have switched to it.
I have used Funtoo for about a decade on a large variety of computers with great success. Then, a few years ago, the package system began to have maintenance problems, like packages that were not updated for long periods, becoming obsolete.
At that time I have switched back to Gentoo, which I am still using.
Due to their great similarity, switching between Gentoo and Funtoo is very easy. Therefore anyone who still uses Funtoo can easily go back to Gentoo with minimal effort.
In any case, I am extremely grateful to Daniel Robbins, the creator of Gentoo and Funtoo, for his work. Like also the work of other great contributors to open-source software, it has saved many years of my life that would have been required to duplicate the results of such work.
https://www.funtoo.org/Support_Matrix
But excluding systemd it wouldn't have been for me. While I am not a fan of all aspects of the project, it's technically superior to the older/simpler systems.
Out of those you listed, I only think readiness notifications are necessary for the existing OpenRC featureset to work more reliably. I'd rather enshrine a different method of doing user services (for example[1]) instead of having openrc require elogind to take over user session tracking, and I'm not a huge fan of socket activation bloating up the amount of running processes the longer the machine runs, outside of the user's control.
Afaik he was hired by Microsoft and didn't have the time to do Gentoo with the new job, so the Gentoo Foundation was setup. When he left Microsoft year or two later he asked to come back but there were disagreements in direction with the Foundation so he created Funtoo instead.
> nor am I interested in trying to find one, or hand the project off to someone else.
Of course if it's all open source anyone can take over the code base. But why saying "not interested to hand it off"?
my guess is that funtoo is not popular enough to have a community from which to recruit a bdfl.
gentoo is not for everybody, but it was a major achievement, and he did it. funtoo was a little sideproject more tailored to his personal interests. there is simply no reason for so many people here to be criticizing him. sheesh.