On Fri, 29 Aug 2025 16:17:35 -0600, Robert Raicer <rjr80...@gmail.com> wrote:
>And, YES, the code had undergone a successful code review. You can't hold people responsible for what they don't know. This thread demonstrates that most people are naive about mixed case. Ask yourself why "regular expressions (regexp)" are pivotal to Unix. Consider the complexity of all Unix messages caused by mixed case (command output, syslog, snmp, logs, error logs, ...) and how you would process those messages in any language (e.g. REXX). Consider the following SNMP message as an example. Without REGEXP, automating this message would be impossible but it comes at a very high price. It would be a mess in REXX without regexp. 02-28-2014 09:00:14 Auth.Notice 192.168.34.31 Feb 27 23:00:24 : 2014/02/27 23:00:24 EST,1,545702,Probe - MAP IP To MAC Failure,0,28,,Switch,192.168.34.1,,Failed to read IP address mappings from device Switch. 1. Without a msgid, you must determine the pattern. 2. With mixed case, you don't know where case can be ignored. 3. You can't uppercase or lowercase the message because case may be important 4. REGEXP is expensive because it of the manipulation required (e.g. uppercase, lowercase & ignore case). Automating Unix events sucks because of mixed case which needs regexp. Automating MVS events is easy because of message ids allow us to either ignore case or know where case is important. >This was not a Severity-1 situation (the customer was very willing Sev1 occurs when a customer believes they are in an immediate intolerable situation. You recognized within 1 minute the cause of the problem. It's not a good look but you could easily explain the problem and apologize. >This is NOT intended to be an argument against using mixed case in I agree. I'm saying we're being cavalier in saying mixed case doesn't deserve respect for its current impacts.
