Discussion

This section describes C language functions that you can use to write UNIX applications that will process input and generate output in a user's native language or cultural environment. It shows you how to use these functions and some associated commands to create programs that make no assumptions about the language environments in which they will be run, and so are portable across these environments.

The basic idea behind the internationalization interface is that at any time a C program has a current ``locale'': a collection of information on which it relies for language- or culture-dependent processing. This information is supplied by implementations and seen by the program only at run time. Because the information is stored externally to the program, applications need not make -- and should not make if they mean to be portable -- any assumptions about

• the code sets used by the implementation in which they are executed. The 7-bit US ASCII code set, for example, cannot represent every member of the Spanish character set; the 8-bit code sets used for most European languages cannot represent every ideogram and phonogram in the Japanese language.

• the cultural and language conventions of the application's users. The same date is formatted in the United States as 6/14/90, in Great Britain as 14/6/90, in Germany as 14.6.90. Similar problems arise in formatting numeric and monetary values. By language conventions we mean, for instance, that the sharp s in German is collated as ss; the character ch in Spanish collated after all other character sequences starting with c.

• the language of the messages in which the program communicates with the user. Interactive applications in an English-speaking setting usually will query users at some point for a yes or no response; in a German-language setting the responses will be ja or nein; in a French one oui or non. Program error messages will differ much more widely than that across languages: File not found, Fichier inexistent, and so on.