I realize there are lots of "what-if" failures that can combine in instrument failure conditions. However the liklihood of such combinations need be consider in the design and mitigated to better then 10E-9 by redundancy or the ability of the pilot to still be able to effect a safe landing.
If we can't rely on the pilot to use his trained resource management than we shouldn't be flying. Now I don't know what has gone wrong in this accident, but before we start blaming the design, we need to find out what went wrong between the interface of the machine and the pilot