Re: Question about legacy coding style

Craig Pelkie wrote:

This is in some old code we are looking at:
C U1
CANNU1 MOVE X2DAT X2DAT 6 0

. . .

It looks like an obscure way to define a field.

Yes, I would say that a MOVE statement copying a variable onto itself, conditioned so it can never execute, is a ridiculous waste of instructions (hopefully, the compiler is smart enough to squeeze them out), when one can simply define the field the first time it appears in the RESULT column.

But statements like MOVE, ADD, Z-ADD, IFEQ, &c., and defining variables inline, are not only inherent to RPG; they are far more idiomatic to RPG than EVAL or IF, and far more idiomatic than defining variables in D-specs. That doesn't mean I don't use EVAL, or IF, or that I don't define variables in D-specs (let's face it, complex calculations and conditions are almost impossible without the former, and if you need a variable to be of a non-default type, you HAVE to do the latter); I just recognize them as non-idiomatic (albeit not nearly as non-idiomatic as "/free" is).

Which brings up a point I keep having to make: if you don't use The Cycle, and you don't like RPG syntax, then why are you bothering to use RPG at all? There *are* other languages. (And if there had been more demand for, say, PL/I, it would have probably been ported to ILE.)

--
JHHL
Proudly a Mixed-Language Programmer Since 1980