wrote an algorithm to calculate grouped weighted means using sql in r, paving the way for analysis of huge survey data sets with free statistical software and limited computing resources.
found a readable discussion of good statistical programming practices.
a person does not really understand something until he can teach it to a computer
your program should not explode, or worse, produce plausible nonsense, but should instead complain noticeably and informatively
insufficiently documented software is indistinguishable from magic
saw u2 in baltimore.
hit a language families gem on wikipedia.
got an awesome postcard.
am infinitely pleased by the new cigarette warning labels.
have gotten into deadwood. best line so far: "announcin your plans is a good way to hear god laugh."