That doesn't mean you shouldn't have a reasonable understanding of what those abstractions are doing, though.

For the purpose of selecting a DOM node in a performant manner, you really don't need to know how CSS/XPath traverse the DOM. You just need to know that querySelector exists.

In general, I think it's enough to know what your abstraction does, rather than how it does it.

There's a huge difference between using an abstraction with vs. without understanding what it does. Often the latter leads to more layers of abstraction piled on top to "fix" the problem, until everything sorta works and is barely workable at the same time.

"All non-trivial abstractions, to some degree, are leaky."

https://en.wikipedia.org/wiki/Leaky_abstraction#The_Law_of_L...

It's build on top of the abstractions; not depend on someone else's implementation of all the abstractions...