The premise appears to be that using an old, simple, easy-to-verify cpu/architecture makes it harder to embed malware in the nuclear-disarmament-verification code that hostile (or at least somewhat untrusting) states would both need to trust.
> ... as the performance of the inspection system is limited, and data acquisition and processing has to be designed and highly optimized accordingly.
I wonder if any potential exploit/workaround/disarmament-dodging would then just move to a different layer.
> ... as the performance of the inspection system is limited, and data acquisition and processing has to be designed and highly optimized accordingly.
I wonder if any potential exploit/workaround/disarmament-dodging would then just move to a different layer.
This also reminds me of the most recent Underhanded C contest at http://www.underhanded-c.org/_page_id_5.html