For All Countries (Country-Independence)

As NLS support is getting more and more important, you may make your program country-dependent (including code-page-, language- or culture-dependent), but such country-dependency should be minimized. Otherwise, you must test your program for each country, although most of your functions are common to all countries. Consider the following:

• Minimize hard-coding of country- or code-page-unique information in your program. Utilize query functions of the system, for example, to get country-unique formats of date and time.

• When you use API's that have code page parameter(s) such as GpiCreateLogFont, do not hard-code code page ID's but get the required ID's from the system,

• If no appropriate system functions are available, isolate and encapsulate country-unique information to an external file for each country and include the country code in the file name, for example, "A081.TBL" for Japan, whose country code is 081. Select the appropriate file querying the current system's country code. This approach eases your future enhancement to add supported countries. Just add new files; no modification to your executable modules is needed.

• A worldwide product need not have the differences in most cases and it is not recommended to do so unless there are some specific requirements. If some functions of your product have these differences, such functions should be tested for each DBCS language.

• Depending on your functions, you may need to support DBCS alphanumeric character handling (such as monocasing, translation between SBCS and DBCS alphanumeric characters). If so, note that each DBCS alphanumeric character is at the same code points among DBCS countries' code pages in EBCDIC DBCS but not necessarily among PC (ASCII) DBCS countries' code pages. See also DBCS Alphanumeric Characters.

If your program becomes country-dependent, let users know what countries your program supports. If your program does not have country-dependencies, you can say that your program supports all the countries that OS/2 supports.

Note: When you document your supporting countries (or languages), you should distinguish NLS support in functions from NLS translation support since you may support NLS functions but may not provided translated PII (or vice versa).