I pointed out the same thing recently in another thread http://www.canonrumors.com/forum/index.php?topic=10866.90
Both manufacturers have had their issues and they will continue to struggle with quality control, lets face it this stuff gets ever more complex and tolerances get ever smaller, manufacturing process upgrading is, like everything else, a very steep curve of quality over cost and as these threads demonstrate consumers don't like paying more!
What does seem surprising is the number of issues that don't show up during real world pre-release product testing, makes you wonder if they do actually get photographers to test half this stuff sometimes. Maybe that is why there is this interminable wait for the 200-400, Canon know this could be a 1D MkIII AF make or break situation.
Although some of the flaws seem obvious by 2020 hindsite, the real issue is the number of users using a camera. A dozen handmade test cameras passed around to photographers to use for 24 hours and that take maybe 25K images each is not going to find any but the most gross issues. They cannot take raw images either, so 0 testing there.
Once the first 100,000 cameras are shipped and delivered over a period of 3-4 weeks, suddenly, there are millions of images taken and the more subtle issues are found. Some are design issues, and some are assembly issues or material issues.
Its unlikely that Canon would put even 1000 hand made cameras out for photographers to test for the 3 weeks or so that it takes to find issues.
QA merely follows a test plan that has been developed by experiences with previous equipment over several years, and are likely not allowed to play around with a camera doing things that were not in the test plan.
In many cases, they merely verify that manufacturing has a good set of instructions for assembling and testing the camera, and that employees have it and know how to use it. They spot check to see that they are using the assembly instructions.
Its a built-in blindness in the system.
Our company let the engineers test and verify subcontracted work, but the engineers could not test or verify internal manufacturing. That was QA/QC responsibility, and they were very possesive. As a result many big dollar assembly problems happened that the designer would have spotted immediately because he understood the function. QA merely verified that the written processes were followed, even if they were wrong, and the designer could not specify many of the common tests because it cost too much!
Even so, the manufacturing folks managed to overcome those issues and make things work properly. I've seen problems crop up after 10 years that should be founnd immediately, but the assembler fixed them for 10 years, and when a new assembler moved to the job, he did not know to fix the issue and things did not work.