In response to your REGION question, nope. REGION[.procstepname]=valueK|valueM
Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Charles Mills Sent: Tuesday, November 04, 2014 4:09 PM To: [email protected] Subject: Re: System vs. user ABEND codes Yeah, 23. I was thinking in base 22. <g> After I wrote it I thought that 23-bit ABEND codes were probably a bit much but 16 bits might have been an improvement on 12, especially since the hardware provides halfword support and lots of things are in halfwords. I have never worked on an octal machine but octal makes sense on a machine with 6-bit characters and words a multiple of 6 bits, right? My favorite octal "story" is that on UNIX to dump a file in hex you use the octal dump program (od with the -x option). Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Paul Gilmartin Sent: Tuesday, November 04, 2014 3:47 PM To: [email protected] Subject: Re: System vs. user ABEND codes On Tue, 4 Nov 2014 14:30:54 -0800, Charles Mills wrote: >I always thought it was the hex just sort of seemed "system-like" and >decimal numbers were, you know, for those COBOL types. <g> > >I always wondered why did they put two more or less mutually-exclusive >data in two different 12-bit fields? If they had devoted 11 bits to the >ABEND code and one bit to system versus user, we could have had ABEND >codes ranging up to S7FFFFF or U8388607. Whether that would have been >good or bad I will leave as an exercise for the reader. > (ITYM "23") For S0Cx, the bottom nybble is the hardware interrupt code. This provides some motivation for hex. But still, why decimal? In a CDC operating system, octal ruled. Job time limits were coded in octal numbers of seconds. The ROT was that 100 (octal) seconds was about a minute, and 10000 (octal) seconds was about an hour. At some point this impelled a naive colleague to ask "Are octal seconds bigger than decimal seconds?" Hey, it makes as much sense as KibiBytes and MebiBytes, abbreviated as K and M. And why not allow a simple unsuffixed decimal number, e.g. REGION=100000000? (Or does that actually work? I haven't tried it.) Can "G" (or "Gi") be far behind? -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
