Globalize your On Demand Business

Messages in the UI usually employ substitution variables to contain variable information. When a message is retrieved dynamically, the current information replaces the substitution variable. However, when a message is translated into another language, the position and order of the substitution variables may have to change in order to meet the syntax requirements of the target language.

Guideline A6

Substitution variables used with messages must be uniquely identifiable, so that when they are repositioned and reordered, there is no requirement to change the program logic.

Example: The following C segment retrieves a message string from an external message file and prints it to the standard output. Depending on the user's preference, the message string is retrieved from either the English or the German message file.

int iReturnCode; /* return code */
char cFunctionName[33]; /* function name */
char *pcFormat; /* printf(...) format string */
nl_catd descCat; /* message catalog descriptor */
:
:
/* Retrieve the printf(...) format string */
pcFormat = catgets( descCat, MSG_SET_ID, MSG_ID, NULL );
if ( pcFormat != NULL ) /* success */
(void)printf( pcFormat, iReturnCode, cFunctionName );

The message in the two external message files is as follows:

When the English format string is retrieved, %d formats the variable iReturnCode as a decimal number, and %s formats the variable pcFunctionName as a character string. When the German format string is retrieved, however, not only are %d and %s erroneously used to format the variables pcFunctionName and iReturnCode, but the resultant message is incorrect because you are misinterpreting the function name as the return code, and the return code as the function name.

The ISO/IEC international standard on the C programming language recognized the importance of this problem, and introduced the %n$ construct to explicitly bind a format flag to a specific argument. Using this feature, the message may be rewritten as: