I was hoping you'd come by. Welcome to the Monastery.

This is 100x better. Well-done. Variable names can be improved, but I can actually scan this code and understand the basic intent. The biggest improvement now would be to skip intermediate variables that only exist once. So, something like:

use strict;
use warnings FATAL => 'all';

use List::Utils qw( sum );

my $input = do {
    local $/;
    <>;
};

my %letters;
$letters{$_}++ for grep /\w/, split //, $input;

my $total = sum values %letters;

for ( sort keys %letters ) {
    printf "%s\t%d\t%.3f\n", $_, $letters{$_}, $letters{$_}/$total;
}

print "$total characters\n";
[download]

The big differences there are:

• Removing intermediate variables
• Scoping the localization of $/
• Using a module to do the summation
• Chaining functions in the grep split.

This would be considered more "perlish". Remember - readablity also includes conciseness. This is why well-written Java is less readable than well-written Perl.