for (qw(CamelCase AStringWithFTP AStringWithFTPOrSFTP)) {
  my @parts = /[A-Z](?:[A-Z]+|[a-z]*)(?=$|[A-Z])/g;
  print "$_ => @parts\n";
}

# outputs:
# CamelCase => Camel Case
# AStringWithFTP => A String With FTP
# AStringWithFTPOrSFTP => A String With FTP Or SFTP
[download]

use strict;
use warnings;

my @strings = (
   q{ThisIsACamelCasedString},
   q{AStringWithFTP},
   q{HeIsANASAAstronaut},
   );

my $rxCamel = qr
   {(?x)
       (?<=[a-z])(?=[A-Z])
       |
       (?<=[A-Z])(?=[A-Z])
   };

foreach my $string ( @strings )
{
    print qq{String: $string\n};
    my @words = split m{$rxCamel}, $string;
    for my $idx ( reverse 1 .. $#words )
    {
        if
        ( 
           $words[$idx] =~ m{^[A-Z]+$}
           and
           $words[$idx - 1] =~ m{^[A-Z]$}
        )
        {
            $words[$idx - 1] .= splice @words, $idx, 1;
        }
    }
    print qq{     $_\n} for @words;
}
[download]

Here's the output.

String: ThisIsACamelCasedString
     This
     Is
     A
     Camel
     Cased
     String
String: AStringWithFTP
     A
     String
     With
     FTP
String: HeIsANASAAstronaut
     He
     Is
     ANASA
     Astronaut
[download]

I think there will be too many corner cases for this task to be be achieved without some comparison with perhaps a dictionary list of acronyms.

I hope this is of use.

Cheers,

JohnGG

First, since there is no seperator, I would not first look to split, but rather to m//g in list context:

my $string = "ThisIsACamelCasedString";
my @split = $string =~ /([A-Z][a-z]*)/g;
print "Parts: @split";
[download]

(This version just skips any part that does not match /[A-Z][a-z]*/; season to taste.)

Dealing with "AStringWithFTP" is trickier, at least without a dictionary. The only heuristic I may suggest is to treat sequences of upper-case characters as a word of its own if in final position or if preceding an uppercase-lowercase sequence:

my $string = "AStringWithFTPAndHTTP";
my @split = $string =~ /([A-Z](?:[A-Z]*(?=$|[A-Z][a-z])|[a-z]*))/g;
print "Parts: @split";
[download]

That this is a heuristic is clear from noting that "ACString" is split as "AC", "String", and not as "A", "C", "String". But frankly, without a dictionary, I don't think there is a real solution.

Update: Oops, pasted the wrong code. Fixed now. Also added prints to make it easier to test that it really is the code I meant to post ...

Update2: Err, pasted the wrong right code too. Sorry, salva, I did not mean to steal it. /me *blushes* (Fixed now.)

Far easier and better to use [[:lower:]] and [[:upper:]] which will work with all lower and upper case characters.

Even with a dictionary, you cannot easily split out:

ACStringContainingTheStandardACVoltageForTheCurrentCountry = "110V";

You'd have to have a context sensitive system capable of understanding English for the really tough examples.

Sorry....meant to include

/[a-z]+|[A-Z](?:[A-Z]+|[a-z]*)(?=$|[A-Z])/g

as possible regex.

for('TheTaoOfProgramming','NowA more,problematic stringExample','Make 
+The Girl Dance - Baby, Baby, Baby_(Audio replaced, horizontally flipp
+ed!)_fmt35'){
    next if $_ =~ /\W/;
    $_  =~ s/_/ /g;
    my @split = $_ =~ /[[:lower:]0-9]+|[[:upper:]0-9](?:[[:upper:]0-9]
++|[[:lower:]0-9]*)(?=$|[[:upper:]0-9])/g;
    $_ = "@split";
}
[download]