FreeBSD system and its share of worldwide use - July 2024

[kpedersen]

Is it developers only?

Pretty much. Which is why it is still barable to use. They work on what they find interesting, with only a small number of projects sponsored by the foundation (generally hardware support which is difficult and hardly rewarding in the long run).

Aren't these things related? More users = more developers.

Sadly not. Just more pressure on the existing developers to cater to less technical users.

The less technical / consumer focused FreeBSD projects such as GhostBSD and NomadBSD is possibly more where your interests in evangelism might be more useful?

I think the best is to have a separate build called for example: FreeBSD-14.1-RELEASE-amd64-desktop.img when new users would like to try FreeBSD for the first time

We already do. Just called slightly daft names like GhostBSD and NomadBSD.

Personally I think the main FreeBSD installer should simplify even more. No curses, just text. The process is so much better on OpenBSD when installing to Raspberry Pis, etc, via serial.

 

[bvdw78]

Let's borrow from the Handbook for a bit:

The goals of the FreeBSD Project are to provide software that may be used for any purpose and without strings attached.Many of us have a significant investment in the code (and project) and would certainly not mind a little financial compensation now and then, but we are definitely not prepared to insist on it.We believe that our first and foremost "mission" is to provide code to any and all comers, and for whatever purpose, so that the code gets the widest possible use and provides the widest possible benefit.This is, we believe, one of the most fundamental goals of Free Software and one that we enthusiastically support.

This does not describe any use case at all for the software that the project delivers. So there it is, the project provides "code" and that's where it ends mission-wise. Mind you: it doesn't say anything about delivering an operating system in any way, shape or form at all. If that is indeed the mission, and I assume the Handbook is correct in this matter, then pushing for use cases is something that third parties should do if they want to. End of story, effectively.

 

[Beastie7]

Who's setting goals for the whole FreeBSD project?

The goal is to maintain continuity of the project. There really isn't a specific "goal" per se. The project exists to serve the needs of its' consumers. They also donate to the Foundation on behalf of the Project. Developers have a vested interest in a specific area of the project, so development happens there. Some of these developers are employed by the Project consumers so research, resources, and manpower get diverted according to their needs. It's just most of these consumers are Enterprises and Embedded companies. BSD Unix grew out of that same environment; so historically that is its status quo.

The Foundation likes to spout a lot of marketing fluff about FreeBSD being a viable desktop; but it's simply not true. There's no vested interest (from the project or its consumers) in developing tier-1 like support for desktop I/O. We don't even have a standard API with which application developers can develop against; it's all a hodge hodge of random s**t ported from Linux. We're still chasing Linux for hardware/software support. When we should have our own APIs and frameworks to facilitate application and driver development. (DRM/Mesa3d/Xorg is really.. really old, archaic garbage.). You want to attract more developers? Give them better tools.

SUN and NeXT were the only two companies that solved this problem. You can guess who won. And the majority of the projects committers are using their (the winner) products.

 

[kpedersen]

The Foundation likes to spout a lot of marketing fluff about FreeBSD being a viable desktop; but it's simply not true.

Agree. Not every operating system needs to be tailored to a consumer desktop. The foundation is sending out mixed messages. Not entirely sure why.

And the majority of the projects committers are using their (the winner) products.

For phones, probably; For desktops, possibly; for workstations, unlikely; for servers, no.

Ironically for kernel development, you don't want to be running the unstable kernel on the development machine itself and risk losing everything. Might as well run on a platform that offers decent virtual machine solutions and a decent POSIX environment. Windows and Linux both fail at this, so macOS is not actually a bad choice (if you can accept the criminal online activation).

DRM/Mesa3d/Xorg is really.. really old, archaic garbage.). You want to attract more developers? Give them better tools.

libdrm is pretty new. It came when people decided that userspace graphics drivers weren't cool anymore. I'm sure we will go in circles in another 20 years.

Mesa3d and Xorg are pretty old and have limitations that people aren't even attempting to resolve via Wayland. Luckily most developers don't care about graphics. You want "amazing graphics"? Buy a Playstation it seems (ironically running a modified FreeBSD).

 

[Crivens]

I'm sure we will go in circles in another 20 years.

Xkcd 2044 anyone?

 

[Jose]

If you'll allow me these two assumptions:

1. Programmers mainly use laptops
2. Many Freebsd programmers don't use Freebsd on their laptops

And your goals are to:

1. Attract more programmers
2. Get them to use Freebsd as their daily driver

You'd be far, far better off targeting

1. Wireless driver support and performance
2. Suspend / resume that works flawlessly
3. Battery life on a par with the major operating systems
4. Bluetooth support

Any programmers you actually want in your project would be capable of installing and configuring whatever DE suits their tastes. I suspect forcing a particular DE as the "default" would actually alienate a lot of people, as they would have to go through the trouble of uninstalling it before configuring their laptop to their liking. DEs is an area where there are a lot of divergent personal preferences. Not the case for things like wireless performance.

 

[ralphbsz]

"All Platforms" should also include:

• Server
• Embedded

In which case I am sure we will see FreeBSD being quite a bit higher. Possibly around 5%.

For server, that's not the case; in practice only Linux and some Windows exist. Obviously, there are quite a few Windows servers used in businesses, and cloud providers (in particular Microsoft Azure, duh) have them available. But the vast bulk of servers are Linux, in particular in two groups that use the vast majority of all servers in the world: Supercomputers (where Linux has 100% marketshare of the 500 largest ones), and the cloud super scalers (FAANG). The only large cloud company that uses any FreeBSD is Netflix, and rumor has it (the Netflix HQ is in the same town I live in, and we know people) that they have been replaced by Linux. I would be surprised if even 1% of all servers worldwide run an OS other than Linux and Windows.

For embedded, that's a really tough question, because many embedded systems do not interact with computer networks, and if they do, only in a limited fashion. For example, my wall thermostat and my WiFi AP all run Linux; the garden sprinkler controller runs Android. But the AP never sends any packets to an outside network, except to its own manufacturer to get firmware updates. And the thermostat and sprinkler speak to their cloud service, and to nobody else (yes, I monitor that occasionally). It is possible that my dishwasher or refrigerator run FreeBSD internally, but how would I ever find out? They have no screen, the only output device is a few indicator LEDs, the only input device a few buttons. It could be that FreeBSD has a 5% market share in this type of application, but why does that matter to the typical user here on the forum, who uses FreeBSD either as a server or as a desktop?

Aren't these things related? More users = more developers.

On the contrary. The four most commonly used OSes for computers with user interfaces are (in that order) Android, iOS, Windows and MacOS. While each of them has a very large number of developers (I bet each of them has many thousand coders), their users are completely not involved in developing. They can't even file bug reports in the usual sense of the word (where you get feedback about the bug), nor can they see OS-level documentation (except for Android, which is partly open source, but doesn't accept much input from outside).

In the specific case of FreeBSD (or Open- or NetBSD): If these OSes got a million new serious desktop users, I think what would happen would be the opposite of having developers have more time for doing development. I think the clueless segment of the user population would clog channels such as PRs and developer mailing lists. Helpful developers (BTDT) would waste their time on helping users; the rest would probably go into isolation or leave the project.

If you want to distribute software to customers or consumers, you have a choice. If you want to provide technical support, you need to create a technical support organization and infrastructure which is capable of handling the workload, including the fact that half of the users have an IQ that's below the median for the group. This is the model used by expensive commercial software: People who spend 5- and 6-digit amounts on software expect their phone calls and e-mails to be answered, and if the question is difficult, they expect the person responding to have a doctoral degree. This model sort of works for cheap consumer software: If you have problems with iOS or MacOS you can walk into an Apple store and get limited help, Intuit and Microsoft at least theoretically have support systems (mostly web-based and automated, and usually of not much help). For free software, that model doesn't work, because the money just isn't there (duh). The other support model is to offer no support. That means support questions get moved to a different place, for example this forum. It also implies that some (charitable and laudable) developers help out with support. But speaking as the manager of these developers: they do have more important things to do. This is the system free Linux offerings use (if you buy RedHat or SUSE, you actually get support, that's where most of your money goes).

The vast majority of users of free software do not turn into developers, nor do they help with developing. There are some exceptions. The first one is large companies. For example, if we could convince IBM to use FreeBSD with a DE of their choice to be the standard OS for the laptops distributed to their employees, then IBM would build a large department to support its internal users, and would put a group of internal developers next to it to improve the software. This exists today (and did even before the RedHat acquisition), and IBM calls it the LTC, or Linux Technology Center. It has thousands of employees, with hundreds or thousands of developers (including dozens or more kernel developers). Together with companies such as Intel and Google, IBM is one of the largest contributors to Linux (both in terms of code and in terms of $$$ to the various foundations). The problem is: How to convince a large company to use FreeBSD in that manner?

The second exception is that a very small fraction of users are actually capable of helping with development tasks. One example: I used to work in research and advanced development, partly on a software product that is sold for 6-digit amounts. Occasionally, I would help with difficult support problems. Part of the software product was distributed as scripts, and a part was open source, so customers had access to a small fraction of the source. We had one customer who would occasionally file bug reports that included patches for the source code they could access, or scripts that were canned work-arounds for bugs or misfeatures. But this kind of "user turned into developer" is exceedingly rare. If you look at people who use Linux as a desktop OS, I would guess that 99.9% are not skilled C programmers.

 

[zsolt]

At the end of the day I think the question is: How to raise those numbers up? (FreeBSD: 0,01%) which by the way are not even rooky numbers If everything stays the same then those numbers may decrease the next year. What will be the situation in ten or twenty years? Will FreeBSD be present in the OS market share? We are also beat by Chrome OS: 1,41%. WFT! Probably we will never conquer the world ever but I think some Linux developers must converted to our side.

 

[sidetone]

I've noticed that a lot of it is, do it yourself, with less and less collaboration, especially when it comes to porting or maintaining ports. There's less and less help for those willing to learn how to maintain ports, than there was in 2016. I understand, RTFM, though porting with the Handbook and with learning GIT is not easy. It takes a couple of times to read over sections, and sometimes doing additional searching. Even with knowing enough, it's still not easy for everyone. While it may be our responsibility, I think there need to be more mentors, because it's quicker to read it first, then ask for when there's specific needs on porting. The purpose is to get something up and running.

When I learn, I share my knowledge to others often in a more clear way than how I learned it. It helps me retain, and helps me with my understanding. If I explain something and miss something I thought I knew, I can realize it when I have to put words together to explain it. When I wrote faq's and howtos, that's one example of me reiterating what I've learned which took a longer time to do from varying sources. Some people act mad that they had to learn something, and they want you to learn something the hard way too. I read the manual a few times over, by the way, and some stuff isn't explained. I want to get something done, and don't want to research it for a week, when I have other things to do, and I'm not a dedicated maintainer. I just learn something, even if it's the hard way, and convey it in a better way for others.

I've learned enough to port and maintain the most basic types of ports, though they may as well drop me from maintaining what I have. I still needed help on those, then. When someone asks, I hope they have the GIT homework, bc I haven't been able to for a while. As, before, when I was set up to do it, I could do it based on their suggestion.

We're lucky when someone is nice enough, knowledgeable enough, has the resources and the time to maintain or port something.

The NetBSD community is overly eager to help on their system with what they understand. They don't understand some issues, such as Raspberry Pi's not having control to shut off.

We do still get a lot of good help from those who port major programs which depend on Python and such. It came slow for a while, because it was a massive project, which had to be fixed first.

Though, at some point, FreeBSD does need to step up, to that the desktop developers use FreeBSD as their primary OS, and that means catching up on features, or using a Raspberry Pi for another OS to perform those tasks which FreeBSD can't.

 

[kpedersen]

For server, that's not the case; in practice only Linux and some Windows exist.

You could say the same for desktop usage, and yet FreeBSD scores 0.1% in the above data.
My logic is that unless you are suggesting that FreeBSD has a bigger penetration on desktops than it does servers, I am fairly certain we would see more representation in the server space. Especially seeing a moderate recession of Windows in that environment.

My estimates are that FreeBSD is used for servers ~50x more than as a client OS. So I crafted my above figures accordingly

 

[cracauer@]

If you'll allow me these two assumptions:

1. Programmers mainly use laptops
2. Many Freebsd programmers don't use Freebsd on their laptops

I have several laptops with FreeBSD, but it is true that my main laptop is a pair of 16" Macbooks.

That is not just an OS thing. The Mac is just the much better laptop than PC laptops:
- better battery life on the Apple Silicon macs
- better speakers (important to me)
- better microphone array
- better, brighter display
- much better trackpad (also important to me)
- can buy one an no surprises about hardware. Very difficult to do in PC land since the turnover of new models is so fast

There also is the aspect that some laptop problems with FreeBSD are hard, too hard for me to do anything about myself. Better wifi? Huge stack of many things to implement. Better battery life due to CPU power states? Maybe doable. Use the idiot dual-mode webcam (with infrared alternative) on a Thinkpad? Not even Linux supports it. Microphone input on Tiger Lake? Again a huge framework to implement, it is not just a straightforward driver.

ETA: even Windows users can have bad hardware surprises when buying laptops. Some idiotic chip that only ever gets amateur level drivers - nothing you can do.

 

[6502]

If you'll allow me these two assumptions:

1. Programmers mainly use laptops

Wrong assumption.

 

[Crivens]

I think there is a sweet spot well below the point occupied with windows and Linux. We may have smaller manpower, but we are not meandering like Linux or rebuilding half the shed every two years like windows. We may not be on the cutting edge for a lot of stuff, but we most likely are a lot more efficient than those two. And that makes us more stable, predictable, fewer kinks in your learning curve. Knowledge stays relevant for much longer.

As for hardware support, I use older ThinkPads. They work pretty decent, can't complain about battery life and suspend resume (since I learned to shut up the windows fast start mode, bloody fools...)

Hardware has the same problem software has, for the last decades. They try to do things differently, but not necessarily better. That means new interface methods, changes to data flow, ... All that makes it more difficult to keep changes small, not changing the design too much. That playing around costs too much time.

 

[Cath O'Deray]

… setting goals for the FreeBSD project?
Is it developers only?

See below.

The core team only?

Yes. Please see FreeBSD Project Administration and Management.

What's the role of the FreeBSD foundation in that process?

None. Please see About the Foundation.

Are these goals always about server usage, or desktop is also taking portion of it?

Neither.

… the Handbook …

It's outdated.

<https://forums.freebsd.org/posts/629088>

 

[BaronBS]

even Windows users can have bad hardware surprises when buying laptops. Some idiotic chip that only ever gets amateur level drivers - nothing you can do

That's why I only buy refurbished office/corporate laptops, like Thinkpads and Dell's. It is like a beta test for hardware. I know that they worked for a long time and if I buy it right I can even get a 1 year warrant from whom refurbished. So f.. apple with their overpriced hardware with a shit unix. I'm gonna use secondhand laptops for the rest of my life, or until risc-v get usable, what comes first.

As for hardware support, I use older ThinkPads. They work pretty decent, can't complain about battery life and suspend resume (since I learned to shut up the windows fast start mode, bloody fools...)

I know, nothing is going to be like T420/X220 era, but the 8gen latitude from Dell are pretty solid.

 

[Cath O'Deray]

general.useragent.override = Mozilla/5.0 (X11; FreeBSD x86_64; rv:129.0) Gecko/20100101 Firefox/129.0

Now let's test it again, better!

Not necessarily better.

If the agent is identifiably on FreeBSD, then there's a greater risk of some sites treating the agent as peculiar, sometimes with negative consequences.

 

[ralphbsz]

At the end of the day I think the question is: How to raise those numbers up? (FreeBSD: 0,01%) which by the way are not even rooky numbers

You missed the even more fundamental question: Do we want to raise those numbers? Does it benefit us if many other people use the same OS as we do?

Below that question is yet another question: What does "we" or "us" in the above line even mean?

My personal answer: I like using FreeBSD for a server. The biggest single reason for that is ZFS; since I have an extensive storage background, I insist on using a storage system that integrates RAID with the file system, uses checksums, and is written by skilled professionals, not drunk college students. The second reason is that FreeBSD is designed more like a cathedral, and less heaped up like a bazaar (I really liked the article by Poul-Henning Kamp linked here recently): I mostly see intelligent people making design decisions, and following good development procedures, not an obnoxious dictator with knee-jerk reactions plus a large army of uncontrolled coders. So I wonder: how will having millions of desktop users help make FreeBSD a better system for me? I think it will not.

Disclaimer: YMMV. I accept that, which is why I'm not trying to stop the freight train, just point out that I don't like its direction.

 

[scottro]

My answer to ralphbsz is a firm, definitive, Err, maybe? There are advantages to popularity, mostly support from hardware and software producers, having more incentive to support FreeBSD for various purposes, e.g., wireless, as they have with Linux. The disadvantages are that it might become more like the GUI linux distributions, all point and drool, which I don't mind, until they start making it difficult to work without the GUI tools. And I don't want that, part of why I, and I'm sure, many others like FreeBSD is because of its usefulness without a GUI. (Hey, today I learned, I think on these forums, that mpv can play a video in console, which I'd never known.)

 

[Beastie7]

better battery life on the Apple Silicon macs

I think a lot of people underestimate how much of a game changer Apple Silicon is. So much that other players in the industry are scrambling to develop and ship out alternatives. The performance and power-per-watt difference between Apple Silicon and a comparable x86 chip is ridiculous. I don't think the open source community is going to be able to keep up with that disparity. Jordan Hubbard tried to help with modernizing the FreeBSD kernel (for hardware like Apple Silicon) but the community was too stubborn (or lazy?) to make use of his proposals. I don't blame him for leaving.

 

[ralphbsz]

There are advantages to popularity, mostly support from hardware and software producers, having more incentive to support FreeBSD for various purposes, e.g., wireless, as they have with Linux.

Several potential answers. First, the area where support from hardware vendors is most needed is the whole graphics and UI area. Which I am personally not interested in. And as Linux shows, support there is spotty too. For an independent vendor of graphics card, mice etc. there is really only one OS that needs supporting: Windows (as Apple builds its own hardware and does its own firmware support). The next largest one, Linux, is already one to two orders of magnitude down in usage, and the lack of support shows. FreeBSD is another two orders of magnitude, so there is even less incentive there.

On wireless, my impression is that the problem is mostly not vendor support, but technical debt of work within FreeBSD, which is now being addressed with foundation funding. And better wireless support is not only useful for laptops, but also for servers (as an AP) and for embedded. So I whole-heartedly support that.

The disadvantages are that it might become more like the GUI linux distributions, all point and drool, ...

And that is exactly my fear. For example, a graphical installer is likely to become a disaster for machines that can't run GUIs, as it will probably suck support from the text-based installer, and there is at some point a zero-sum game in developer effort. And a graphical installer is simply not needed (even if the resulting install is intended to be a desktop). And this is just one example; many others can be constructed.

Interestingly, Linux is a very good counter-example: In spite of it having a strong focus in its developer community on being a desktop system, and in spite of most Linux developers using it as the daily driver, it has also remained a superb server and embedded OS. Part of the reason is that in servers it is the only game in town except for specialized uses, and in embedded it has by far the best price for usefulness, so there is lots of corporate support for making sure the FAANGs of the world can run their millions of servers well on it. FreeBSD does not have that support behind its server usage.

 

[Jose]

That is not just an OS thing. The Mac is just the much better laptop than PC laptops:

I don't think we can expect the Freebsd foundation with its limited resources to come up with new hardware I agree that Apple hardware is really sweet. I just don't think it's worth the money anymore. Will someone else step in and sell comparable hardware? I can dream.

There also is the aspect that some laptop problems with FreeBSD are hard, too hard for me to do anything about myself. Better wifi? Huge stack of many things to implement. Better battery life due to CPU power states? Maybe doable. Use the idiot dual-mode webcam (with infrared alternative) on a Thinkpad? Not even Linux supports it. Microphone input on Tiger Lake? Again a huge framework to implement, it is not just a straightforward driver.

I did not mean to imply that Freebsd developers in general or any Freebsd developer in particular should be working on anything or should be to blame for anything. I was hoping to articulate reasonable goals for the Freebsd project. I suspect that if enough programmers used Freebsd as a daily driver all the things we think are "good" like a good desktop experience, etc., would follow.

ETA: even Windows users can have bad hardware surprises when buying laptops. Some idiotic chip that only ever gets amateur level drivers - nothing you can do.

This is so true. I still have a couple of Windows machines around, and that's no picnic. Getting worse over time, too.

 

[teo]

FreeBSD is a "base" operating system. It has a bunch of DEs and WMs like Mate, KDE, CDE, Xfce, twm, and much more.

And which DEs/WMs are wanted is depending on users.
If defaulting to single DE, users who want it would attracted, but others would silently go away.
Let other projects like GhostBSD or something to do so, and keep FreeBSD itself as a "base" looks good to me. If any of GUI variants attracts a huge amaout of users, some of them are happy to pay for it, and the fact attracts commercial developers to earn by "support business" and feedbacks a bunch of really good codes as BSD-compatible license, FreeBSD can get the advantages as their base.
There can be per-GUI (DEs/WMs) variants, theoretically.

Most important things for FreeBSD itself would be up-to-date hardware supports, stability, safety and speed. Of course, hardware support should include not only data-center grade but also cutting edge notebooks and gaming PCs (as usually has most computing [CPU/GPU] power within consumer grade hardwares).

The FreeBSD project has a long history, over 30 years as the default base system for servers, yet its usability and marketability to the world over the years has drastically declined to the point of jeopardizing its very existence as an operating system. Over those years, it has been shown that none of its derivatives have shone as an attractive desktop environment system to gain users in the world which is a loss for FreeBSD in every way, and the few new users who have tried and liked FreeBSD (in the absence of a default graphical environment system) prefer to fight with the machine trying to install a decent 100% FreeBSD based synchronized and updated system, and find themselves with obsolete software due to lack of desktop environment maintenance or lack of software/hardware and end up moving to another system.

Hopefully the foundation will listen to the suggestions of many users who over 30 years have suggested a FreeBSD system with a graphical desktop environment by default.

 

[T-Aoki]

The FreeBSD project has a long history, over 30 years as the default base system for servers, yet its usability and marketability to the world over the years has declined drastically to the point of shaking its very existence as an operating system. Over those years, it has been shown that none of its derivatives have shone as an attractive desktop environment system to gain users in the world which is a loss for FreeBSD in every way, and the few new users who have tried and liked FreeBSD (in the absence of a default graphical environment system) prefer to fight with the machine trying to install a decent 100% FreeBSD based synchronized and updated system, and run into obsolete software due to lack of maintenance for the desktop environment or lack of software/hardware and end up moving on to another system.

Hopefully the foundation will listen to the suggestions of many users who over 30 years have suggested a FreeBSD system with a graphical desktop environment by default.

Think about Linux world.
Linux is just a kernel part of the distributions. And every distros choose which version of Linux to use, which userland (usually Gnu libraries and core utils) to use, choose whether based on one of other distros or go their own way including package manager, choose CUI or GUI, if GUI, which DE or WM to use, and finally, what apps to bundle by default.

With this synonims, FreeBSD is the first 2 things plus ports/pkg infrastructure. Does Linus himself providing any distro? What about Gnu project? None. What you are saying is equivalent to require Linus to create his own distro, which ALL OTHER DISTROS SHALL BASED ON.

What you're stating is on distros, like GhostBSD, PC-BSD or something else.

Maybe most succeeding distro based on FreeBSD would be PlayStation4 and later.

And (possibly discontinued) interesting effort was Debian/kFreeBSD, which used FreeBSD kernel instead of Linux and (maybe specially crafted) Debian userland. I love genuine FreeBSD better, though.

 

[Cath O'Deray]

… Jordan Hubbard tried to help with modernizing the FreeBSD kernel (for hardware like Apple Silicon) but the community was too stubborn (or lazy?) to make use of his proposals. I don't blame him for leaving.

What's the supposed reason for his 2001 or 2002 departure?

The Project's Contributors to FreeBSD page lists him as having a commit bit, which contradicts the access files for the three main trees (there was a relevant change in 2017); his Wikipedia page has a few bad links; and so on.

In June 2001 he became Engineering Manager of the BSD Group at Apple, Inc.

I see a copy of his April 2002 resignation from FreeBSD core team (whether it was published with his permission, I don't know) … "The first and certainly foremost reason is a lack of time and energy.".

In May 2002 he wrote publicly (with his FreeBSD ID):

… I think that if FreeBSD is to succeed in the much longer-term, it needs to transcend its "old guard" beginnings and transition to being governed by a new and entirely more youthful set of folks, it perhaps being a maxim for the future that if you've been in core before then you don't qualify, no hard feelings or offense meant.

Of course, I also expect the old guard to fight this tooth and nail with lots of dire predictions as to what will happen if anyone under 30 gets into core or why only a long and hoary track-record truly qualifies one for the job. As an old guardsman myself, I say ``bah!'' to that sort of self-serving "logic" since we've already proven over the last two rounds that experienced people are more than capable of making mistakes or being blind to all the relevant signs and portents when things are going off the rails. Even if it came with a penalty of inexperience and beginner's mistakes, I wouldn't mind the fresh energy and enthusiasm that a totally new slate would bring. It's not like those of us in 1992 were entirely blessed with loads of experience at doing this when we started and look what we managed to create.

– and:

… given that this is now a democracy, I guess that's all for the committers to decide. I've merely spoken my piece regarding what I hope they'll choose and what I'll certainly be biased in favor of when presented with the next slate for election.

I'm now listening to the 2022 fireside chat …

 

[kpedersen]

The performance and power-per-watt difference between Apple Silicon and a comparable x86 chip is ridiculous. I don't think the open source community is going to be able to keep up with that disparity.

I don't think we need to keep up. Since Ventura, macOS users are required to activate online via Apple's DRM servers in order to reinstall the OS. Not only does that make it a ticking time bomb, but also makes it unsuitable for isolated networks.

Basically FreeBSD will win by default. That can be installed on a mac hardware even once Apple's servers stop answering activation requests for a given EOL device.

In other words, in 20 years, FreeBSD (and other FOSS) will be the only choice to install on the current mac you own now.