> - Legacy code. Bob was the only one around when this code was written and it hasn't been touched in years, now it's breaking and no one understands why. (In an ideal world, there wouldn't be knowledge silos like this, but there always are)
Not sure if valid, if some legacy code was written a few years before, probably even author will have no clue what's wrong.
The author might have a higher chance of figuring it out. This is playing the likelihood game though. In perfect world, the code would be tight and small enough that you wouldn't take long to understand it, and then the complex components (there are always a bunch of these) are well tested and documented. Especially rationale for design decisions or implementation quirks.
Not sure if valid, if some legacy code was written a few years before, probably even author will have no clue what's wrong.