You can design for unknowns, e.g. whether Huygens lands in a liquid versus on a solid surface. The designer simply determines the worse case scenario and design the vehicle to survive that. However unknown-unknowns are a real gottcha.

What's an unknown-unknow? Accidently designing an electrical switch so it can be installed backwards is a classic example (they get it right with the prototype and then screwup with the flight vehicle). Proper mechanical design requires that assembly only be possible the right way. It's relatively easy to do this with an automobile where thousands of cars are made and the design can be continually tweaked. However with a single spacecraft, this sort of problem is hard to defend against.

Huygens is Europe's first interplanetary lander (Beagle 2 wasn't really an ESA project). There are lots of opportunities for unknown-unknows when there is almost no flight heritage. JPL got it right with Viking but there were lots of problems and close calls before Viking. In the beginning there was the Ranger program (a Moon crash lander). Many Rangers failed because JPL needed time to get their act together. There were still issues with the early Surveyor program (the Moon soft lander) but finally JPL got their act together. That heritage of success was passed onto the Viking program.

One advantage that ESA had with teething issues was in NASA looking over ESA's shoulder and helping out whenever NASA saw a problem, e.g. the CN radiative heat flux SNAFU.

It's vital that Huygens succeed. I wish the United States had built Huygens and the Europeans had their learning experience with a lunar lander. However there was not enough money for the U.S. to do Cassini as a completely American spacecraft (it barely survived politically as a joint European/American spacecraft).

Gary