I'd like to add to the previous post. The Y2K problem is controversial
precisely because we don't know whether it will be "a snap" or not. Some of
the errors can be subtle - for example, lots of companies seem unaware that
2000 is a leap year. It will take a lot of hard work to ensure operations
are going to make the transition. Things like computers and shrinkwrapped
software are easy. Instruments and other equipment with embedded
microprocessors and the code to run them are a nightmare. Therefore,
reports and data based on that equipment may be at risk. And something like
90% of all microprocessors built since 1980 are NOT in PCs or workstations.
They're embedded as control subsystems in other devices.

"Your potential problem areas are in executable software or
research output that depends on someone else's software"

Agreed. Its vital that organizations examine their whole information supply
chain for Y2K, not just their internal operations.  Executable software
must be tested thoroughly, and there is no automated tool that will do it
all for you.