you lying, non-ascii bastards

grep -l $'[\x80-\xff]' * > nonascii.txt

brain dump

What about a collection of micro-apps that extract linked data from epidoc, a la SNERT/OC? One for date info normalization, one to spit back pleiades, etc.