You are right, my reasoning is as follows (though still 100% personal bias).

For writing a web server library, like oak or flask or etc etc:

First, resources own the methods, methods don't own the resources, so the path comes first, easy enough, its just a string with potentially placeholders whose values need to be shunted down the line

Next is the method, for some reason a lot of libraries like to hardcode these? Immediately jumping away from comprehensive, unless you have a language that makes identifier-like things possible, e.g. symbols in ruby, or `__getattr__` in python, or proxies in javascript, of the 3, proxies are probably the most cumbersome in implementation.

then the request handling, that is fine for all languages

Templating is more of a bother: Either you use JSX, which in most cases is not XML, but instead DOM, subtle but a pain, additionally requires a compilation step before you can run the code, and also requires a DOM or DOM-lite library. Or you do chained calls with an ending call, e.g. https://sonnet.js.org/, or you do something like lit-html which is quite practical. You can't do something like title("hi") vs title({attr:123}, "hi"), as everything is an object and {} are objects, so you need to have fixed function signatures unless you want to have a lot of bother in implementing things. This is all less of a problem in other languages (e.g. Python, Ruby, Raku)

In short, Javascript does not flex when you push on it, that is great for a lot of tasks, but not so much for building frameworks where instead of having systematic ways of extending the language, either tricks are used which break other things down the line ( https://news.ycombinator.com/item?id=45472119 ), or you have opinionated narrow frameworks, which results in more churn as instead of frameworks, its more, frame-specific-use-cases.

it feels like a fad to hate on js and js frameworks; esp from people who parrot what others are saying without any actual insight