It is self-evident that any and all paths must be open to a researcher during the actual course of his [or her] investigations.
A mathematician who is not also something of a poet will never be a complete mathematician.