Because NASA's planetary missions on the whole have been so incredibly successful, many of us outside the process can forget how incredibly complicated it can be. A just released audit (see post on the journal Nature's blog, Space News, and Space Ref) gives examples of the issues still to be addressed on the Mars Science Laboratory as it prepares to launch. From what I've read, this seems normal at this stage (read Steven Squyres' book on the Mars Exploration Rovers for more examples of pre-launch issues).
I once had lunch with a planetary scientist who began his career in the 1950s. He told me that in the late 1950s, he came to believe we would never send probes to other planets because statisticians had shown that you could never achieve the reliability necessary for the missions to succeed. Of course, the various space agencies proved him wrong (often with dramatic failures that served as learning opportunities). When I think of the great advances brought by the space age, I am less impressed by the technical wonders and more impressed by the management systems that can deal with these levels of complexity, especially for missions such as MSL that expand the technical capabilities on so many fronts. This audit is part of the management system, in this case providing an outside review of issues.