The web (and CGIs) are stateless, so "taking out the printed lines each time" isn't as easy as you think.

A better option would be to abuse the magical $. variable, that tells you which line you're on when iterating over a file handle, so if your CGI submits an offset and number of lines to display, you can do something like...

print header;

my $start = param('start_line');
my $end   = $start + param('show_lines') - 1; 

if ( open TEXT, "< /path/to/your/file.txt" ) {
    while ( <TEXT> ) {
        last  if $. >= $end;
        print if $. >= $start;
    }
} else {
    print "Oops, there was a problem!";
}
[download]

See also HTML::Pager (alt.).

    --k.