I nearly forgot that today is Don Quixote's 400th birthday. It's worth noting, as any introduction to the novel will, that there is hardly a literary theory or construct that Cervantes did not prefigure in Don Quixote. (Modernist and postmodernist strategies not excluded, given the enduring theme of inescapable madness and the book's unreliable, polylingual narrators, metacritical narrative, and self-reflexive structure.) A fact made all the more signficant by Cervantes's indisputable contribution to the canon as its first member, his book being the first novel. I've heard that both Faulker and Tolstoy read it once a year.
Of course you don't reach for Don Quixote for all that, you read it because it's enchanting, hilarious, terrifying, everything you want from a novel, and any person whose nose doesn't scrunch up a little whenever Quixote describes his love for Dulcinea lacks soul.
Any time DQ comes up, it's noted how few people get around to reading it—even among good readers. There hasn't been a more appropriate year to resolve to read it in the last century. I recommend Burton Raffel's edition, not only because it's the version I've read, but because it has a strong supporting cast, including: other writings by Cervantes (including "Portrait of the Artist as an Old Man"), source material (selections from de Montalvo's Amadis of Gaul and Ariosto's Orlando Furioso), apocrypha (the Prologue to de Avellaneda's "false" Quijote), and scholarship, both good (Foucault and Borges, María Antonia Garceés's timely essay on Andalusia and the veil) and bad (Harold Bloom).