On Wed, 29 Aug 2007 10:49:15 -0500, Paul Gilmartin <[EMAIL PROTECTED]> wrote:
>On Wed, 29 Aug 2007 15:02:27 +0000, Ted MacNEIL wrote: > >>This has been discussed many times. >>IBM's design choice is based on a simple premise: >>When do you convert the system symbols? >>On the submitting system? >>On the converting system? >>On the executing system? >>What about NJE/NJP? >> >Rhetorical questions: Why not, then, allow the programmer, through a >control statement or symbol qualifier, to choose among those three >alternatives? And why does NJE make a difference? (NJP is not in my >vocabulary). > >>Any choice can/will be wrong, so why open the can of worms? >> >Many programmers feel that even if IBM were to choose one of the >alternatives above, they would benefit; it would be right for them. >Those to whom the facility would be no benefit would be not be harmed >by it if it were provided: they could continue to use batch JCL with >no system symbol references as they do today. In practice, some users do not read the documentation, and even if they did they might not realize the problems lurking when interpretation (or conversion?) and execution happen on different systems. Then one day it would matter, and they would call the Support Center, sure that IBM code was broken, because it "always worked before." Perhaps they had only 1 system, and could not have problems, and then one day added a second system. Then suddenly some number of their jobs would break. And again they would call the Support Center. And very likely one of those jobs that broke would be critical, for some large customer, and cause a big problem for IBM because we implemented something that let them shoot themselves in the foot. -- Walt ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
