For “alphanumeric” data?
For “alphanumeric” data (or alphanumeric edited data), you should be OK, if you simply take the default of USAGE DISPLAY for your data. In other words, the definition of COBOL insures (more or less) that such data is pretty portable and that each symbol in the PICTURE clause takes 1 byte of storage. HOWEVER, the only way to really insure that such data is completely portable is to: Use the Standard-1 (or Standard-2) Collating Sequence Use the Standard-1 (or Standard-2) CODE-SET for input/output files Not rely on the order of keys of indexed files However, in most environments, you should be fairly safe in using the “native” coding system for alphanumeric items – as long as you have a method of converting such data from environment to environment (for example – an upload or download system that converts from ASCII to EBCDIC) if this is now or may ever be an issue for your application.
