Re: The report of my death was an exaggeration

[ralphbsz]

Mark Phillips just posted an opinion piece in the top section of the forum. Since I can't reply to it there, let me do it here.

First, he claims that FreeBSD is doing well (meaning: not dying), because the number of Google searches for it has a slight uptrend, while the number of Google searches for Linux is constant. So let's look at the data, which he clearly got from Google Trends. Here is the number of searches if you make the search window 5 years:

Neither graph shows a trend, unless one carefully selects a time period. Any notion that the number of searches goes up or down by more than a few percent a year is not supported by the data. But what he omits in there is that the number of searches for the two terms is RADICALLY different. For fun, let's put them on the same graph:

The red line is searches for "linux", the blue line for "freebsd". Even if the searches for FreeBSD went up by a few percent per year, it would take a century to catch up to Linux. So his attempt at using data to make an argument falls flat.

Next, he claims that Linux has succeeded in the market place because the word "Linux" has become synonymous with "operating system", because of the "Kleenex" effect (he uses the examples Hoover, Google, and Zoom for products where the brand name has become a description of the product). He conveniently ignores the fact that Hoover has ceased to exist (it's today only a brand name used in a part of the world by a Chinese manufacturer, with a small market share), and that other cases of the effect ended up self-defeating (Kodak is the textbook example).

But even worse: Linux is exclusively used by nearly all supercomputers (100% of the top500 list), nearly all hyperscalers, and the bulk of servers in the world. Claiming that this dominance is caused by a brand name effect, and not by technical considerations, is claiming that people who spend somewhere between tens of millions and tens of billions on computers make decisions based on emotions and familiarity with brand names. That seems far fetched, and is quite insulting to decision makers. If someone told the folks who just bought El Capitan, or who run tens of millions of FAANG servers on Linux (and few or none on FreeBSD) that they ignored technical merit, and spent $600M or $1B/year based on brand name familiarity only, they would get very mad, justifiably so.

But then he (correctly) points out that there are some companies actually using FreeBSD, and points to a Register (!) interview that talks about FreeBSD actually being used, and pointing at the usual success stories: NetApp, Netflix, ignoring conveniently that a large fraction of Netflix' machines actually run Linux, and FreeBSD is only used there for the streaming servers. Well great, so why are Google and Amazon and Baidu not using FreeBSD? Each of them is 100x larger than Netflix?

Finally, he puts his fingers on what he thinks the reason is that nobody talks about the great successes of FreeBSD: It's the license! Under the BSD license, you can modify the source, and don't have to contribute it back. True, but wholly irrelevant to the question of FreeBSD's mind share. Companies who use or modify FreeBSD (and equally Linux) are not required to be "talking about it" because of either license.

Then he claims that under the GPL you have to share derivative works. As a generalization, that is untrue: If you ship a product that employs a modified version of Linux, you also have to share the modifications you made to Linux, to the customers or users of that product. But you don't have to contribute the modifications back (although you can't stop your users/customers from doing so). And if you use Linux in-house, you don't have to share the modifications at all. And the really big users of Linux I mentioned above use the OS in-house. So no, many derivative works do not have to be shared.

There is an even more fundamental question though. Mark Phillips seems to be advocating that the BSD license is what is holding FreeBSD back. I find that quite an astonishing viewpoint from someone representing the Foundation.

In my (not at all humble opinion), that opinion piece is all around very bad.

 

[tanis]

From my point of view, there is the lack of widely adoption in all use cases, what is holding back FreeBSD. More users are required, who are deploying FreeBSD in real life scenarios, which would in turn increase development. What is the USP of FreeBSD ? Why should I choose FreeBSD over any other OS ? There are more and more software packages like SLURM which move in a more linux feature centric direction, hence there is no one using FreeBSD in that context, FreeBSD falls further behind. I'm pretty sure there are other examples as well.

Regarding the license "advantage", I consider the license aspect last in a long list I have to work through, "license" does matter, but way less then anyone believes. I would recommend to drop the argument about the license and look for the real value besides the "license". If the license is the only thing left FreeBSD got going for it, it is indeed dead.

 

[kpedersen]

Good rebuttal. As a FreeBSD fan, I am going to put my fingers in my ears and go with Mark's more optimistic analysis but you certainly aren't wrong. Some points:

1) FreeBSD doesn't look like its dying but...
2) We are so far off Linux.
3) Internal usage by companies is a key part of the market and license does not affect any of this.

4) I do believe brand name may play a part more in education and juniors who *bring* it with them as they take more senior positions within their career. One of Apple's best strategies to penetrate the market was to advertise to students who very much did take the brand (and their experience with it) rapidly into the market with them. I am not quite so sure this is what is happening with Linux however. It has had to fight tooth and nail to compete against commercial UNIX (and Windows) so in many ways is very deserving.

 

[ShelLuser]

But even worse: Linux is exclusively used by nearly all supercomputers (100% of the top500 list), nearly all hyperscalers, and the bulk of servers in the world. Claiming that this dominance is caused by a brand name effect, and not by technical considerations, is claiming that people who spend somewhere between tens of millions and tens of billions on computers make decisions based on emotions and familiarity with brand names.

You'd be surprised, because that's definitely a thing.

I've been there... Unix? Isn't that this ancient OS that was awesome in the past and has been surpassed by Windows and MacOS? (Free)BSD? What's that?

Linux? Oh, right! That's the free (!) unix thingie that's now supported by Microsoft. Yah, I can run that straight from my console app.

Guess which Unix-like solution you're most likely to sell to the higher ups?

Technical considerations? There's also the issue of how those get presented to the higher ups, plenty of managers in tech companies actually have 0 to none affinity with that themselves. So... it also relies on how well the tech staff will explain and relay all their information. And let's not forget: it's the managers ass on the line, so once again... would they put their trust into some weird thingamagick which is claimed to run better, or would they go for the stuff they already know?

Of course... we're talking exclusives here. You're directing one, I the other. But that doesn't make either argument invalid; context applies. And yes, I've been there in several situations where I could "sell" Linux, but a better FreeBSD solution was immediately shot down.

That seems far fetched, and is quite insulting to decision makers. If someone told the folks who just bought El Capitan, or who run tens of millions of FAANG servers on Linux (and few or none on FreeBSD) that they ignored technical merit, and spent $600M or $1B/year based on brand name familiarity only, they would get very mad, justifiably so.

Just because dozens of good decision makers exist doesn't imply that they're all good.

Heck... I'll do you one more: why insist on running a more expensive Windows server environment instead of using Linux and/or FreeBSD? Exact same story most of the time: Windows is what people used at home, so "obviously" that would be the better tool for the job. Linux? naaah.

I can't comment on the rest because I didn't read the article, but this detail stood out for me.

 

[Maturin]

1) FreeBSD doesn't look like its dying

Just by my personal stomach feeling (so very reliably scientific indeed) my impression is, FreeBSD is growing.

In above statistics I imply three critical points not to take it too seriously:
1. Don't overestimate Google's statistics. The amount of searches on Google is not the same as the number of users using FreeBSD; it may correlate, but it's for sure not the same.
Plus: as far as I know FreeBSD's users typical search engine is duckduckgo.com. I know it's a metaengine, but I wouldn't be surprised if this may the cause or anyways Google's numbers do not completely representing the whole world's absolute truth.

2. You said it yourself:
The curve is tight. Too quickly things are interpreted into shapes which will later reveal as nothing but normal statistical fluctuation and noise. If you had just seen the dropping part before "Note Nov 21, 2021" only, extrapolated from that, only, you had to get to the conclusion: FreeBSD will be completely dead app. March 2022
And if you extrapolated the few monthes after that point your conclusion had been: Shortly there will be only FreeBSD - Windows, Apple, Linux, all dead.
By Google search requests.

3. You said it yourself:
Because of the license many FreeBSD usage is not known as it, which doesn't mean it's less usage but just it's simply not known.

The seducing trap of statistics are the numbers are often not seen as what they actually are, only, but interpreted as what they might represent. Since any one can interpret anything in everything all bets are off.

Referring to my first line:
Ask the forum's admins if the numbers of users, new-registrations, or better the foundation if the amount of donations are dropping, rising, or are stable.
At least I would give more on such numbers to represent the status of FreeBSD than on the amount of Google searches.

 

[cracauer@]

ralphbsz I share your concerns about this writeup. Well laid out.

And the tendency to substitute the designation of a thing by a specific brand name is mostly a US thing. Even if the argument was true (which I don't think it is) then it wouldn't explain what is happening in the rest of the world.

 

[freethread]

I don't know if I will be able to express my thoughts in an understandable way (in English)...

If you look at the rough data of the market, Windows is at the top of sales, followed by macOS and finally Linux. The market is not only made up of super computer experts who tinker with corporate and commercial data networks, clouds and server farms. Most users are gamers (video games), ordinary people who save photos from their phones, spend a week on Office/LibreOffice/OpenOffice just to write a resume, visit all the most useless sites on the internet, buy/sell stuff and millions of other things of this type.

Windows (DOS) made its fortune because it cost relatively little (hw & sw), Macintosh cost double but was still accessible, for Linux it took a little longer at least for the users mentioned above, but on the server side it had a huge success (because it was cheap and it was UNIX), now it is also gaining on the desktop market since the installation is now simple and the DEs are largely easy to use, after all even on this forum many talk about similarities between Windows and Linux.

Now before concluding we should think about what is meant specifically by the market of an operating system when there are at least three major types of their use:

1. server
2. embedded
3. desktop/laptop

If we take into consideration servers then ousting Linux could be a problem. I would leave embedded alone because I think it is a world apart and in any case even there, at least on PLCs, Linux has a good share. From the desktop/laptop point of view FreeBSD could gain points, but it could compete if the installation with a graphical environment was at least on par with Linux (which has less market than Windows and macOS). I know that many here do not like it (and to tell the truth I am also fine with it as it is) but if the intent is to gain market then you have to rely on the consumer market users (above). Easy installation, graphical environment, applications that are more or less useless and futile. Then there is the work of updating the system and applications (ports/packages) that should be transparent to the end user. If these goals are not achieved FreeBSD will remain a niche operating system, from a technical point of view (my personal point of view) FreeBSD is the best because it can be used at best as a server, as a desktop and also in embedded applications (at least on hardware where it is possible to compile it) but those who use it must have the knowledge and experience to use it at best. Installing a WM or DE is not easy for users who are not familiar with UNIX and the command line (as of today I also think many Linux users, the youngest, have troubles to install FreeBSD because Linux is changed a lot) and... some traps that hardware can reserve, this means that you have to tweak some settings and most of them, including system services and drivers, need to be set and tweaked manually. These things a consumer user does not accept.

 

[ralphbsz]

From my point of view, there is the lack of widely adoption in all use cases, what is holding back FreeBSD.

But wide adoption does not mean lots of "script kiddies" or "distro hoppers" running it. Those tend to not do anything positive for the OS, they just use it, and the clog the volunteer support channels (such as this forum, or Reddit and Discord) with questions.

More users are required, who are deploying FreeBSD in real life scenarios, which would in turn increase development.

Yes, but: the kind of users that turn into development are either amateur software developers (typically software engineers / computer scientists looking for a hobby), or big corporations who can fund developers. Remember, companies like Intel or IBM employe hundreds of dedicated Linux developers.

What is the USP of FreeBSD ? Why should I choose FreeBSD over any other OS ?

As you said, that depends on context. For my use case (as an amateur running a few servers, at home and in the cloud), the Unique Selling Proposition is that FreeBSD is well engineered (not slapped together with little oversight and little coherent design), strives for high quality and stability, and has good integration with the best freely available storage backend (ZFS). From that viewpoint, FreeBSD isn't dead at all. I completely understand that other users have different requirements; that's why 99% of all servers in the world don't run FreeBSD.

 

[ralphbsz]

Heck... I'll do you one more: why insist on running a more expensive Windows server environment instead of using Linux and/or FreeBSD? Exact same story most of the time: Windows is what people used at home, so "obviously" that would be the better tool for the job. Linux? naaah.

That may happen in smaller companies, where technical leadership is not focused on computers. The classic example is an SMB (small or medium business) that needs a few servers and desktops, for things like order entry and fulfillment, e-mail, billing, accounts receivable. But that is actually a surprisingly small part of the computer market. The bulk of all computers (and I count cell phones and tablets in there) are either servers, with most of them in use by the FAANG and hyper scalers, and single-user machines (the laptop at home, the desktop in the office, the cell phone in the pocket).

 

[ralphbsz]

If you look at the rough data of the market, Windows is at the top of sales, followed by macOS and finally Linux. ... Most users are gamers (video games), ordinary people who save photos from their phones, spend a week on Office/LibreOffice/OpenOffice just to write a resume, visit all the most useless sites on the internet, buy/sell stuff and millions of other things of this type.

That is the market for personal machines, the ones that humans interact with via screen and keyboard. Including gamers and home users, as you said. Today, this is the smaller part of the computer market, less than half. The bigger ones are servers:

The market is not only made up of super computer experts who tinker with corporate and commercial data networks, clouds and server farms.

The larger half of all computer in the world are there. The hyper-scalers (a.k.a. FAANG, although that one misses for example Microsoft/Azure, Baidu and TenCent) alone account for the majority of those,.

In neither the desktop/laptop/personal/portable market nor the server market does FreeBSD have a significant presence. And ultimately, that's perfectly fine with me: I'd rather have FreeBSD be a niche product, used by people who value certain of its characteristics.

 

[Espionage724]

Because of the license many FreeBSD usage is not known as it, which doesn't mean it's less usage but just it's simply not known.

Not directly related, but I was surprised to see so many BSD mentions on my modem's config page earlier:

___

I'm into performance. Windows does that good with games, but I have concerns about Linux and Wayland's push that makes it not future-appealing if I'm serious about game performance.

Xorg support on FreeBSD is what got me in; it works, I liked what I saw of FreeBSD overall, and recently switched my homelab to it (from Windows, after years of Linux and questioning desktop-side recently). I like that FreeBSD maintains their Xorg support, and I'd like to eventually contribute code to both in-support! (I'm interested in C/C++ but have to learn it )

Why should I choose FreeBSD over any other OS ?

If someone asked me that, I'd tell them it's faster

Compiling a game server was 40 mins Windows, 35 mins Fedora Linux, 29 mins FreeBSD. And after switching my homelab over I get instant page loads now even without Zend OPcache. FreeBSD was better in my minor usage and I'm confident to recommend it to others now, and would use it in Enterprise where possible.

 

[ralphbsz]

Not directly related, but I was surprised to see so many BSD mentions on my modem's config page earlier:

Looking at the names, that's probably not the modem containing the BSD source code itself, but instead containing other software that happens to use the BSD license (or more accurately various flavors of the BSD license). Still a good thing.

If someone asked me that, I'd tell them it's faster

See, there are good subjective and objective reasons to run FreeBSD. You just mentioned one, and you have a number to back it up.

It turns out my personal reason is also a mix of subjective and objective. I like to use BSD for servers, because they tend to be well engineered and well designed, unlike Linux, which fluctuates a lot, and always has components that have divergent styles (even if they function well, systemd is an example of something that works, but doesn't fit in). That's a subjective reason, and not the main one even for me. For me, the main reason is ZFS. So someone could say "but you can get ZFS in Linux". True, but in Linux it is always a second-class citizen, and that is a subjective reason.

Someone could say "You can use ext4 and the MD RAID layer". And now that is an objective statement: ZFS is better than ext4 + MD. Not because ext4 is a bad file system (it is an excellent file system, and I'm lucky to consider Ted Ts'o to be a friend), but because ZFS's integration of file system and RAID layer accomplishes something that ext4+MD can not do, by design. It is the integration of file system allocation table with RAID. Here is the logic behind it: The reliability of a RAID system is driven by (among other factors) how long it takes to repair after a disk failure (the formula for that is even in Garth Gibson's original paper), that's known as the MTTR. The MTTR depends on how much needs to be repaired. In a traditional RAID system, such as ext4+MD, the RAID layer needs to repair all blocks that have at least one fault. ZFS does not need to do all those blocks, but only those that have data allocated on them. So if a ZFS file system is 50% full, only half the RAID repairs need to be done, meaning the MTTR is twice as fast (for a single-fault tolerant layout), meaning the probability of a multiple fault causing data loss is only half (for higher-fault tolerant RAID systems, the math gets harder, but the result is even better). So ZFS is objectively better at providing data durability than many other RAID systems, such as Linux' MD layer.

There are other storage system that share this objective advantage of ZFS, but I don't know of any others that are freely available (GPFS = Storage Scale is not free), and suitable for a single-node server (neither Ceph nor Lustre are practical in a small installation). So FreeBSD it is for me.

(By the way, if Ted is reading this: I know you have a day job and little time to mess with free software, but modifying/improving the Linux RAID system might be a nice retirement project. I might help after retirement too.)

 

[rootbert]

I do work in the software industry for more than 20 years now, and to be honest, FreeBSD is not dying, however, it caters just a tiny niche, and that niche will not grow but shrink. Do I prefer FreeBSD over Linux? - Yes.

Please do not feel offended, this is just a description of my situation, no complaint about the best operating system in the world and its contributors. However, let me describe you my situation: when I start a project and I have to tick the requirements with team members, Linux just wins most of the time in my field (over the years: public transport, ISPs, governmental organizations, insurances and banks, medical/pharma). 10-15 years ago, in many cases I could argue to use FreeBSD. Nowadays not so any more. FreeBSD simply lacks many features to use it in my job, and a few years ago I have even switched my personal notebooks due to the sad state of FreeBSDs wifi stack. Quite often I need Linux DRBD/Ceph as high available storage, but mostly application containers is the utmost feature that kills almost every attempt to bring FreeBSD into the game. The OCI infrastructure in Linux is amazing, you get security-patched software instantly from the vendor and do not have to wait for your Linux distribution or FreeBSD ports to incorporate them. For Windows fileshares, Samba server has incorporated many linuxisms, and thus naturally works better on Linux and offers more recent versions than FreeBSD. Hosting VMs? Linux KVMs live migration is a great feature, add a clustered filesystem like OCFS or glusterfs, which just works. ZFS has also been available to Linux for a long time now, so there is basically "no feature" that could convince team members to choose FreeBSD, even though in a two teams there are more FreeBSD fans than Linux fans. For certain use-cases I could recommend TrueNAS, but they have also switched to Linux recently. Firewalls/router for smaller companies? opnsense can only be recommended when you use something else for wifi. Regarding hosting? - Hetzner, a big player in Germany cut FreeBSD installation support for its baremetal servers a few years back, and various hosting providers in central european countries do not offer an option to install FreeBSD; some offer their (costly) remote hands to install FreeBSD, and with some you can at least mount ISO files as virtual cdrom drives. IoT development with ARM? Quite sure you will have more headache with FreeBSD, our projects heavily rely on Bluetooth so FreeBSD is naturally not on the list of options. For some of our projects it is relevant that FreeBSD is just a Tier 3 supported platform of the Python Software Foundation, and with Rust for amd64 it is Tier 2 whereas for arm its just Tier 3 - we work in small teams and unluckily do not have the manpower to improve that situation somehow.

I'd love to see FreeBSD in more of my projects, not only because I prefer to use it, but also because we would contribute to improving the (eco)system. It's a chicken-egg problem: developers are attracted to Linux because more software is available, and easy to install and up to date, and easy to produce and distribute. Thus FreeBSD lacks the manpower, solving some of its problems in order to compete in more areas.

Sorry for this rant about my situation, to sum it up I just hate that monopoly of Linux.

 

[ralphbsz]

I agree with all of your points.

Regarding hosting? - Hetzner, a big player in Germany ...

Amazon AWS, Google Cloud and Microsoft Azure still offer FreeBSD hosts pre-installed on their compute platforms. Or at least they did last time I re-evaluated my cloud-hosts machine, half a year ago. I understand that using the US-based hyperscalers may be not an option in some geographic areas or some legal systems.

Sorry for this rant about my situation, to sum it up I just hate that monopoly of Linux.

I don't hate it, just like I don't hate the law of gravity. Sure, when I drop a wine glass and have to mop up the shards I might be a tad angry, but on the whole, Linux machines have served me well. Even at home I have a handful (the particular use case is one you mentioned, IoT on ARM, where FreeBSD barely works, Linux works smoothly and flawlessly). At work, I tend to have between dozens and millions of Linux machines available, and they do they job just fine.

I think it was Esther Dyson who coined a phrase along these lines (in the 1990s, when it was thought that Microsoft's market share would grow from 90% to 100%): "Microsoft is not a dictatorship of the computer industry; it was democratically elected by the users. Except that the constitution of the computer industry only allows for one election." Well, she was wrong, but she was also right: right now, Linux is the de-facto dictatorship of non-desktop non-personal systems; and again, it was democratically elected by users. There is no point bemoaning it. There is no point pretending that it is otherwise. I think it would be very unwise for the FreeBSD community to try to challenge Linux head on in its home turf, or for FreeBSD to throw its advantages (beginning with its license!) overboard to try to chase after Linux' success. It is a niche OS, but it is really good in its niche.

 

[atax1a]

is Netcraft still around? normally you need them to confirm an operating system's death.

 

[freezr]

I would say that yesterday as well as for today is a technical choice to use FreeBSD instead of Linux; Today using Linux on the server side is not anymore a technical choice as much as using Windows on the desktop. Modern Linux is bloated like an hell exactly as Windows, it is full of proprietary blobs exactly as Windows; actually today Linux and Windows are totally intertwined. These are the tools that corporations have decided to use and to support, and specifically for Linux the reason was the license.

Since the developing of Linux is on the hands of few (and very evil) corporations any time you need to approach something differently you do a technical choice, and perhaps you need a better OS, and the only viable options are the BSD based operative systems. I do believe the more Linux become messy and hypertrophy the more is better for FreeBSD and for anyone is looking for a real technical choice.

 

[USerID]

These days, building internet services doesn’t require much thought about the underlying systems.
With containers and cloud platforms, development has moved far from the hardware. Operating systems aren’t top of mind —
so people default to what’s familiar. And when they do think about the OS, it’s usually Linux.

I completely agree. The operating system is hidden behind a dense abstract layer. Programmers sit on this layer and slap code. System administrators sit in dashboards. Data center maintenance personnel sit in containers, behind these abstract cocoons. There are also supreme owners - investors, banks, corporations, offshore companies, federal systems, government and quasi-government organizations, special services. There are architects of this virtualization. There are managers of abstract levels, there are philosophers, psychologists, etc. This is a food chain, like in nature. And very rigidly algorithmic. So pushing FreeBSD aside with niche users and fans for these investment businessmen is not a tricky business. They have money, power, influence on the masses (on the crowd). We do not shape the consciousness of the masses. They need the crowd, they need us as a service, a product, a service. In their little world there is a concept of power, pressure, money. Opinion leaders (large holdings, emulation and virtualization platforms, BigData organizations) can break and absorb the lower ones. These platforms are already too powerful for any product that gets out of their control to be able to break through to the top or become a competitor for their Oracles, Redhats! They will not allow competitors to enter the market. They will subjugate competitors, like they subjugated Debian for money (systemd). Prostitution. And no offense. We do not have the capital to control the situation. All we can do for now is to master small amounts of money to support laptop Wi-Fi whistles, Bluetooth keyboards, etc.

I chose FreeBSD because now I can at least somehow take control of my systems and understand the essence of things.
That's it, I've spoken out.

The availability heuristic is a fascinating mental shortcut.

There is a lot of nonsense in the world. In addition to this psychological concept, there is, for example,
a terry, classical, indestructible mathematical category: "Game Theory". In my opinion, this is a charlatan model
of herd management. The leading bigwigs (in politics, in financial and industrial groups, in the banking sector, in
retail, investments, etc.) use these methodologies in their own interests, but the herd does not understand this.
This is all manipulation of consciousness and even more. Game Theory is often used against us. If it is necessary
to remove a competitor from the market, then for good money, a corporation can hire a staff of mathematicians and
pay them a good reward. It is not even necessary to officially employ these mathematicians. Why? After all,
cryptocurrency was invented.
So they throw money into their wallets for mathematically stable algorithms.
I was philosophizing.

 

[eseipi]

There is roughly three statements being discussed in the OP:

1. Number of Google searches;
2. "Kleenex" effect;
3. License.

I mostly don't care agree with tanis, kpedersen and ShelLuser regarding the last two points, but let me throw a couple of quotes here about the first one.

First, he claims that FreeBSD is doing well (meaning: not dying), because the number of Google searches for it has a slight uptrend, while the number of Google searches for Linux is constant.

Mark didn't say anything about "doing well", but yeah, that data was used as an evidence of "FreeBSD isn't dying" claim.

But what he omits in there is that the number of searches for the two terms is RADICALLY different. [...] Even if the searches for FreeBSD went up by a few percent per year, it would take a century to catch up to Linux. So his attempt at using data to make an argument falls flat.

I disagree with the latter. Since when dying is the synonym to not having the biggest or even considerable marketshare? By that logic most projects born already dead, which just isn't true. We can say that dying is about negative dynamics, but it's certainly not the state compared to others, so I don't think the argument falls flat in this regard. And by the way, although Mark obviously didn't show the vivid chart of it, he said at least twice about FreeBSD being outnumbered, so there's no need in accusations:

Absolute numbers may be considerably different, but the trend shows a clear flatline for Linux and a gentle up trend for FreeBSD. [...] And when they do think about the OS, it’s usually Linux.

Overall, I'm so tired of this marketshare/growth/competition obsession, that now I think I would personally prefer FreeBSD (along with a lot of other good things I use) to be considered dead by those who value those metrics above all others, so they can stop convincing everyone around them of that. The original article in its way was meant to promote the OS that still works great in many cases and that it's totally fine to be used by companies as well, while the folks from its own community are saying consistently and with passion that it just doesn't deserve wider adoption. I don't see how this can be helpful to anyone.

 

[wolffnx]

Allways the same...is a question of follow the trends..
today , in 2025 , I dear to anyone to compare FreeBSD to Linux , in terms of performance and flexibility to make projects (mean in a good way,not the "fbsd is better, Linux is bulshi#..no)

yes,somethings some drivers are not in fbsd, but for servers is %90 there

In my job interviews, when the job is not Linux and only Linux cult
I say, let me show you with teorical and tecnical presentation why FreeBSD is better than Linux and as gif you save resources(hardware for example)

but is a big NO , and is not teorical is real

when Linux become like windows..lets see
for example, now I am working for a one of importants banks here , they stands on redhat(the linux infraestructure) .. on everything, all is licenced and pay , red hat make a ecosystem of bullsh# and fancy terms and words
they are most and most getaway from reality , blowing and make it more fat..and more fat..."Linux"
one example is ansible...ok...fancy tool, but I can make the same with a sh script, but noo
the "guru" of the job says "is profesional so use it,lets pay big money for that"

 

[johnjohn]

I'm a FreeBSD newbie and you probably do not care about my opinion. However, i never care much what others think of me to a certain degree, of course. I just have to point something out (it bothers me a bit): some members here are a bit rude to newbies and immediately instruct newbies to use Windows, Mac or Linux instead. that's like working at Microsoft and telling customers to buy Apple. And the website address is forums.FREEBSD.ORG. It doesn't help the matter. I think that the forum has had its share of Windows/Linux sideshows that never comeback or what but, still, noone should be encouraging newbies to give up on FreeBSD.

As a Windows user of 26 years, even Linux pales in comparison to some of the features of Windows. I have used 6 different Linux distros since April and they all suck in my opinion. I actually hate Linux since i have started using FreeBSD. Anyway, we all know what motivates the masses: easy and simple. Bring home a new pc/laptop and everything just works. FreeBSD shouldn't try to mock this to increase usage in my opinion. easy and simple is some of the reason that Microsoft sucks. Windows is more like computing for the computing illiterates. You know what i am talking about. But it is really nice to have drivers for almost everything within a certain timeframe and it is also nice to be able to double-click install software and never hear the word 'dependency'. I've never heard this word using Windows. Software just works, but then Microsoft has all of the necessary libraries sitting in system32. Also noone needs to type a single command in Windows. The masses have a comfortable environment and Microsoft capitalizes on it in every imagineable evil way. Remember the 90s/early 2000s nickname for MS? The Evil Empire.

I really do not know why so many people turn to Linux but it is good for developers, hackers, students and privacy buffs. Better than Windows or Mac in my opinion. I like FreeBSD and i think that if some of the driver problems disappear, more people will be using it. Also, simplifying some of the desktop setup will certainly attract desktop environment users. But not all software is available for FreeBSD and that could make a difference with some people. If FreeBSD had more of the simpleton draw then more people will come. Social media crap like filters for videos to compete with tik-tok. Some cool hacking tool that is only available on The Beast. A cool new web browser that really doesn't spy or track its users. Things like this go a long way in attracting users. Remember the old expression: attract more flies with honey...

Windows, Mac and Linux remind me of the old Greek/Roman Trojan Horse story...

 

[wolffnx]

I'm a FreeBSD newbie and you probably do not care about my opinion. However, i never care much what others think of me to a certain degree, of course. I just have to point something out (it bothers me a bit): some members here are a bit rude to newbies and immediately instruct newbies to use Windows, Mac or Linux instead. that's like working at Microsoft and telling customers to buy Apple. And the website address is forums.FREEBSD.ORG. It doesn't help the matter. I think that the forum has had its share of Windows/Linux sideshows that never comeback or what but, still, noone should be encouraging newbies to give up on FreeBSD.

Yes, sometimes happens, but analize that cases, one user came with actitude (and words) like "how to do this"
"dont, I dont wanna do it" ,the 90% of the users that are "rejected" cames in with a bad actitude
that is bad, I am some or way new too, but first is the respect
and you got luck that where are not in the 90 (in that time I even know what FreeBSD is) but the era forums,chats,etc
are full of gurus in that moment...and that whas a bad..bad experience as a newbie

but here,in my short time,was good, but you have to know some "base", not be a Linux guru,even fbsd
but you cant ask "for what is ssh used for" "and what is tcp4" for example

 

[Maturin]

and i think that if some of the driver problems disappear,

If you keep using FreeBSD, next hardware decision you make check if it's supported by FreeBSD. Not vice versa: buy something and then complain there is no driver. One can be satisfied with the hardware FreeBSD already supports, and it's becoming more.

 

[cracauer@]

If you keep using FreeBSD, next hardware decision you make check if it's supported by FreeBSD. Not vice versa: buy something and then complain there is no driver. One can be satisfied with the hardware FreeBSD already supports, and it's becoming more.

Except that it is very hard to find out whether any random computer in a shop is made of hardware that is supported, especially for laptops. Tech specs are a joke these days.

 

[jardows]

If you keep using FreeBSD, next hardware decision you make check if it's supported by FreeBSD. Not vice versa: buy something and then complain there is no driver. One can be satisfied with the hardware FreeBSD already supports, and it's becoming more.

I respectfully disagree, to a point, with this sentiment. If we only purchase hardware that is already supports FreeBSD, we soon will find ourselves limited to only old, outdated hardware.

I have seen cases where individuals have "complained" about non-supported hardware, and someone in the community was able to tackle the issue and release patches to enable support for that hardware. I was a part of the ones looking for hardware support and did some testing for what became the rtsx driver for SD card readers.

At the same time, the "complaints" need to be properly directed - towards the manufacturers of the hardware. We users need to pressure hardware companies to develop native drivers for FreeBSD. Nvidia does it for their GPU drivers, why not AMD and Intel? (That is a rhetorical question, I'm not looking for actual answers) Microsoft writes very few drivers, and it is the responsibility for the hardware companies to write drivers. Of course they have a significant incentive to write drivers for Windows, but if enough people lobby the manufacturers, we could convince more to write for FreeBSD as well.

 

[cracauer@]

We users need to pressure hardware companies to develop native drivers for FreeBSD.

I wouldn't use them.