Yeah, the my thing was really confusing me. Thanks for the clarification.

How large a text file is too large to be practical, roughly speaking? I know even less about databases than I do about Perl, so I'd like to stick with some semblance of my original approach if it's feasible.

Last, I found the random_line subroutine here: http://search.cpan.org/~bigj/File-Random-0.17/Random.pm#FUNCTION_random_line(%24filename_%5B%2C_%24nr_of_lines%5D)

What do I need to add to my script to make it work?

(And as I continue to struggle with this project, the option of simplifying things down and abandoning the randomizing device becomes more and more attractive.)