my @ids       = qw(a b c d e f g h);
my $chosen_id = '';

# Flat file!
# each line is a time and an id separated by a space
open my $recent_fh, '<', 'recent'
    or die "Can't read recent: $!";
my @recent_lines = <$recent_fh>;
close $recent_fh or die "Can't close recent: $!";
chomp @recent_lines;
my %recent_ids = map { split } @recent_lines;

# remove the recent ids from the id list
my @possible_ids = grep { ! $recent_ids{$_} } @ids;
$chosen_id = $possible_ids[rand @possible_ids];

# process it
# remove the oldest, as before

open $recent_fh, '>', 'recent'
    or die "Can't write recent: $!";
print $recent_fh
    map { "$_ " . $recent_ids{$_} . "\n" }
        keys %recent_ids;
close $recent_fh or die "Can't close recent: $!";
[download]

Note there's the possibility of clobbering the recent file if it gets killed while writing it. Also, there's no provision for multiple instances writing to the same file (that is, there's no file locking).