[Espressosaurus]

I agree with this. I also like to divide things into "this requires research/invention" and "this does not require research/invention, just work".

Anything requiring invention by definition is not straightforward because its potential timeline may be functionally infinite if it cannot be solved with the tools at hand.

A task like changing all APIs to accept a different set of parameters might take 6 months, but is straightforward.

"Easy" is the one line change that has no side effects, and is uncommon in practice. After all, sometimes it is that one line change, but there are ramifications elsewhere in the system that need to be thought through.