[Perl] How to do it better

There are some really helpful people in the Perl community. I advertised the original posting on comp.lang.perl.misc and received some very useful responses from John W. Krahn and Michele Dondi, as below.

With respect to the rules, John wrote:

# Rule 1. Remove all 's' sub rule1 { ( my $arg = shift ) =~ tr/s//d; return $arg; } # Rule 2. Sort the characters of the word into alphabetic order sub rule2 { return join '', sort split //, shift; }

John> Why [were] you converting the '|' character and the 'e' character to 'e'?

# Rule 3. Convert all vowels to 'e sub rule3 { ( my $arg = shift ) =~ tr/aiou/e/; return $arg; }

John> The . character class matches a lot more than just letters, or did you really mean "replace any first character except newline with 'n'".

# Rule 4. Replace the first letter with 'n' sub rule4 { ( my $arg = shift ) =~ s/\A[[:alpha:]]/n/; return $arg; }

John> The . character class matches a lot more than just letters.

# Rule 5. Drop the last letter sub rule5 { ( my $arg = shift ) =~ s/[[:alpha:]]\z//; return $arg; } # Rule 6. Replace letter pairs with 'ow' sub rule6 { ( my $arg = shift ) =~ s/([[:lower:]])\1/ow/g; return $arg; }

Then, with respect to the string eval() of each rule, John said, "Ouch! Use a dispatch table instead of string eval()."

my %rule = ( 1 => sub { ( my $arg = shift ) =~ tr/s//d; return $arg; }, 2 => sub { return join '', sort split //, shift; }, 3 => sub { ( my $arg = shift ) =~ tr/aiou/e/; return $arg; }, # more rules here ); # and later on, use rules $res = $rule{ $j }( $arg );

At this point Michele Dondi chipped in with, ">my %rule = (
>  1 => sub {
>    ( my $arg = shift ) =~ tr/s//d;
>    return $arg;
>    },
>  2 => sub {
>    return join '', sort split //, shift;
>    },

Since the keys are numbers, an array may be appropriate."

So there you have it, better rules, anonymous subroutines, and hashes. Powerful stuff.

I wonder how Tcl and the other languages would have done it. Or Protium, for that matter. Any takers?

© Copyright Bruce M. Axtens, 2008