There are Forth words with generally agreed-on definitions but it seems it's not designed to grow a large, common vocabulary in the way that modern package managers enable.
There are Forth words with generally agreed-on definitions but it seems it's not designed to grow a large, common vocabulary in the way that modern package managers enable.
Sometimes... you don't need to understand anyone else, or to be understood by anyone else. Sometimes you just want to do what you need to do.
The goal of replicating understanding is all well and good, but that shouldn't be the primary goal of programming language development, nor should it be the primary goal of programming.
I don't think that anyone is suggesting that we start rebuilding enterprise service frameworks in Forth.
Arduino itself is a good example of how a decent, common abstraction can help users get started.
One of the big advantages of Forth is that it's (apparently) pretty trivial to implement on your own, in whatever environment you choose.
When that is layered on top of overly generic libraries/frameworks, you may end up using only parts of the underlying vocabulary. Now the project is stuck with two vocabularies. Learning the generic one doesn't get you far and may be actively misleading ("yeah, we never use features X and Y of the framework, instead we've overridden Z with our own implementation that does all this plus W").
But these problems are, in their way, a sign of success at the package management level. If you don't have good package management then people probably aren't having much success at sharing code.
We should still aim for building and sharing stable, well-understood, high-quality code libraries. Unfortunately, sometimes the way to get there is to go through a period of instability.
And I agree with that 100% - but that isn't the solution to every single problem. That's all I'm saying.