Comment Re:vice as virtue (Score 3, Informative) 107
In the early 90s AT&T sued Berkley Software Design because BSDs originated with AT&T Unix. This led to distrust of the BSDs at the precise moment where Linux rose.
Part of it. It was a stumbling block out of the gate. I tried the Jolix BSD (before all the others) and it didn't work on my 486. Linux did. FreeBSD and BSDi didn't exist quite yet. I didn't look at the BSDs again for 10 years. By that time I had introduced Linux into my work (I was a sysadmin). I was lucky to work at a place that used OpenBSD for firewalls and FreeBSD for web servers. Linux was everywhere else.
I think there were other factors. Red Hat and SUSE certainly drove adoption of Linux and I don't think any of the BSDs had an equivalent corporate engine driving adoption forward. Then there's the marketing aspect. Once Tux started showing up Linux really started being seen as a viable "brand." That may seem silly but branding works—it instills curiosity in potential consumers and trust among those who make purchasing decisions.
Sure. Some wouldn't touch FreeBSD because of the Daemon mascot
I would never argue that the license had nothing to do with the success of Linux, but I do not think its success was dependent on the license. At the time, people had a lot more faith in the GPL than they do today and it certainly drove a lot of developers to Linux.
Linux was easier to contribute to. It seems like you have to be part of the "inner circle" to add to BSD. Plus there are 3 of them. When you contribute to the Linux kernel, it goes to all Linux distributions.
But things tend to happen for a confluence of reasons and I think that's especially the case here. Linux was in the right place at the right time. Your Oracle example is one of many things that happened to go right for Linux.
I remember an Oracle developer at a Linux user group saying he compiled Oracle on Linux & it didn't need any changes. This was before it was offered for sale. When Oracle started selling for Linux, it drove sales.
The dotcom collapse also meant there was no VC capital to buy "proper" Unix systems. They had to max out their credit cards so they bought x86 and put Linux on them. In most cases they found the AMP stack was faster on Linux than on Solaris that they were paying $$$ for during the dotcom.
But I also don't like the OP's assertion that "Linux won," as if this is a zero-sum game. My laptops and desktops use Linux because it works best for those applications. My servers (web server, file server) use FreeBSD because I prefer FreeBSD for servers. For small Raspberry Pi projects I use FreeBSD and I prefer FreeBSD for routers. At work I use a Mac for most things, a Linux box for other things, a FreeBSD server hosts our internal documentation, AlmaLinux servers, and a TrueNAS file server (used to be BSD-base and now Debian). FreeBSD has its niche and it does it well, and being niche shouldn't be equated with losing.
Having multiple systems can help w/ security. But having 1 OS means you need one set of skills. I don't want to go back to the admin differences between Solaris, SunOS, OSF/Digital Unix/Tru64, HP-UX, Irix, Xenix, Ultrix. Linux is basically rpm/yum and deb/apt flavors.
The BSDs might be better for some things but you always adapt Linux. OpenZFS is a good example. It's often a source code download that gets built locally and installed because of the CDDL license where in BSD, it's in the kernel and distributed as a binary.
Maybe containers can be added to BSD, I think they're on MacOSX. They're on Windows (they have a Linux layer to do it!).