2026-06-09 Implementation monocultures

It seems that lately the word monoculture is popping up more in my online readings.

Is this a good thing? Let's break this down.

I want to scope the term "implementation monoculture" to mean software implementations. Not other things that could be monoculture (like... bananas). If you (can) look under the hood of the software that you use everyday you will see quite a few of these "implementation monocultures" lurking but quite often it is invisible until you actively start looking.

The higher occurrences of people talking about monocultures I think is a positive thing. Which is the opposite of my opinion of monocultures (skepticism).

So what does monocultures have to do FreeBSD? Isn't this kind of abstract? Maybe. cks blogged about his observations on the phenomena of monocultures in UNIX (err Linux) land. In today's supposed "fast-paced" tech world when it comes to implementations of something, in a commercial setting, who has time to support anything but the most 'popular' platforms?

Whether intentional or not, that decision tends compound and push software ecosystems to monoculture implementations (oh hi skia) because the engineering cost of supporting another platform becomes tied with financial, popularity, and/or availability concerns.

As a FreeBSD user, this choice costs oneself down the line when you find an interesting piece of software or framework and then find out due to reasons it is not supported at all. I think that is unfortunate but it is the state of things (currently).

Using FreeBSD tends to be a choice these days, a choice that has the seed of resistance to the notion of monoculture implementations. Is this something to be nurtured?