i used to have a secret slight disdain for reading Very New Bleeding Edge Research papers because i always felt like they built atop so many mountains of research that i would never begin to comprehend them without starting with some more fundamental paper that developed the seed ideas to begin with. but now that i've read a bunch of "classic papers" for karl's class, i'm starting to realize that sometimes the papers that
build on a body of work give a much cleaner, more concise presentation of the old ideas, and sometimes these old papers are really darn crufty, in the sense that the meat of them is addressing an audience that needed to be convinced that these ideas were worth their salt, and these days we either think it's obvious (because they've become mainstream) or nonsensical (because they never caught on). i still enjoy it, but it's certainly in large part a historical undertaking, and i think only the
gentzen paper (on natural deduction and sequent calculus) actually met my idealized vision of "pure, beautiful, old ideas, though new at the time, presented in a way that makes them crystal clear that they are Great."