3 ms·

But would this prevent the next Spectre? Not necessarily. Hardware vendors would continue to optimize execution of machine code in unexpected ways, including wa

But would this prevent the next Spectre? Not necessarily. Hardware vendors would continue to optimize execution of machine code in unexpected ways, including ways that allow for side channel attacks. They wouldn't be optimizing for C code, but they would still be optimizing for some portable low-level language. Unless we made that low-level language constantly change as hardware microarchitecture improves (thus giving up on portability), I think we'd be back with the same problem of a mismatch between low-level languages and underlying CPU architecture.

cryptonector 7y ago
TFA doesn't even propose a non-serial programming paradigm that programmers can easily learn. I suppose array languages, maybe? It's hard to imagine a CPU architecture where there's no branches, or minimal branches, or a language that greatly minimizes them yet can be used to implement the sorts of software we're fond of. A lot of what we do with computers is highly parallel, but also highly serial with lots of logic -- is that the fault of C, or is that natural and C is just a scapegoat? My money is on the latter.