i have a confession to make. i almost obliterated a good chunk of kucinich supporters (well, their email addresses). i was trying to be helpful, offering to do some data mangling / cleanup tasks as a token of appreciation for henri, who graciously hosted me in san francisco.
so i sat down with the openoffice spreadsheet and went into tinkering mode. after a while, i had the procedure down, and was happily getting rid of dupes and sorting the addresses in various ways. scanning through the dataset, i noticed something odd: entire letters were missing from the alphabetically ordered list. openoffice had somehow munged ~30% of all records (~100k email addresses). fortunately, i had backups. redid the job with excel, without a glitch.