Capitalizing words by uppercasing only the first letter

Question

In Perl, there's the ucfirst function.

Is it this the equivalent to this:

sub uppercase {     
    my ($W) = @_;       
    $$W = uc(substr($$W,0,1)).substr($$W,1);        
}

Does it matter across Perl version?

---

Contextualizing the question, https://github.com/moses-smt/mosesdecoder/pull/206/files#diff-876e51db2a1ab71c1ae736182d1e5e04R63 ,

Previously, the usage of uppercase is as such:

sub process {
    my $line = $_[0];
    chomp($line);
    $line =~ s/^\s+//;
    $line =~ s/\s+$//;
    my @WORD  = split(/\s+/,$line);

    # uppercase at sentence start
    my $sentence_start = 1;
    for(my $i=0;$i<scalar(@WORD);$i++) {
      &uppercase(\$WORD[$i]) if $sentence_start;
      if (defined($SENTENCE_END{ $WORD[$i] })) { $sentence_start = 1; }
      elsif (!defined($DELAYED_SENTENCE_START{$WORD[$i] })) { $sentence_start = 0; }
    }

    # uppercase headlines {
    if (defined($SRC) && $HEADLINE[$sentence]) {
        foreach (@WORD) {
            &uppercase(\$_) unless $ALWAYS_LOWER{$_};
        }
    }

But it seems like replacing &uppercase(\$WORD[$i]) and &uppercase(\$_) with ucfirst(\$WORD[$i]) and ucfirst(\$_) is different.

Answer 1

ucfirst is not equivalent to the following:

sub uppercase {     
    my ($W) = @_;       
    $$W = uc(substr($$W,0,1)).substr($$W,1);        
}

ucfirst is mostly^[1] equivalent to the following:

sub ucfirst {     
    my ($W) = @_;       
    return uc(substr($W,0,1)).substr($W,1);        
}

If you wanted to rewrite uppercase in terms of ucfirst, it would look like this:

sub uppercase {     
    my ($W) = @_;
    $$W = ucfirst($$W);    
}

uppercase(\$string);

That means that if you wanted to eliminate uppercase entirely, you'd replace

uppercase(\$string);

with

$string = ucfirst($string);     # Correct

You tried using

ucfirst(\$string);              # Wrong

---

1. ucfirst actually does a better job of handling more esoteric characters such as U+01F3 LATIN SMALL LETTER DZ ("dz").

Answer 2

In Perl, there's the ucfirst function.

Is it this the equivalent to this:

Let's find out...

$ cat testuc
use strict;
use warnings;
use Test::More; 

sub uppercase {
  my ($w) = @_;
  return uc(substr($w, 0, 1)) . substr($w, 1);
}

my @tests = qw[foobar Foobar FOOBar fOObar fOObAR FOOBAR];

for (@tests) {
  is(ucfirst($_), uppercase($_), "correct for $_");
}

done_testing;

$ prove -v testuc
testuc ..
ok 1 - correct for foobar
ok 2 - correct for Foobar
ok 3 - correct for FOOBar
ok 4 - correct for fOObar
ok 5 - correct for fOObAR
ok 6 - correct for FOOBAR
1..6
ok
All tests successful.
Files=1, Tests=6,  0 wallclock secs ( 0.04 usr  0.03 sys +  0.03 cusr  0.04 csys =  0.14 CPU)
Result: PASS

So, yes, it looks like they're the same thing (at least for my rather limited set of tests).

I'm using Perl 5.26.1 - but I think this will work fine for all Perl versions back to at least 5.10.

Update:

I made a silent edit to your code which I forgot to mention. You code originally worked on a reference to a scalar, but I changed it to work on a scalar ($W instead of $$W). I assumed that would be a harmless substitution.

But now you've shown us your change in context and I can see what's going on.

You had:

&uppercase(\$WORD[$i])

And you changed that to:

ucfirst(\$WORD[$i])

This doesn't work as ucfirst() doesn't change its argument; it returns the changed value. So you actually want:

$WORD[$i] = ucfirst($WORD[$i]);

That will then work as expected (modulo the Unicode character issues mentioned in other answers.

Your whole loop can be simplified if you move away from the C-style for loop.

for my $w (@WORD) {
  $w = ucfirst($w) if $sentence_start;

  if (defined $SENTENCE_END{ $w }) {
    $sentence_start = 1;
  } elsif (!defined $DELAYED_SENTENCE_START{ $w }) {
    $sentence_start = 0;
  }
}

Answer 3

The functions are not equivalent because of some Unicode details, especially dealing with digraphs.

For example, the Hungarian language uses the digraph "DZ", which is considered a single letter of the alphabet, and so can optionally be represented using the Unicode code points:

• U+01F1: DZ
• U+01F2: Dz
• U+01F3: dz

So

my $text1 = "\x{1f3}won";
my $text2 = $text1;
$text1 = ucfirst($text1);
uppercase(\$text2);
print($text1 eq $text2 ? "same\n" : "different\n");

prints "different".