Sorting
String sorting is used in a number of places in the Macintosh Operating
System (for example, in a standard file dialog box) and in applications (for
instance, spreadsheets). When performing such sorting, it is important to
order strings in the manner expected by the user-that is, according to the
rules of the language and region for which the system is localized. The
two strings and indicate whether the first should be sorted before, after, or at
the same place as the second string. For details, see
Using the International Utilities Package Routines.
Sorting or comparing strings can be an extremely intricate operation. Subtle
issues like expansion, contraction, ignorable characters, and exceptional
words may be taken into account. Sorting cannot be done properly by a simple
table look-up, even for such straightforward cases as English. Sorting depends
not just on the script, but on the individual language. While broad similarities
in sorting exist between languages that share the same script, definite
variations between languages must be taken into account.
international resource 'itl2' have long provided for many sorting issues,
including primary or secondary order, expansion, contraction, and ignorable
characters. With system software version 7.0, several new sorting
capabilities provide support for systems with multiple installed scripts and
language capabilities.
• You can sort strings in different scripts and languages.
• A new international resource, 'itlm', indicates the preferred sorting
order for scripts, languages, and region codes, and indicates how to map
region codes to languages and language codes to scripts. See
• You can explicitly specify the handle of the resource to be used for
sorting. This is helpful for multilingual systems. See the routines
• 'itl2' and 'itl4' resource handles for all active scripts are cached by
application-supplied resources can be used. See the section entitled,
Primary or Secondary Order
Sorting order is determined by a ranking of the entire standard Roman
character set. This ranking can be thought of as a two-dimensional table. Each
row is a class of characters-for example, all of the forms of uppercase and
lowercase A with and without various diacritical marks. The characters are ordered within the row, but that ordering is secondary to the primary
ordering of the rows themselves. For example, all of the forms of A precede all of the forms of B, as follows:
B < b
Primary sorting characteristics denote a strong ranking; if any primary
differences are present, all secondary differences are ignored. For instance,
only primary sorting is needed to determine that abc precedes bc. Secondary
sorting characteristics indicate that if certain differences are present, a
second pass is made that introduces a weak ordering. Here's an example:
abc < åbc
Expansion
A single character may be sorted as if it were a sequence of characters. First,
the single character is expanded; then the primary sorting occurs based on
this expansion. In the secondary sorting, the characters are recombined. For
instance, ä in German may be sorted as if it were the two characters ae, as in
this example:
bäk < baek < bäks
Contraction
A sequence of characters may be sorted as a single character. For instance, ch
in Spanish may be sorted as if it were one character that sorts after c, as in
this example:
czar < char< dar
Ignorable Characters
Certain characters should be ignored unless the strings are otherwise equal.
In other words, they have no effect on primary sorting, but they do influence
secondary sorting. Examples of ignorable characters in English are hyphens,
apostrophes, and spaces. Here is an example of how a hyphen influences
secondary sorting:
blackbird < black-bird < blackbirds
Exceptional Words
Sometimes the sorting order changes drastically for special cases. For
instance, when words are understood to be abbreviations, the strings are
sorted as if they were spelled out.
McDonald < Mary McDonald is treated as MacDonald ¯
Thus, MacDonald < Mary
St. James < Smith St. is an abbreviation for Saint ¯h
Saint James < Smith
Easy Step < Easy St. St. is an abbreviation for Street ¯n
Easy Step < Easy Street
Such cases require a direct dictionary look-up and are not handled by the
Macintosh Script Management System. Note that abbreviations are
context-dependent; for example, St. may denote Saint or Street, depending on
the meaning of the adjacent text.