A few approaches, some already covered by others:
-
The split approach produces a lot of 'noise' (even when using a more reasonable /\W+/ split pattern) that must be removed with further processing.
-
A tricky approach is to try to figure out just what a 'player' is and define a regex to extract those substrings.
-
Maybe the easiest and most reliable approach is to go to the dramatis personae list at the beginning of the play, look at all the players found there, and make a regex of that.
>perl -wMstrict -le
"my $char_list = 'Exit Cassio; Enter Iago, Othello, and others';
;;
my @words = split /\W+/, $char_list;
printf qq{'$_' } for @words;
print '';
;;
;;
my $not_player = qr{ (?! Enter | Exit) }xms;
my $player = qr{ \b $not_player [[:upper:]] [[:lower:]]+ }xms;
;;
my @players = $char_list =~ m{ $player }xmsg;
printf qq{'$_' } for @players;
print '';
;;
;;
my @dramatis_personae = qw(Cassio Iago Othello);
my ($character) =
map qr{ \b (?: $_) \b }xms,
join '|',
@dramatis_personae
;
;;
@players = $char_list =~ m{ $character }xmsg;
printf qq{'$_' } for @players;
"
'Exit' 'Cassio' 'Enter' 'Iago' 'Othello' 'and' 'others'
'Cassio' 'Iago' 'Othello'
'Cassio' 'Iago' 'Othello'