I was just told a story about Lotus Notes (owned by IBM Corporation). Let's call the person who told me this 'John'.
John was having problems with Lotus Notes about once a week where it would simply crash. Well, it didn't exactly crash crash. Apparently Lotus Notes utilizes WER (Windows Error Reporting) or an equivalent:
So when the application crashes, it sends a crash dump of the program to John's IT Helpdesk (or IBM's Helpdesk - I'm not sure how the system works). John doesn't really care. Except John got curious because the application says to contact the corporate IT helpdesk about the issue. So, John calls the corporate IT helpdesk one day and describes the problem. Let's say the IT helpdesk person's name is Sue. Sue asks the following question, "Does Lotus Notes crash on Tuesdays right around lunch time?" Thinking for a moment, John says, "Not sure about the day, but that sounds like it might be the right time of day." Sue then sends John an attachment in an e-mail with a button labeled "Fix Tuesdays at noon". John clicks the button and the crashes go away. In actuality it should have said, "Fix Tuesdays at 11:45".
At this point every last programmer out there is asking, "What the..?" and "Who the..?" How do you introduce a bug into a program that causes it to crash exactly at 11:45 a.m. every Tuesday and then have a binary patch for just that bug? I can think of only one way: Hardcoding it (unless the code base for Lotus Notes is so obtuse that such things just happen). Intentionally writing crash bugs is a terrible approach to software development. Possibly criminal. I mean, what if a life support system depended on Lotus Notes. Every Tuesday at 11:45 there would be a disasterous rise in the number of fatalities.
Some people live on e-mail. Literally. Crazy people.