On Wed, 11 Mar 2020 22:09:28 +0000, Jesse 1 Robinson wrote: >APAR for failure to follow doc. Gotta be careful though. Could turn into a doc >APAR. ;-( > When that happens, I truly wish I could see the unpublished design specs on which both are based. Yes, I'm suspicious that they took the feckless course.
I was once confronted by a True Blue user who discovered a discrepancy in a program I maintained, requesting that I adjust the doc accordingly as he had come to expect. I recalled my (mental) design notes and concluded that the doc matched my intent and repaired the program, disregarding my user's expectation. Besides, it was easier to repair the bug than to fully describe the chaotic behavior he had encountered. Worst case: a defect has gone unreported so long and so many users have accommodated it that fixing it would be more disruptive than tolerating it. IBM is too often in that position. IBM once repaired a long-standing bug in assembler code generation. I mentioned this to our project manager who decreed that all code in our product be re-certified lest it exhibit an undesirable behavior change. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
