Hard to have an opinion without knowing the details but from my little experience I feel like the reason why these big project fail is not so much because they are too complex, I am sure a good (not excellent) developper could come up with an algorithm and data structure generic enough to accomodate all corner cases. But because these big projects start from a preexisting solution (PeopleSoft in this case), that wasn’t designed for this particular complexity, so it requires orders more energy to go around its limitations, and the people staffed by the likes of IBM on these projects are usually project managers with little to no programming skills and little domain knowledge. So you end up with a disaster for things that aren’t that complicated (custom shifts for coast guards and prison guards doesn’t look like a hard computer science problem).