For “less portable numeric” data?
USAGE COMP (or COMPUTATIONAL) is a Standard definition (unlike COMP-1, COMP-2, … COMP-n or COMP-x). HOWEVER, what it means is totally implementor defined. There is a common MISCONCEPTION that the Standard requires USAGE COMP to be the “most efficient” USAGE for numeric fields (I have even seen this quoted in some manuals). However, this requirement certainly does NOT exist in the current ANSI/ISO Standard. Nevertheless, often (but definitely NOT always) this USAGE is the most efficient for numeric fields used in frequent arithmetic operations.
