One morning, your head programmer runs in to your office wildly waving a handful of papers above his head. He slams it down on your desk and blurts out, "We have a problem."
It turns out that there was an obscure bug in KAREL, hiding beneath layers of code. Under certain conditions, KAREL will enter into a constant reboot cycle, bringing the whole hospital to a stand still. Your staff tells you that the chances of ever encountering this bug, though, is about once every twenty years, far longer than KAREL was designed for. Additionally, it is not a hard matter to avoid the KAREL bug, but it would require retraining the hospital staff and reprinting the manuals to inform them of how to avoid it and what to do if KAREL ever crashes.
You are faced with the following decision:
You can send the code back to the programmers and tell them to weed out and rewrite the faulty code. This would take time, but at least you can be more confident about KAREL's reliability.
You can retrain the hospital staff and reprint the manuals to inform them about the bug. This will cost money and will also delay the release of KAREL. The bug will still be in your final version, but at least the hospital staff will know it's there. In the meantime, you can work on a patch for the bug and release that whenever it's ready. The downside is that it will be a long time before you will ever be able to start work on a competitor for LOGO the Medic.
You can also ignore the bug. Your engineers said that there is almost no chance that it will ever come up in normal operations, so you can probably get by for at least a few years without major incident. You expect to have KAREL II, the market version of KAREL designed to compete with LOGO, out in a few years, anyway. You figure you can just upgrade the hospital's version of KAREL to KAREL II at that time.
What do you want to do?
