I believe what I said was "I make crashbugs on production machines go away". This is so that they don't crash anymore. Often, this means any number of the following
- the OS is unstable
- I have a VT110 connection
- I have exactly 23 lines and 80 columns
- I'm probably over a crappy connection
- the CPU is going nuts
Basically, I have to take the errant application down NOW and make sure it doesn't come back up till I can fix it, preferably no earlier than Monday morning. And, of course you fix it in dev, write the appropriate tests, run it through QA, then put it back on prod. It's not a matter of hacking till it works in production - that's called DIP and is a very apropos acronym. And, shutting the machine down isn't always an option, either. You don't always have the luxury of running just one app on a given machine.
My criteria for good software:
- Does it work?
- Can someone else come in, make a change, and be reasonably certain no bugs were introduced?
| [reply] |