Yawn. The answers are predictable (Linux tries to emulate Windows, Linux driver quality is bad), often incorrect (FreeBSD on mainframes… right) and not very insightful. FreeBSD currently has virtually no desktop market share compared to Windows, macOS, and Linux, because:
They have better support from software and hardware vendors.
They are easier to install and configure for the average person.
They are easier to use for the average person.
Most people don’t even know FreeBSD and don’t care about their OS.
Inertia.
Of course, the more interesting question is why Linux became more popular than FreeBSD, despite FreeBSD having a more friendly license for commercial/proprietary use.
Of course, the more interesting question is why Linux became more popular than FreeBSD, despite FreeBSD having a more friendly license for commercial/proprietary use.
I think there’s a better answer to that on Server Fault. UC Berkeley was fighting off a lawsuit from AT&T over BSD, and by the time all of that was resolved Linux had already gotten off the ground and achieved sufficient popularity that the SCO lawsuit couldn’t stop its momentum.
This is often used as one of the explanations. I am sure that it is one of the factors, but the lawsuit was already settled in 1994. I remember buying a FreeBSD 2.1.5 CD set in 1996, long after the lawsuit was settled. In 1996 Linux was still very primitive and a hobbyist thing. Slackware still reigned, SuSE had just moved from Slackware to Jurix as its base, RPM did not even exist yet. I was surprised at the time how much better FreeBSD was - technically, it’s ports collection, the documentation, etc. Also, FreeBSD and BSD/OS were still much more popular on ‘serious’ servers at the time.
I think there are other important (internal) factors. E.g., the development model (outside OpenBSD) favored long-running stable branches and only branching from -current every 2-4 years, whereas Linux distributions were always pushing the latest (except uneven kernel versions), allowing Linux to surpass the BSDs in driver support, etc. Also, the Linux distributions at the time already focused on a wider user base, e.g. Caldera and others had graphical installers near the end of the nineties. And due to many distributions being commercial, they had more incentive pushing Linux boxes to stores and do marketing. E.g., local book stores in The Netherlands would carry Red Hat, SUSE, etc.
the development model (outside OpenBSD) favored long-running stable branches and only branching from -current every 2-4 years, whereas Linux distributions were always pushing the latest (except uneven kernel versions), allowing Linux to surpass the BSDs in driver support, etc.
Both Richard Gabriel’s Worse is Better and entrepreneurs’ highlighting execution over ideas/quality show that this strategy by itself could cause a lot of the momentum of Linux. Also, Caldera was the first one I used since I could buy a CD with graphical installer at Best Buy for $20.
While I see where you are getting to I think those are at least partly myths. I have yet to see a person who can use Linux on the desktop on their own and cannot use FreeBSD or OpenBSD.
While I hear these arguments over and over I just don’t see them mapping to the real life. When people start using OpenBSD or FreeBSD they usually end up thinking it would be a lot harder, because of these myths.
Now I don’t wanna say they are easy to install, but if you want to use one of these systems for day to day life, they are certainly more friendly then Debian and Arch Linux for example. About others one might argue, but really, the first half of the Windows install was about as hard as installing either Linux (aside from Gentoo, Arch, etc.) until fairly recently. I really do thing that the effect of the initial installation is overrated.
What is a bigger problem of course is support for recent hardware. Looking at how far FreeBSD lags behind with Intel graphics (OpenBSD and DragonFly do way better here) or its sometimes desktop-unfriendly defaults (changing a sysctl to make Chrome run correctly) are bigger issues. One can be mitigated by using a recent Apple laptop or an older generation Thinkpad, the other by using a “distribution”.
I think a big reason is that all the commercial interest in something for end users was around Playstations. There is no SuSE, no RedHat, no Ubuntu, all having at least some money and grip to push their systems to the desktop, and be it just to get future sysadmins, selling their products for them.
Right now I think the comparison that would make more sense is Arch Linux vs FreeBSD, simply because these are a lot more similar, than Ubuntu which won by a huge initial investment in tech, branding and marketing, more than anything.
Arch Linux and FreeBSD have a lot more in common - speaking purely about desktop. They have a somewhat technical users in mind, they are not backed by some big organization, they value certain forms of simplicity (not exactly the one OpenBSD is thinking about, but still), they both have huge repositories of easy to install and very up to date software, that can be either taken from packages or source, they like to tune, configure and optimize, they enjoy having packages close to upstream, etc.
My best guess here would be stuff like steam and other things that became available as Linux blobs, that were “made possible” due to the investments from various other companies, which started out in the B2B field. Now one might ask why the BSDs don’t have strong companies in that sector, but at least to me it seems that the idea of using BSD outside of networks (routers, servers, etc.) infrastructure and the need of having something GPL-free (gaming consoles, etc.) just never occurred to people until Linux did lift off.
The reasons to use BSD often were a lot more more pragmatic and there weren’t really people with that dream of one day replacing Windows, in which Linux so far succeeded on the phone, but more in a way that one could say it’s Linux + lots of BSD code and macOS and iOS are BSD after all.
Even though BSD people often don’t want to hear that, but the license might be a part, especially on the hardware support side and you simply have code flowing in, for mostly legal reasons that one at least can look at.
I am sure that’s not the only reason and of course it will be a mixture, but the person running Linux on their own free will likely won’t decide against FreeBSD because it looks text based, especially not your average Arch Linux, Gentoo, Debian, Slackware. They might even find it more convenient.
Knowledge is especially historically a huge factor. I haven’t heard about BSD at all before the day I first installed it in 2005, which I think was because I read that Gentoo’s portage was inspired by it.
Extremely subjective, but Linux seems to really have a lot more missionary stuff going on. I have met more than one person who was about to duck and cover when I mentioned Linux, fearing a speech about moral and technical reasons on why they should switch. This used to be worse though. I think with the growth of the Linux community people feel a lot less like they have to defend their decision. There are barely any flame wars about Windows vs Linux vs macOS these days.
So I think marketing and in general network and social effects, as well as a hype and a nice story together with quite a bit of ideological undertone make up a large portion, of the history leading to status quo. I know a few people that tried BSD liked it and only switched back for ideological reasons.
While the BSDs are certainly not easy to use compared to macOS or Windows, I think argument is not holding true as big driving factor at all, when comparing to Linux in general and longer term desktop usage. Even holding on to Ubuntu for an extended period of time (upgrading from one release to another) will require a similar level of interest.
I think a lot of it may also have had to do with GCC being so popular, and the push from the GNU folks towards Linux (at least “until Hurd is ready”). Combine that with Linux often being positioned as Anti-Windows by users (I remember a /lot/ of zealous propaganda back in the day), it certainly started to pick up mindshare quickly on college campuses in the mid 90s.
Some quick googling gives me https://netmarketshare.com/linux-market-share - it says < 3% and the other results also all say < 5%. And I think Ubuntu is “kinda mainstream” at least as much as it gets.
Sure, 5% is a lot more than 0.5% but maybe the people using Linux at work still have to fight enough to not have to use Windows or OS X, in most positions it’s not worth the fight.
Even in my most “I can do whatever I want” positions I still had a pretty hard requirement on either VirtualBox for local virtualization or writing Java code or using Google Hangouts daily (ok, these last two might work flawlessly on FreeBSD) or some other stuff that only works on Linux without pain.
In my current job it’s again some stuff I’m not sure I could get to work easily on FreeBSD, and I’d say I’m a lot closer to using a BSD again than others who have no experience and have never tried it.
Yawn. The answers are predictable (Linux tries to emulate Windows, Linux driver quality is bad), often incorrect (FreeBSD on mainframes… right) and not very insightful. FreeBSD currently has virtually no desktop market share compared to Windows, macOS, and Linux, because:
Of course, the more interesting question is why Linux became more popular than FreeBSD, despite FreeBSD having a more friendly license for commercial/proprietary use.
I think there’s a better answer to that on Server Fault. UC Berkeley was fighting off a lawsuit from AT&T over BSD, and by the time all of that was resolved Linux had already gotten off the ground and achieved sufficient popularity that the SCO lawsuit couldn’t stop its momentum.
This is often used as one of the explanations. I am sure that it is one of the factors, but the lawsuit was already settled in 1994. I remember buying a FreeBSD 2.1.5 CD set in 1996, long after the lawsuit was settled. In 1996 Linux was still very primitive and a hobbyist thing. Slackware still reigned, SuSE had just moved from Slackware to Jurix as its base, RPM did not even exist yet. I was surprised at the time how much better FreeBSD was - technically, it’s ports collection, the documentation, etc. Also, FreeBSD and BSD/OS were still much more popular on ‘serious’ servers at the time.
I think there are other important (internal) factors. E.g., the development model (outside OpenBSD) favored long-running stable branches and only branching from -current every 2-4 years, whereas Linux distributions were always pushing the latest (except uneven kernel versions), allowing Linux to surpass the BSDs in driver support, etc. Also, the Linux distributions at the time already focused on a wider user base, e.g. Caldera and others had graphical installers near the end of the nineties. And due to many distributions being commercial, they had more incentive pushing Linux boxes to stores and do marketing. E.g., local book stores in The Netherlands would carry Red Hat, SUSE, etc.
Good points here.
Both Richard Gabriel’s Worse is Better and entrepreneurs’ highlighting execution over ideas/quality show that this strategy by itself could cause a lot of the momentum of Linux. Also, Caldera was the first one I used since I could buy a CD with graphical installer at Best Buy for $20.
Citation needed. ;)
While I see where you are getting to I think those are at least partly myths. I have yet to see a person who can use Linux on the desktop on their own and cannot use FreeBSD or OpenBSD.
While I hear these arguments over and over I just don’t see them mapping to the real life. When people start using OpenBSD or FreeBSD they usually end up thinking it would be a lot harder, because of these myths.
Now I don’t wanna say they are easy to install, but if you want to use one of these systems for day to day life, they are certainly more friendly then Debian and Arch Linux for example. About others one might argue, but really, the first half of the Windows install was about as hard as installing either Linux (aside from Gentoo, Arch, etc.) until fairly recently. I really do thing that the effect of the initial installation is overrated.
What is a bigger problem of course is support for recent hardware. Looking at how far FreeBSD lags behind with Intel graphics (OpenBSD and DragonFly do way better here) or its sometimes desktop-unfriendly defaults (changing a sysctl to make Chrome run correctly) are bigger issues. One can be mitigated by using a recent Apple laptop or an older generation Thinkpad, the other by using a “distribution”.
I think a big reason is that all the commercial interest in something for end users was around Playstations. There is no SuSE, no RedHat, no Ubuntu, all having at least some money and grip to push their systems to the desktop, and be it just to get future sysadmins, selling their products for them.
Right now I think the comparison that would make more sense is Arch Linux vs FreeBSD, simply because these are a lot more similar, than Ubuntu which won by a huge initial investment in tech, branding and marketing, more than anything.
Arch Linux and FreeBSD have a lot more in common - speaking purely about desktop. They have a somewhat technical users in mind, they are not backed by some big organization, they value certain forms of simplicity (not exactly the one OpenBSD is thinking about, but still), they both have huge repositories of easy to install and very up to date software, that can be either taken from packages or source, they like to tune, configure and optimize, they enjoy having packages close to upstream, etc.
My best guess here would be stuff like steam and other things that became available as Linux blobs, that were “made possible” due to the investments from various other companies, which started out in the B2B field. Now one might ask why the BSDs don’t have strong companies in that sector, but at least to me it seems that the idea of using BSD outside of networks (routers, servers, etc.) infrastructure and the need of having something GPL-free (gaming consoles, etc.) just never occurred to people until Linux did lift off.
The reasons to use BSD often were a lot more more pragmatic and there weren’t really people with that dream of one day replacing Windows, in which Linux so far succeeded on the phone, but more in a way that one could say it’s Linux + lots of BSD code and macOS and iOS are BSD after all.
Even though BSD people often don’t want to hear that, but the license might be a part, especially on the hardware support side and you simply have code flowing in, for mostly legal reasons that one at least can look at.
I am sure that’s not the only reason and of course it will be a mixture, but the person running Linux on their own free will likely won’t decide against FreeBSD because it looks text based, especially not your average Arch Linux, Gentoo, Debian, Slackware. They might even find it more convenient.
Knowledge is especially historically a huge factor. I haven’t heard about BSD at all before the day I first installed it in 2005, which I think was because I read that Gentoo’s portage was inspired by it.
Extremely subjective, but Linux seems to really have a lot more missionary stuff going on. I have met more than one person who was about to duck and cover when I mentioned Linux, fearing a speech about moral and technical reasons on why they should switch. This used to be worse though. I think with the growth of the Linux community people feel a lot less like they have to defend their decision. There are barely any flame wars about Windows vs Linux vs macOS these days.
So I think marketing and in general network and social effects, as well as a hype and a nice story together with quite a bit of ideological undertone make up a large portion, of the history leading to status quo. I know a few people that tried BSD liked it and only switched back for ideological reasons.
While the BSDs are certainly not easy to use compared to macOS or Windows, I think argument is not holding true as big driving factor at all, when comparing to Linux in general and longer term desktop usage. Even holding on to Ubuntu for an extended period of time (upgrading from one release to another) will require a similar level of interest.
I think a lot of it may also have had to do with GCC being so popular, and the push from the GNU folks towards Linux (at least “until Hurd is ready”). Combine that with Linux often being positioned as Anti-Windows by users (I remember a /lot/ of zealous propaganda back in the day), it certainly started to pick up mindshare quickly on college campuses in the mid 90s.
Does Linux even have a real market share?
Some quick googling gives me https://netmarketshare.com/linux-market-share - it says < 3% and the other results also all say < 5%. And I think Ubuntu is “kinda mainstream” at least as much as it gets.
Sure, 5% is a lot more than 0.5% but maybe the people using Linux at work still have to fight enough to not have to use Windows or OS X, in most positions it’s not worth the fight. Even in my most “I can do whatever I want” positions I still had a pretty hard requirement on either VirtualBox for local virtualization or writing Java code or using Google Hangouts daily (ok, these last two might work flawlessly on FreeBSD) or some other stuff that only works on Linux without pain. In my current job it’s again some stuff I’m not sure I could get to work easily on FreeBSD, and I’d say I’m a lot closer to using a BSD again than others who have no experience and have never tried it.