I think of Feynman as the archetypal mad scientist, and while I don't think he happened to love fanfiction (and actually, don't we mean "writes fanfiction"?), I wouldn't have been surprised to have found out that he did and I wouldn't have thought less of him if he did.

I think the real issue is not that "writes fanfiction" is not part of the archetype but that you have (or think others will have) some kind of moral/emotional reaction to "writes fanfiction" that causes you to think about it in different terms than "writes poetry" or "loves functional programming" or "loves stamp collecting" or "loves civil war re-enactments" or whatever.

I think the underlying question is how inauthentic one should be willing to be in order to "present the best image." You and I both love functional programming, but there are many "Enterprise Architects" that would find passion for functional programming weird and suspect, deeming it pointless love of complexity for the sake of obfuscation. Imagine you were a public figure for a software company that marketed mostly to Enterprise Java shops, and somebody tells you that you should consider avoiding writing publicly about functional programming, working on xmonad, participating in haskell-cafe, because it might give potential customers the wrong impression (however stupid that "wrong" impression might be). If you think that "functional programming" and "stamp collecting" and "writing poetry" are more valid "side passions" than writing scifi or fanfiction, can you give a good explanation for why, or is it just a matter of "what most people would think"?