Interfaces/traits in julia 2.0 and multiple inheritance

[Mason]

@jakobnissen

Broadly speaking, I agree that Julia’s inheritance model is not particularly good. Most of the interesting properties of types are not monophyletic.
But I do believe you’re underestimating the problem of ambiguity that would result from crossing Julian multiple dispatch with a trait system. I fear method ambiguities would explode.

For what it’s worth, we had this conversation already rather extensively in Why did Julia choose nominal typing over structural typing/traits? and the response was more or less that the OP did not believe that this would actually be a problem.