The latest Russian Soyuz spacecraft just returned from the international space station, arriving via an emergency “ballistic descent”. No one knew where they were for a half hour, during which they were lying in a burning field where car loads of locals showing up to help them out of their craft; not exactly the homecoming most of them envisioned I’m sure. MSNBC has an article detailing the multiple system failures that must have occurred for the emergency descent (autopilot), losing them (ground based radar), and not being able to talk to them (radio beacons on Soyuz) until one of the Cosmonauts stumbled out and phoned home on a satellite phone. While these are all valid concerns to have if you are planning on riding in one of these things, it is hard to criticize a program for a few non-fatal glitches when your country is operating the most dangerous space craft to ever fly. It shows the robustness of the initial design that it can suffer all these problems and still have a perfectly safe landing.
When approaching a difficult problem sound design choices are paramount. Intelligent initial design choices will save untold work and problems later on, brute force solutions are almost never desirable. While many of the ideas for the space shuttle were innovative and good, they were apparently to ambitious and didn’t mesh with reality.
This is just as true for smaller projects, people waste a lot of time trying to following initial faulty decisions instead of figuring out a more intelligent approach that would dramatically improve reliability, functionality, amount of time to develop, or all three. Work smart not hard.