ADD 1 TO A.

COBOL supports many ways of storing and defining numeric variables. Numeric values as expressed in the DATA DIVISION (where all 😎data is defined) can carry many forms from things like BINARY and BINARY-LONG to “$$$$$$$$9.99”, “99999999999.99” or just “9999999999999”; native binary versus edited numeric for display, numeric with precision, and just numeric. And in modern (ISO) COBOL we get into all of that plus National (8-bit code page) and multi-byte character sets (16 and 32-bit characters for UTF). Oh, I forgot to leading, trailing or no sign designators.

The 1 isn’t the problem! 😉 So we’re (Bob’s) working through from the simplest, improving performance on that. Data format conversion is one of the foundation capabilities of a COBOL compiler and it supports more data formats, explicitly, than any other language.

This is a critical aspect of GCC COBOL’s (gcobol’s) performance improvement. Our scan of some code available to us (around 30M lines of COBOL not counting copybooks) says that the most often used stateents (like “MOVE A to B” and “IF C EQUALS D”) all involve the probability of such conversions.

Thanks for reading the inside-COBOL babble! 😎