Hmmm... Good point. Well, I think in a case like this, I'd still like to try to take advantage of some degree of factoring:

c:\@Work\Perl\monks>perl -wMstrict -le
"use charnames ':full';
 ;;
 use Test::More tests => 2;
 ;;
 my $in = qq{\N{LATIN SMALL LETTER C WITH CEDILLA}};
 ;;
 like ($in, qr{\A[A-Za-z0-9](?:[A-Za-z0-9_.-]*?[A-Za-z0-9]|)\z}, 'kcot
+t');
 ;;
 my $alpha      = qr{ [A-Za-z0-9]    }xms;
 my $alpha_plus = qr{ $alpha | [_.-] }xms;
 like ($in, qr{ \A $alpha (?: $alpha_plus* $alpha)? \z }xms, 'Anomalou
+sMonk');
"
1..2
not ok 1 - kcott
#   Failed test 'kcott'
#   at -e line 1.
#                   'ô'
#     doesn't match '(?^:\A[A-Za-z0-9](?:[A-Za-z0-9_.-]*?[A-Za-z0-9]|)
+\z)'
not ok 2 - AnomalousMonk
#   Failed test 'AnomalousMonk'
#   at -e line 1.
#                   'ô'
#     doesn't match '(?^msx: \A (?^msx: [A-Za-z0-9]    ) (?: (?^msx: (
+?^msx: [A-Za-z0-9]    ) | [_.-] )* (?^msx: [A-Za-z0-9]    ))? \z )'
# Looks like you failed 2 tests of 2.
[download]

(Both the  /a modifier and the  (?a) embedded modifier seem to work with the original
    qr{ \A [[:alnum:]] (?: [[:alnum:]_.-]* [[:alnum:]])? \z }xms
regex to suppress extended Unicode matching, but I don't fully understand the interaction of this and related flags with POSIX character classes. And it's one more modifier to remember!)