Programming Perl, 4th Edition

A Case of Mistaken Identity

If you only know ASCII, nearly all your assumptions about how text behaves with respect to case will be invalid in Unicode. In ASCII, there are uppercase letters and lowercase letters, but in Unicode, there is also a third sort of case, called titlecase. This isn’t something we make much use of in English, but it does occur in various other writing systems derived from Latin or Greek.

Usually titlecase is the same as uppercase, but not always. It’s used when only the first letter should be capitalized but not the rest. Some Unicode codepoints look like two letters printed side by side, but they are really just one codepoint. When used on a word that’s supposed to have only its first part capitalized but not the rest, the titlecase version only capitalizes the appropriate part. These mostly exist to support legacy encodings, and today it is more common to use codepoints whose titlecase map produces two separate codepoints, one each in uppercase and lowercase. Here’s an example of one of the legacy characters:

use v5.14;
use charnames qw(:full);
my $beast = "\N{LATIN SMALL LETTER DZ}ur";
say for $beast, ucfirst($beast), uc($beast);

That prints out “dzur”, “Dzur”, and “DZUR”, each of which is only three codepoints long.

Some letters have no case, and some nonletters do have case. Lettercase is comparatively rare in the world’s writing systems. Only eight out of Unicode v6.0’s nearly 100 supported scripts have cased characters in them: Armenian, Coptic, Cyrillic, ...