Should AbstractPolynomial subtype Function?

[nsajko]

Since AbstractPolynomial types are required to be callable, wouldn't it make sense to change the definition from

abstract type AbstractPolynomial{T,X} end

to

abstract type AbstractPolynomial{T,X} <: Function end

Then polynomials would be accepted anywhere functions are.

[jverzani]

Yeah, this has come up with plotting recipes as useful. I've always been wary, as I'm a bit biased towards the interpretation that polynomials are more like abstract vectors with an offset.

[nsajko]

Makes sense.