Could perhaps be posted as a response to some of the other mixin discussions, but...

I had a thought on the use of mixins and on syntactical requirements following from that. In my mind, ideally, you would define classes and structs as usual. Then, later on, perhaps in some other project, you would find yourself in a situation where you realize that the class you're writing needs this, this and that bit, all pieces of functionality that you had already written before. So you say something along the lines of:

class X {
    reuse class Y, Z;    // or "aggregate" or "mixin", I don't really care
at the moment
}

My point is, ideally you would be able to postpone the decision to make some functionality a 'mixin' until you find a use for it as such. This means that there should be no special syntax to declare some class as a 'mixin', only special syntax to use some class as a mixin

There could be restrictions, like "a class can only be used as a mixin if it has no datamembers or virtual methods", so implicitly you would have to foresee potential future use of a class as mixin.

Does any of this make sense?