On 2017-05-05, at 10:05, Jesse 1 Robinson wrote: > IBM's reluctance to provide system symbol support in *batch* is well > motivated. From the beginning of system symbols, they were honored only in > STC and TSU. In batch they were ignored. The problem with batch is that a > number of key factors affecting symbols are unpredictable. So when and how > would a symbol be set? > > At submit time? > At conversion time? > At NJE transmission time? > At execution time? > Make it an option; give the programmer control.
> ... > What if a system symbol happens to clash with an existing JCL or PROC > variable? > And here, we have the horrible, irreparable design error in JCL syntax. Reference to an undefined symbol should have *always* been a syntax error. Otherwise, JCL or PROC variables should dominate system symbols. There should be a construct to restrict scope of symbols defined in JCLLIB members so they don't clash with symbols in open code. I hate JCL! -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
