teach computers to read the internet (open-source)
structure > semi-structured > unstructured

1. parse what you expect; 2. see what else is there; 3. repeat real fast
parse: grammar says what to expect
see: sample what you got from visualization, saving best to a file

factbook has characters
factbook has keys and values
factbook key capitalization means something
factbook has a few keys used everywhere

real fast trying experiments
real fast reading files
read all of wikipedia in under an hour

exploratory mode works because of innovation in 2004 (2 minutes)
this is not the stuff you used to learn in compiler class


scanner v. parser v. PEG
wiki syntax == repeated scanning by php


preview: restoring old programs (45 seconds)
preview: reinventing wiki for open-data

QR Code (15 seconds)