Paul Fultz II

> What I proposing to you don't includes too much meta-programming. It is a quite bit of metaprogramming, not just internally, but the interface would need to provide these meta...

Ok i've added `construct_meta` and documention.

> I don't believe construct_meta is the solution. IMHO, we need just a function that do whatever is needed. I don't know what you mean. What should the solution be?

> should not work as there is a vector constructor Nope `construct()` does not work. I don't see that as a problem. The template version should be used when you...

> IIUC, what you mean is that the the Tconversion operator of implicit()(xs...) behaves like F()(xs...)) Yes, that's correct. I am not sure how to make the semantics clearer. >...

So it looks like `__is_literal_type` intrinsic is still available, so I should use that instead.

Actually, I would like to implement an adaptor to aid multivisitation. One approach is this [here](http://mpark.github.io/programming/2015/07/07/variant-visitation/) which is interesting. I envision a `matrix` decorator that could be called like `matrix(n1,...

> Hi, in P0050R0 I suggest a select function applied to any and taking a variadic list of types to inspect. select(a) behaves like a nullable variant, variant. Now we...

> FIT_STATIC_FUNCTION(visit_all) = compose(capture(flip(apply))(compress), partial); So that won't work since the callback must always return a fixed type. I will have to think about that more.

> I can't help but think that this is some form of lifted function application on variants. Basically a generalized ap to many arguments. Perhaps this is an interesting point...