A recent story of a non-feathered dinosaur in a presumably feathered clade is stirring the pot a bit right now. The importance of shared derived characters in systematics is one reason that something like feathers tends to elicit a response from many scientists, since one character might result in the reworking of the tree of life. But there are other ways to study feathers aside from paleontology, in particular, I point you to the work of Richard Prum (in particular, this paper) and Matthew Harris. Development and genetics can be crucial supplements to the fossil evidence in this case, though dinosaurs are no longer around, birds and the crocodilian outgroup are.