Goran said:
Best response so far! Except that better word for 3) is "reporting"
option.
Indeed. I realized I had written the wrong category the next day whereas
usually I'm all over this stuff. In a nutshell, to get errors under
control, one has to understand and address all of the following:
identification/categorization, detection, propagation, handling,
reporting. The first step (identification/categorization) is analysis of
what kinds of errors the program/system will encounter and helps plan the
other stages in the "error lifecycle".
I probably subliminally avoided saying "reporting" in my OP because some
people use it to mean propagation. Perhaps "notification" would be a
better word, but maybe not since that is a means of reporting just like
logging is... oh well.