Good to see Pete Geddes testing. The idea of a number of "complicated" random data sets for testing is a good one.
Just four comments: a. It would be good for the further testing if you could create a zip file with the data and upload to the bug/development system - see the link on www.epidata.dk/testing.php
b. It seems you are reading the dates as : mdy, but looking at the data the dates are ymd
00063364;.........;1949/07/29;N;1958/04/19;CHE;75683
which might explain the strange date values you have seen.
c. Regarding UTF-8 This really does only come to play with localised alphabets/letters.
11 d1 Date of Birth Date (mdy) 10
12 v1 Gender Upper-case text 1
13 i11 Date Entered Today (mdy) 10
d. regarding Operating systems The linux version. We will create a list of dependencies. Next in line is testing a Mac version - but awaits arrival of hardware.
regards Jens Lauritsen EpiData Association