Quote:
Originally Posted by RogerEvans
Error codes were incomprehensible except for the closing phrase which always said ' probable user error'. IBM JCL made vi editor commands look like nursery rhymes and the floating point exponent range was so limited that you were forever chasing underflow / overflow errors.. It did however have a mind boggling 256kB of memory and ran at 1MIP (360/195).
|
Ah yes, those error codes. Something like "Byte-level inconsistency in xxnvpqjt_yyy when executing sqdlprj", followed by a dump of the registers. The only vaguely comprehensible one was the one which ended "probable cause: the line printer has run out of paper". Only operators had access to the line printer so that didn't help much. However much JCL you had there was always at least one statement missing. Our 360/65 had a whole MB of memory but divided into partitions so you could never get more than about 50kB. I had one program which required 500kB. It took them a month to schedule setting up the partition and then they had lost the input tape.
Then there were the spoof assembler instructions such as WWLR (write wrong length record), PED (punch and eject disk) and CFS (catch fire and stop).