On Mon, 22 Nov 2010 10:57:34 -0400, Clark Morris wrote: >> >>The biggest challenge is likely to be the limited resource in TSO (the IKJ >>component), but I'll see what I can do... It certainly wouldn't hurt to >>resubmit >>it as part of the SHARE requirements clean-up. >> >>W. Kevin Kelley -- IBM POK Lab -- z/OS Core Technical Development > So it's permissible to market defective software provided the excuse is resource constraint? In this case, the defect is of significant consequence to the customer, and a remedy wouldn't seem technically difficult to implement.
>In general, if a message doesn't specifically tell you what is wrong, >it is useless. Back in the 1980s I justified the effort in going to >MVS SP 1.3.3 (from 1.3.2 IIRC) because at that level the error message >finally told people WHICH symbol was not defined in the procedure. As I think you're referring to a notorious JCL message. It's still wrong. Regardless of what it says, it means that the symbol is defined but not _applied_ in the procedure. Why is this even considered an error? In contrast, assembler doesn't regard it as an error if a symbol is defined but never applied. This merits a warning or informative, not an error message. >a semi-retired person who can't justify the cost of going to SHARE (No >contracts or revenue and no system to apply information to), I would >urge a general requirement that any message must show the erroneous >data (duplicate symbol, etc.) in the message. Further the requirement >should state that any message that does not contain the specific >information needed to isolate the problem highlighted (actual >erroneous value, statement number, etc.) is aparable and maybe >automatically escalates at 6 month intervals. > A regular contributor to this list lauds JCL's syntax for its similarity to assembler syntax. A couple things that JCL could further adopt from HLASM: o Place JCL error messages inline, as HLASM does, so the programmer would be spared the need to swap between JESYSMSGS and JESJCL. (In some cases, as the example above, the earliest the error could be detected and the message issued in a single-pass translator is at the PEND statement. I'd accept that.) o Identify, as HLSAM does, the source line (relative if unnumbered) and member (if from PROCLIB or JCLLIB) of the offending statement. This would enable the programmer more easily to locate the line using an editor. The "L" in "JCL" stands for "Language". JCL should be subject to at least some of the design principles of other language processors. -- gil ---------------------------------------------------------------------- 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
