Good point biker.

You can actually start out using DBD::CSV which actually uses text files and then move onto a real database file. That would allow you to focus on learning just the DBI/DBD part of the piece and then when you are comfortable with it, move on to putting in a real database.

Disclaimer: This is in theory, ie. this is how it should work. Never tried it so am not sure if there are any practical problems involved. Would someone who has attempted something similar please speak up.