Ten Challenges

The deliberate sequel to Ten Great Books two months earlier. That list was the comfortable one — books finished, cleanly recommended. This one is harder and more honest: ten books Steve is in the middle of, mostly because finishing them is a multi-year project. Gödel, Escher, Bach sits at #1 with the warning that nearly nobody finishes it; SICP at #2 with the confession that the condescending tone almost made him quit before he realised what he had in his hands. Knuth's Digital Typography is the unexpected pick (a book about typography typeset in the program it documents). The compiler/language stack — Scott's Programming Language Pragmatics, Friedman/Wand/Haynes, Pierce on types, the two Scheme books — collectively form the curriculum Steve thinks the field skipped. The closing turn is Okasaki's Purely Functional Data Structures, on the list because of why Steve and his colleague JG keep coming back to it — they've been running a private research program at Amazon for a year on how to escape the J2EE/C++ swamp, and the literature keeps pointing the same direction. A higher-level language abstraction backed by functional-programming machinery, on top of the network as a single programmable substrate.

January 2005, with Erlang's revival, Scala, Clojure, and the whole map/reduce/actor wave still several years out. Steve is reading the road signs and the essay closes on that note: books as a survival kit for a distributed-computing world that hasn't arrived yet.