I'm not convinced it is a realistic goal to attract more new desktop users, given that what would get /me/ to consider moving is if FreeBSD offered a better UX than the big 2, Linux and macOS. As it is, it doesn't: performance is less, feature set is less, software availability is less; and as a result, developer enthusiasm to fix any of these is less. Libraries and workflows are already tuned to work on either x86_64 Linux or ARM64 macOS. Second system effect will always get you, especially when you're the fourth system! We've already seen this play out with OS/2, a technically superior system at that point in time which lost the long race and only bought itself some life support with Windows compatibility, ATM install base, and general corporate torpor. Your product must win on its merits, but those merits have to be ones that appeal to buyers, not you-the-product's-developer.
Where I find FreeBSD especially suited, and what I think FreeBSD should lean into is making its platform THE attractive embedded/specialist/RTOS(?) option for companies building products they don't necessarily want to open source. Go do what Linux did to Solaris and Irix, and put VxWorks and QNX out of business. The BSD license and the current 'fair source' trend (responding to the friction between Stallmanesque Open Source and capitalism) seems to me a synergistic pairing. Get funding from companies to maintain the base so they don't have to, and charge them handily for integration consultancy. Use that to bootstrap achieving parity with Linux, and then perhaps more folks will climb on the bandwagon.
You mean what Intel/AMD/Dell/Nvidia did to Solaris and Irix.
SGI couldn’t sell enough hardware, the OS was of no consequence. Back when SGI was circling the drain, most of their customers were perfectly OK to migrate over to NT.
Linux eventually fucked over Microsoft. Because Microsoft kept the desktop but (mostly) lost the datacenter.
> put VxWorks and QNX out of business.
Numerous potential customers of these products use Linux (the kernel GPL often isn’t an issue). Those that can’t are unlikely to migrate to FreeBSD either. Their remaining incumbent customer base is often in an industry where migration costs are higher (qualifications, certifications), those companies provide value adds (GUIs, toolkits, specialized hardware and simulation libraries) far beyond just the kernel (Wind River will even gladly sell you Linux if you want). Undercutting a little on a kernel license isn't compelling.
And it’s not like FreeBSD is a great RT kernel or has much to help with safety certification. FreeBSD is usually not technically appropriate to replace vxWorks, and when it is, Linux will usually do just fine.
Where I find FreeBSD especially suited, and what I think FreeBSD should lean into is making its platform THE attractive embedded/specialist/RTOS(?) option for companies building products they don't necessarily want to open source. Go do what Linux did to Solaris and Irix, and put VxWorks and QNX out of business. The BSD license and the current 'fair source' trend (responding to the friction between Stallmanesque Open Source and capitalism) seems to me a synergistic pairing. Get funding from companies to maintain the base so they don't have to, and charge them handily for integration consultancy. Use that to bootstrap achieving parity with Linux, and then perhaps more folks will climb on the bandwagon.