> The problem with reified generics is that the same variance model must be adopted by all guest languages on the runtime. I think this is a fair observation,

> The problem with reified generics is that the same variance model must be adopted by all guest languages on the runtime. I think this is a fair observation, but it really boils down to "a dynamically typed compilation target is an easier target", which isn't all that surprising. > such as Scala to port to the CLR I am not as familiar with Scala's saga here, but I've heard multiple conflicting reports from Scala insiders, so I think this is a more complicated issue than just the generics model of the CLR.