Abstract: A Universal NLS ("UNLS") paradigm is disclosed to test the integration of localized multi-national software application programs. A user prepares a UNLS test script using a plurality of UNLS operators. The UNLS operators identify non-ASCII and multibyte characters, as well as provide other functions. The UNLS test script created by the user is provided to a UNLS translator which translates the UNLS test script into any one of a number of desired target languages for a selected target computer platform. The UNLS translator translates the UNLS script by translating ASCII character strings in accordance with the UNLS operators. Each ASCII character in an identified string is translated to a predefined "problem" character in the target language. A maximal coverage lookup table maps ASCII characters to problem characters. A "problem" character is defined as a character in the target language which has historically presented a greater likelihood of error when used by a localized computer program.