Yes, both choices are possible. And a choice has already been made. This allows Base to have fallback methods.
Can you explain what you don’t see here in the previous post? You can also just look at the docstrings for isless, <, <=, etc.
To be precise, my claim was about isless and <=, not < and <=. And to be less precise, I did use scare quotes around fundamental. I would not say one of these concepts are inherently more fundamental than the other, but a choice has been made to define (by default) comparisons in terms of isless, and not e.g. to define isless in terms of <=.
A user is able to make another choice for their own type. They could define <= first, and then define isless in terms of <= and ==. Their method definition for isless will look generic, but because the user has gone against the design intention of the generic functions, these methods must be limited to their own type, otherwise there will be method ambiguities.
I am really struggling to understand the confusion here. Some people are wishing that issorted was defined in terms of something like <=. Okay, but it’s not. That it is defined in terms of something like isless is reasonable.